Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

When ‘I’ belong I don’t care about ‘you’ : the role of self-construal and social inclusion in pro-social… South, Cluny 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_september_south_cluny.pdf [ 8.5MB ]
Metadata
JSON: 24-1.0371191.json
JSON-LD: 24-1.0371191-ld.json
RDF/XML (Pretty): 24-1.0371191-rdf.xml
RDF/JSON: 24-1.0371191-rdf.json
Turtle: 24-1.0371191-turtle.txt
N-Triples: 24-1.0371191-rdf-ntriples.txt
Original Record: 24-1.0371191-source.json
Full Text
24-1.0371191-fulltext.txt
Citation
24-1.0371191.ris

Full Text

		 WHEN	‘I’	BELONG	I	DON'T	CARE	ABOUT	‘YOU’:	THE	ROLE	OF	SELF-CONSTRUAL	AND	SOCIAL	INCLUSION	IN	PRO-SOCIAL	BEHAVIOUR	DIRECTED	AT	ANIMAL	OUT-GROUP	RECIPIENTS	by		CLUNY	SOUTH		B.A.	(Hon.),	Central	St	Martins,	The	University	of	the	Arts	London,	1987		A	THESIS	SUBMITTED	IN	PARTIAL	FULFILLMENT	OF		THE	REQUIREMENTS	FOR	THE	DEGREE	OF	DOCTOR	OF	PHILOSOPHY	in	THE	FACULTY	OF	GRADUATE	AND	POSTDOCTORAL	STUDIES	(Interdisciplinary	Studies)	[Marketing/	Forest	Resources	Management	/	Psychology]		THE	UNIVERSITY	OF	BRITISH	COLUMBIA	(Vancouver)		August	2018	©	Cluny	South,	2018		 ii	The	following	individuals	certify	that	they	have	read,	and	recommend	to	the	Faculty	of	Graduate	and	Postdoctoral	Studies	for	acceptance,	the	dissertation	entitled:		When	‘I’	belong	I	don’t	care	about	‘you’:	The	role	of	self-construal	and	social	inclusion	in	pro-social	behaviour	directed	at	animal	out-group	recipients.			submitted	by	 Cluny	South	 	 in	partial	fulfillment	of	the	requirements	for	the	degree	of	 Doctor	of	Philosophy	in	 Interdisciplinary	Studies	(Marketing,	Forest	Resources	Management,	Psychology)		Examining	Committee:	Michael	Meitner,	Forest	Resources	Management	Co-supervisor	Katherine	White,	Sauder	School	of	Business	Co-supervisor		Hal	Herzog,	Psychology,	WCU	Supervisory	Committee	Member	Mark	Schaller,	Psychology	University	Examiner	Marc-David	L.	Seidel,	Sauder	School	of	Business	University	Examiner			Additional	Supervisory	Committee	Members:		Supervisory	Committee	Member				 		 iii	Abstract		This	dissertation	is	aimed	at	improving	understanding	of	the	mechanisms	at	play	in	prosocial	behaviour,	as	a	function	of	recipient	group	identity	(i.e.	in-group	vs.	out-group	member),	self-construal,	and	social	inclusion.	While	some	research	has	examined	the	impacts	of	self-construal	on	prosocial	behaviour,	as	well	as	the	downstream	reactions	to	threats	of	exclusion,	little	research	has	looked	at	the	impact	that	perceptions	of	inclusion	may	have	on	prosocial	behaviour.	Moreover	even	less	research	has	looked	at	how	promises,	or	reminders,	of	social	inclusion	may	be	specifically	experienced	by	individuals	with	high	independent	self-construal,	and	how	this	may	impact	their	subsequent	prosocial	behaviour	towards	out-group	(vs.	in-group)	targets	of	concern.	The	goal	of	this	dissertation	is	to	explore	the	interaction	between	inclusion	and	independent	self-construal,	with	a	focus	on	the	downstream	consequences	for	prosocial	behaviour	directed	at	out-groups.				Across	five	experimental	studies	I	demonstrate	that	individuals	with	high	independent	self-construal	may	behave	more	prosocially	towards	out-group	targets	(more	inclusively	in	two	grouping	tasks,	and	more	prosocially	in	two	donation	tasks)	under	normal	conditions,	in	comparison	with	individuals	high	in	interdependent	self-construal.		I	also	offer	evidence	that	following	an	affirmation	of	social	inclusion	status	the	pattern	reverses	and	individuals	with	high	independent	self-construal	behave	less	prosocially	towards	out-group	targets.	Furthermore,	I	provide	some	tentative	evidence	to	support	the	argument	that	individuals	with	high		 iv	independent	self-construal	may	be	motivated	to	behave	prosocially	towards	out-group	targets	in	order	to	maximise	social	connection	potential,	and	that	feelings	of	similarity	may	increase	this.	Finally,	I	demonstrate	that	feelings	of	connection	to	cause	may	mediate	these	mechanisms	in	the	case	of	donation	intentions.				Taken	together	this	dissertation	builds	on	previous	research,	and	then	extends	it	to	demonstrate	that	while	individuals	with	high	independent	self-construal	may	behave	more	prosocially	to	out-groups	under	normal	circumstances,	promises	or	reminders	of	inclusion	may	reverse	this	pattern,	decreasing	prosocial	behaviour.	I	provide	some	preliminary	evidence	that	the	increase	(decrease)	in	prosocial	behaviour	is	as	a	result	of	increased	(decreased)	motivation	for	social	interaction.						 		 v	Lay	Summary	In	this	research	I	demonstrate	that	under	normal	circumstances,	individuals	that	are	high	in	individual	self-construal	(having	a	sense	of	self	that	is	more	self	focused	as	opposed	to	more	other	focused)	behave	in	a	more	prosocial	way	towards	out-groups	and	distant	others.	However,	following	a	reminder,	or	assurance,	that	they	belong	and	are	socially	included,	individuals	that	are	high	in	individual	self-construal	behave	in	a	less	prosocial	way	towards	out-groups	and	distant	others.	I	show	that	perceptions	of	similarity	and	connection	between	donors	and	recipients	of	prosocial	behaviour	alter	the	responses	by	donors,	and	argue	that	their	behaviour	may	be	driven,	in	part,	by	a	desire	to	maintain	a	sense	of	belonging,	or	to	improve	social	connection	with	others.		       	 		 vi	Preface	I	am	the	primary	author	of	the	work	presented	in	this	Ph.D.	dissertation.	I	was	responsible	for	conducting	the	literature	review,	developing	the	hypotheses,	designing	the	experiments,	collecting	the	data,	analyzing	the	data,	and	preparing	the	manuscript.	Katherine	White	and	Michael	Meitner	assisted	in	designing	the	experiments.	Katherine	White,	Michael	Meitner,	and	Hal	Herzog	provided	intellectual	contributions.	None	of	the	text	of	the	dissertation	is	taken	directly	from	previously	published	or	collaborative	articles.		The	research	presented	in	this	dissertation	was	supported	by	Standard	Research	Grants	from	the	Social	Sciences	and	Humanities	Research	Council	(SSHRC),	awarded	to	Katherine	White.			Ethical	Approval	for	all	experimental	studies	was	obtained	from	the	UBC	Office	of	Research	Services	Behavioural	Review	Board	(Human	Ethics)	under	the	following	certificate	number:	H14-00190.			 		 vii	Table	of	Contents		Abstract………………………………..…………………………………………………………………..……iii	Lay	Summary……..…………………..………………………………………………………………………..v	Preface………………………………..………………………………………………………………………....vi	Table	of	Contents………………………………..……………………………………………………..….vii	List	of	Figures………………………………..………………………………………………………...……xii	List	of	Acronyms………………………………………………………………………………….…….….xv	Acknowledgements………………………………..……………………………………………..….…..xvi	Dedication………………………………..……………………………………………………………..….xviii	Chapter	1:	Introduction………………………………..………………………………………………....1	1.1	Overview	of	the	Dissertation……………………………………………………………..….4	Chapter	2:	When	I	Belong	I	Don't	Care	About	You:	The	Role	of	Self-Construal	and	Social	Inclusion	in	Prosocial	Behaviour	Directed	at	Animal	Out-group	Recipients…………………………………………………………………………………………………….…8	2.1	Introduction	of	Dissertation	Topic…………………………………………………..……8	2.2	Conceptual	Background……………………………………………………………………...11	2.2.1	Social	Identity	and	Self-Construal:	Our	Relationships	with	Others…11	2.2.2	Exclusion	and	Belonging:	Impacts	of	Inclusion	on	Sense	of	Self…..…24	2.2.3	Prosocial	Behaviour:	Mechanisms	and	Motivations	for	Donation	Behaviour	and	Inclusive	Attitudes	Explored…………………………………………34	Chapter	3:	Empirical	Investigation……………………………………………………………..…47	3.1	Pre-test	Study………………………………………………………………………………….…47		viii	3.1.1	Overview…………………………………………………………………………………….47	3.1.2	Procedure……………………………………………………………………………….......48	3.1.3	Results………………………………………………………………………………………..49	3.1.4	Discussion…………………………………………………………………………………..51	3.2	Study	1………………………………………………………………………………………………51	3.2.1	Overview…………………………………………………………………………………….51	3.2.2	Procedure…………………………………………………………………………………...55	3.2.3	Results………………………………………………………………………………………..58	3.2.4	Discussion…………………………………………………………………………………..60	3.3	Study	2………………………………………………………………………………………………63	3.3.1	Overview…………………………………………………………………………………….63	3.3.2	Procedure…………………………………………………………………………………...65	3.3.3	Results………………………………………………………………………………………..71	3.3.3.1	Role	of	Exclusion	Manipulation	in	Grouping,	and	Anthropomorphism	Tasks……………………………………………………………….74	3.3.3.2	Role	of	Inclusion	Manipulation	in	Grouping,	and	Anthropomorphism	Tasks……………………………………………………………….78	3.3.4	Discussion…………………………………………………………………………………..83	3.4	Study	3………………………………………………………………………………………………91	3.4.1	Overview…………………………………………………………………………………….91	3.4.2	Procedure……………………………………………………………………………….......94	3.4.3	Results…………………………….………………………………………………….……....98		 ix	3.4.3.1	Role	of	Positive	Affect	Manipulation	in	Donation	Intentions……………………………………………………………………………………..102	3.4.3.2	Role	of	Inclusion	Manipulation	(vs.	Control	+	Affect)	in	Donation	Intentions……………………………………………………………………………………..103	3.4.3.3	Role	of	Inclusion	Manipulation	(vs.	Control	+	Affect)	on	Empathy	and	Connection	to	Cause	as	Dependent	Variables………………………..….106	3.4.3.4	Tests	of	Mediated	Moderation……………………………………….….…109	3.4.4	Discussion…………………………………………………………………………….…..112	3.5	Study	4……………………….……………………………………………………………….…...120	3.5.1	Overview…………………………………………………………………………………..120	3.5.2	Procedure………………………………………………………………………………....126	3.5.3	Results……………………………………………………………………………………...131	3.5.3.1	Role	of	Inclusion	Manipulation,	Poster	Frame,	and	Interdependent	Self-Construal	in	Donation	Intentions…………………....135	3.5.3.2	Role	of	Inclusion	Manipulation	(vs.	Control),	Poster	Frame,	and	Independent	Self-Construal	in	Donation	Intentions………………………...137	3.5.3.3	Role	of	Inclusion	Manipulation	(vs.	Control),	Poster	Frame,	and	Independent	Self-Construal	in	Actual	Cash	Donation………………………141	3.5.3.4	Role	of	Inclusion	Manipulation	(vs.	Control),	Poster	Frame,	and	Independent	Self-Construal	in	Connection	to	Cause………………………..144	3.5.3.5	Role	of	Inclusion	Manipulation	(vs.	Control),	Poster	Frame,	and	Independent	Self-Construal	in	Empathy	for	Cause…………………………..148			 x	3.5.3.6	Role	of	Inclusion	Manipulation	(vs.	Control),	Poster	Frame,	and	Independent	Self-Construal	in	Similarity	to	Cause…………………………..153	3.5.3.7	Tests	of	Mediated	Moderation……………………………………………..158	3.5.4	Discussion…………………………………………………………………………………163	3.6	Study	5…………………………………………………………………………………………….173	3.6.1	Overview…………………………………………………………………………………..173	3.6.2	Procedure…………………………………………………………………………………173	3.6.3	Results……………………………………………………………………………………...176	3.6.3.1	Role	of	Poster	and	Self-Construal	in	Donation	Intentions,	Cash	Donation,	and	Empathy…………………………………………………………………179	3.6.4	Discussion…………………………………………………………………………………183	Chapter	4:	General	Discussion……………………………………………………………...........187	4.1	General	Discussion…………………………………………………………………………...187	4.2	Theoretical	Implications……………………………………………………………….…..190	4.3	Practical	Implications……………………………………………………………………….192	4.4	Research	Limitations	and	Directions	for	Future	Research….………………195	4.5	Concluding	Remarks………………………………………………………………………...200	Chapter	5:	Figures……………………………………………………………………………………….202	References…………………………………………………………………………………………………..239	Appendix:	Experimental	Materials	&	Follow-Up	Analysis……………………………263	A.1	Pre-test:	Follow-Up	Analysis…………………………………………………………….263	A.2	Self-Construal	Scale………………………………………………………………………….264	A.3	Study	1:	Scales…………………………………………………………………………………265		 xi	A.4	Study	1:	Experimental	Grouping	Task	Instructions……………………………266	A.5	Study	2:	Scales…………………………………………………………………………………267	A.6	Study	2:	Future	Life	Manipulation…………………………………………………….268	A.7	Study	2:	Anthropomorphism	Picture	&	Task	Instructions………………….269	A.8	Study	2:	Follow-Up	Analysis……………………………………………………………..270	A.9	Study	3:	Scales…………………………………………………………………………………291	A.10	Study	3:	Writing	Manipulation	Task	Instructions…………………………….292	A.11	Study	3:	Charity	Poster	&	Task	Instructions…………………………………….293	A.12	Study	3:	Follow-Up	Analysis…………………………………………………………...294	A.13	Study	4:	Need	Fulfillment	Scale………………………………………………………308	A.14	Study	4:	Poster	Frame.……………………………………………………………………309	A.15	Study	4:	Follow-Up	Analysis………………………………………………………...…310	A.16	Study	5:	Poster	Manipulation…………………………………………………….……320	A.17	Power	Analysis	for	Studies……………………………………………………………..321												 xii	List	of	Figures		Figure	1.	Social	Identity	vs.	Personal	Identity	.………………………………………………...…202	Figure	2.	Optimal	Distinctiveness	Model	(ODM)	of	the	Self………………………………...203	Figure	3.	Construals	of	the	Self………………………………………………………………………….204	Figure	4.	Self-Construal	and	Prosocial	Behaviour	Path	(Studies	1,	2,	and	3)………..205	Figure	5.	Self-Construal	and	Prosocial	Behaviour	Path	(Study	4)………………………..206	Figure	6.	Sample	Ethnicity	by	Percentage	(Pre-test	Study)…………………………………207	Figure	7.	Sample	Age	by	Percentage	(Pre-test	Study)…………………………………………208	Figure	8.	Sample	Ethnicity	by	Percentage	(Study	1)………………………………………...…209	Figure	9.	Sample	Age	by	Percentage	(Study	1)………………………………………………...…210	Figure	10.	Sample	Ethnicity	by	Percentage	(Study	2)…………………………………………211	Figure	11.	Proportion	of	In-group	made	up	of	Animals	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	2)…………………………………………..212	Figure	12.	Floodlight	Analysis	of	Proportion	of	In-group	made	up	of	Animals	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	2)…………213	Figure	13.	Entitativity	Scores	of	the	In-group	as	a	Function	of	Condition	(Inclusion	vs.	Control)	and	Independent	Self-Construal	(Study	2)………………………………………214	Figure	14.	Anthropomorphism	of	Dog	Picture	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	2)…………………………………………215	Figure	15.	Floodlight	Analysis	of	Ratings	of	Anthropomorphism	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	2)……….216	Figure	16.	Sample	Ethnicity	by	Percentage	(Study	3)…………………………………...…….217		xiii	Figure	17.	Sample	Age	by	Percentage	(Study	3)……………………………………..…………..218	Figure	18.	Donation	Intention	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	3)……………………………………………………………….219	Figure	19.	Floodlight	Analysis	of	Donation	Intentions	as	a	Function	of	Condition	(Control	+	Affect	vs.	Inclusion)	and	Independent	Self-Construal	(Study	3)………….220	Figure	20.	Connection	to	Cause	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	3)……………………………………………………………………………..221	Figure	21.	Sample	Ethnicity	by	Percentage	(Study	4)………………………………………....222	Figure	22.	Sample	Age	by	Percentage	(Study	4)……………………………………………...….223	Figure	23.	Donation	to	Cause	as	a	Function	of	Poster	Frame	(Similar	vs.	Different)	and	Interdependent	Self-Construal	(Study	4)…………………………………………………….224	Figure	24.	Floodlight	Analysis	of	Donation	Intentions	as	a	Function	of	Poster	Style	(Similar	vs.	Different)	and	Interdependent	Self-Construal	(Study	4)…………………...225	Figure	25.	Donation	to	Cause	as	a	Function	of	Condition	(Control	vs.	Inclusion),	Poster	Type	(Similar	vs.	Different),	and	Independent	Self-Construal	(Study	4)…....226	Figure	26.	Floodlight	Analysis	of	Donation	Intention	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)………………………………………………………………………………………..…227	Figure	27.	Actual	Cash	Donation	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)…...…228	Figure	28.	Connection	to	Cause	as	a	Function	of	Independent	Self-Construal,	Inclusion	Manipulation,	and	Similar	vs.	Different	Poster	Frame	(Study	4)……….....229		xiv	Figure	29.	Floodlight	Analysis	of	Connection	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)………………………………………………………………………………………..…230	Figure	30.	Empathy	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)……...231	Figure	31.	Floodlight	Analysis	of	Empathy	for	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)…………………………………………………………………………………………..232	Figure	32.	Similarity	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)………233	Figure	33.	Floodlight	Analysis	of	Similarity	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion),	and	Independent	Self-Construal	(Study	4)…………………………………………………………………………………………..234	Figure	34.	Sample	Age	by	Percentage	(Study	5)…………………………………………………236	Figure	35.	Conceptual	Models	(Hayes,	2013;	2018)	Used	in	Studies	2-5…………...…237										 		 xv	List	of	Acronyms		ANOVA	 analysis	of	variance	BC	SPCA	 British	Columbia	Society	for	the	Prevention	of	Cruelty	to	Animals	FNE		 	 fear	of	negative	evaluations	(scale)	HAI	 	 human	animal	interactions	IDA	 	 individual	differences	in	anthropomorphism	IDAQ	 	 individual	differences	in	anthropomorphism	questionnaire	(scale)	IHSR		 	 Intergroup	Helping	as	Status	Relations	INQ	 	 interpersonal	needs	questionnaire	(scale)	JN	 	 Johnson-Neyman	MMR	 	 moderated	multiple	regression	MTurk		 Mechanical	Turk	NTB	 	 need	to	belong	(scale)	ODT	 	 Optimal	Distinctiveness	Theory	SCT	 	 Self-Categorization	Theory	SD	 	 social	desirability	(scale)	SIT	 	 Social	Identity	Theory	SMS	 	 Social Monitoring System	SPSS	 	 Statistical	Package	for	the	Social	Sciences	THWB		 thwarted	belonging	(scale)	WEIRD	 western,	educated,	industrialised,	rich	and	democratic	(societies)		 		xvi	Acknowledgements		These	years	spent	at	UBC	came	at	a	time	in	my	life	when	I	was	changing	direction	and	embarking	on	a	new	adventure.	Many	people	have	supported	and	encouraged	me	along	the	way.		To	my	supervisor,	Kate	White,	I	owe	enormous	thanks.	I	know	I	wasn't	always	the	easiest	student,	but	you	nonetheless	offered	me	amazing	support,	resources	and	encouragement	over	the	many	years	it	has	taken	to	get	here.	You've	shaken	me	out	of	my	comfort	zone	time	and	again,	and	encouraged	me	to	challenge	myself	and	to	achieve	more	than	I	ever	thought	would	be	possible.	It	has	been	really	exciting	learning	with	you	and	you	have	opened	my	mind	to	a	fascinating	body	of	research.	Thank	you	so	very	much.	I	would	also	like	to	thank	my	other	supervisor	Michael	Meitner.	Mike,	you	were	more	than	open	minded	and	generous	as	I	blundered	around,	trying	to	find	the	right	direction	to	take	my	research.	You	offered	me	many	wonderful	opportunities,	and	opened	my	mind	to	interesting	and	engaging	projects.	It’s	always	been	a	lot	of	fun	working	with	you,	and	I	am	grateful	for	the	time	I	have	spent	doing	so,	as	well	as	the	many	chances	you	have	given	me	to	learn	and	grow,	sometimes	at	your	expense.	I	owe	you!	To	my	outside	supervisor,	Hal	Herzog,	I	thank	you	for	your	outrageous	enthusiasm	and	fantastic	sense	of	humour.	Your	inspirational	presence	in	the	field	of	anthrozoology	both	lit	my	thirst	for	research	in	this	area	and	then	fanned	the	flames.	As	well	as	encouraging	me	to	be	brave	and	bold,	you've	been	a	fantastically	supportive	member	of	my	committee	and	a	steadying	influence	at	times.	I	am	deeply	honoured,	and	very	lucky,	to	have	had	you	on	my	team.		Finally,	a	special	thanks	goes	to	my	very	first	supervisor,	Paul	Wood,	who	believed	in	me	from	the	outset	and	encouraged	me	to	take	this	path	from	the	beginning.		xvii	Although	you	were	not	with	me	at	the	end,	I	thank	you	immensely	for	your	very	kind	and	personal	support	that	set	me	on	this	adventure.		In	addition	to	my	supervisory	committee,	I	would	like	to	say	thank	you	to	everyone	in	the	Marketing	Division	at	Sauder,	especially	Darren	Dahl,	Joey	Hoegg	and	Sandra	Robinson,	for	their	kind	support	over	the	years.	Thank	you	also	to	Hillel	Goelman,	and	more	recently	Barbara	Weber,	at	the	ISGP.	I’d	also	like	to	thank	all	the	staff	at	Sauder,	Forestry	and	ISGP.	In	particular,	I	would	like	to	thank	Florence	Yen	for	her	wonderful	support	and	kindness	at	Sauder,	and	for	taking	care	of	me.		A	final	UBC	shout	out	must	also	go	to	my	fellow	PhD	students,	and	especially	Yoonji	Shim	and	Johannes	Boegershausen,	who	were	always	very	generous	with	their	time.				Lastly,	I	would	like	to	thank	all	my	friends	and	family,	who	have	supported	me	and	cheered	me	on	over	the	years.	Thank	you	to	my	partner,	Rick,	who	was	always	there	at	my	darkest	hours;	you	are	my	rock,	as	well	as	the	love	of	my	life.	Endless	thanks	also	to	my	parents	for	encouraging	me	to	be	brave	and	fearless	and,	in	the	words	of	my	father,	to	“murder	the	bums.”		I	may	not	be	making	the	front	page	of	the	San	Francisco	Chronicle,	but	I	think	I	won	the	game!	I	also	want	to	pay	a	posthumous	tribute	to	my	grandfather,	Charles	Fairfoot,	who	always	encouraged	me	to	keep	on	going	even	when	life	was	tough,	and	whose	immortal	phrase	“don’t	let	the	buggers	get	you	down!”	still	spurs	me	on	when	times	are	tough.	You	were	the	best	granddad.	Thanks	also	to	my	kids	for	putting	up	with	a	student	mother,	which	can	be	even	more	embarrassing	than	having	a	non-student	mother	J.	Finally,	thanks	to	my	friends	in	Canada	and	England	who	encouraged	me	on	my	crazy	dream—I	am	lucky	to	have	you	all.	I	hope	I	have	shown	you	that	by	doing	this	that	pretty	much	anything	is	possible,	if	you	try	hard	enough!					 		xviii	Dedication		This	dissertation	is	dedicated	to	my	family,	without	whom	I	would	be	a	lesser	person.	To	my	partner,	Rick	Gibson,	my	children,	Orlanda	South	and	Arlo	Maguire,	and	my	parents,	Peter	and	Marie	South,	I	would	like	to	acknowledge	the	part	you	all	played,	and	the	price	you	all	paid,	for	this	research	work.		Any	success	I	accrue	as	a	result	of	this	work,	I	share	with	you.	Any	pride	I	acknowledge	as	a	result	of	this	work,	I	share	with	you.		For,	in	the	words	of	the	English	poet	John	Donne,	no	man	is	an	island.			"No	man	is	an	island	entire	of	itself;	every	man		is	a	piece	of	the	continent,	a	part	of	the	main;		if	a	clod	be	washed	away	by	the	sea,	Europe		is	the	less,	as	well	as	if	a	promontory	were,	as		well	as	any	manner	of	thy	friends	or	of	thine		own	were;	any	man's	death	diminishes	me,		because	I	am	involved	in	mankind.		And	therefore	never	send	to	know	for	whom		the	bell	tolls;	it	tolls	for	thee."		 John	Donne,	1624		 1	Chapter	1:	Introduction			Over	a	quarter	of	a	century	of	research	(see	Baumeister	&	Leary,	1995;	Williams,	2009;	Leary	&	Cottrell,	2013	for	narrative	literature	reviews)	has	been	devoted	to	examining	the	effects	of	social	isolation	on	the	human	psyche,	and	in	particular	the	impact	that	social	exclusion	may	have	for	subsequent	social	interactions,	such	as	prosocial	behaviour.		A	few	researchers	(e.g.,	Brewer,	1991;	DeWall,	Baumeister,	&	Vohs,	2008)	have	also	looked	at	the	impact	that	belonging	satiation	and	reassurances	of	social	inclusion	may	have	for	subsequent	interactions,	including	attitudes	to	distant	humans	and	even	other	species	(Waytz	&	Epley,	2012;	Waytz,	2013).	Results	from	the	research	into	the	impacts	of	exclusion	and	inclusion	on	downstream	reactions,	however,	have	been	mixed.		Some	research	has	found	that	exclusion	increases	antisocial	behaviour	(Twenge,	Baumeister,	Tice,	&	Stucke,	2001),	and	reduces	prosocial	behaviour	and	cooperation	(Catanese	&	Tice,	2005;	Twenge,	Baumeister,	DeWall,	Ciarocco,	&	Bartels,	2007;	van	Beest	&	Williams,	2011),	whilst	other	research	has	demonstrated	that	prosocial	behaviour	may	increase	following	exclusion	(Lee	&	Shrum,	2012),	especially	if	an	opportunity	to	reconnect	still	exists	(Maner,	DeWall,	Baumeister,	&	Schaller,	2007;	Cuadrado,	Tabernero,	&	Steinel,	2015).				Alongside	this	work	on	social	exclusion	and	belonging,	another	separate	but	related	body	of	research	has	been	explicitly	examining	the	motives	for	prosocial	behaviour	(Batson,	1987;	Batson,	Ahmad,	&	Stocks,	2011;	Duclos	&	Barasch,	2014;	Martela	&		 2	Ryan,	2016).	In	particular,	researchers	in	this	stream	have	focused	on	how	prosocial	motivation	may	differ	depending	on	the	status	and	group	membership	of	both	donor	and	target	(Brewer,	1979;	Hopkins,	Reicher,	Harrison,	Cassidy,	Bull,	&	Levine,	2007;	Jonas,	Schimel,	Greenberg,	&	Pyszczynski,	2002;	Leeuwen	&	Tauber,	2012;	Sturmer	&	Snyder,	2010;	Tajfel	&	Turner,	1986),	as	well	as	examining	how	behaviour	may	be	impacted	by	perceived	similarity	or	closeness	to	the	cause	(Burnstein,	Crandall,	&	Kitayama,	1994;	Sturmer	&	Snyder,	2010).				Further	recent	interest	has	focused	on	the	moderating	effect	that	self-construal	orientation	may	have,	both	on	threats	to	belonging	(Pfundmair,	Graupmann,	Du,	Frey,	&	Aydin,	2014;	Pfundmair,	Graupmann,	Frey,	&	Aydin,	2015;	Ren,	Wesselmann,	&	Williams,	2013),	as	well	as	on	group	behaviour	(Yuki	&	Takemura,	2013),	and	prosocial	behaviour	directed	towards	both	in	and	out-group	targets	(Duclos	&	Barasch,	2014;	Kemmelmeier,	Jambor,	&	Letner,	2006).	Self-construal	may	best	be	defined	as	what	people	believe	about	the	relationship	between	the	self	and	others,	and	specifically	how	separate	or	connected	an	individual	is	to	others	around	them.	Two	views	of	self-construal	are	usually	distinguished;	that	of	independent	self-construal,	which	concerns	an	individual’s	self-determination	and	distinctiveness,	vs.	that	of	interdependent	self-construal,	which	concerns	an	individual’s	shared	self	and	maintenance	of	relationships	with	others	(Markus	&	Kitayama,	1991;	2010).	Traditionally	individuals	with	highly	independent	self-construal	are	considered	to	be	more	likely	to	behave	according	to	their	own	preferences	and	for	their	personal	pleasure,	compared	to	those	with	high		 3	interdependent	self-construal,	who	are	thought	to	be	more	likely	to	value	interpersonal	relationships,	and	others’	feelings	and	evaluations	(Sung,	Choi,	&	Tinkham,	2012).	However,	when	it	comes	to	prosocial	behaviour,	it	appears	that	individuals	with	high	independent	self-construal	are	just	as,	if	not	more,	charitable	towards	others	(Kemmelmeier,	Jambor,	&	Letner,	2006),	and	are	distinctly	more	likely	to	behave	prosocially	to	out-group	members.	Individuals	high	in	interdependent	self-construal,	on	the	other	hand,	appear	to	show	a	marked	in-group	bias	(Duclos	&	Barasch,	2014).			This	dissertation	is	focused	on	the	intersection	of	these	bodies	of	research;	prosocial	behaviour	directed	towards	out-groups,	as	a	factor	of	social	inclusion	status,	donor	self-construal	orientation,	and	social	identity.	The	hope	of	this	research	is	to	forward	an	argument	as	to	why	prior	research,	investigating	the	downstream	prosocial	responses	following	inclusion	and	exclusion,	may	have	delivered	divergent	results,	and	to	make	some	progress	towards	filling	in	gaps	in	our	knowledge	regarding	this	important	area	of	research.	Specifically,	I	propose	that	at	least	some	of	the	divergent	past	findings	on	prosocial	behaviour	patterns	towards	out-groups	may	stem	from	a	failure	to	account	for	the	interaction	between	self-construal	and	social	inclusion	status	and,	in	particular,	the	motivational	impact	that	a	need	for	social	inclusion	has	for	individuals	that	are	high	in	independent	self-construal.				 4	Furthermore,	I	argue	that	the	social	identity	of	the	recipient	of	the	prosocial	behaviour	(as	compared	to	the	donor)	is	of	crucial	importance.	In	addition,	I	offer	forward	evidence	that	although	individuals	with	high	independent	self-construal	are	considered	to	show	less	in-group	bias	(Duclos	&	Barasch,	2014),	they	may	nonetheless	behave	as	poorly	towards	out-group	targets	as	do	individuals	with	high	interdependent	self-construal,	under	specific	(inclusion)	conditions.		Finally,	I	bring	a	novel	out-group	target,	non-human	animals,	to	the	table,	in	order	to	investigate	how	perceptions	of	similarity	to	an	out-group	may	mediate	prosocial	intentions	for	individuals	that	are	high	in	independent	self-construal.	In	summary,	my	research	seeks	to	further	the	understanding	of	the	impact	that	self-construal	orientation	has	on	inclusive	behaviours	and	prosocial	intentions	towards	out-group	members,	as	a	result	of	social	inclusion	status.			1.1	Overview	of	Dissertation			This	dissertation	is	comprised	of	one	essay	that	includes	six	studies	(one	pre-test	study,	one	correlational	study,	and	four	experimental	studies,	including	one	experimental	field	study).		The	essay	seeks	to	build	on	previous	research	and	further	probe	the	motivations	driving	prosocial	behaviour	towards	distant	others,	including	non-humans,	as	a	function	of	social	inclusion	status,	and	to	examine	the	moderating	influence	of	self-construal	orientation.			 5	To	this	end,	I	first	present	evidence	that	individuals	show	a	propensity	to	gauge	others	around	them,	including	both	humans	and	animals,	in	terms	of	their	potential	to	offer	opportunities	for	social	interaction	and	connection	(pre-test	study).	I	also	show	that	independent	self-construal	correlates	with	anthropomorphic	tendencies		(assignment	of	human	attributes	to	non-human	others).	I	then	seek	to	demonstrate	how	self-construal	orientation	may	impact	the	formation	of	in-groups	and	out-groups	(study	1),	and	show	that	a	positive	relationship	exists,	between	independent	self-construal	orientation	and	the	inclusion	of	animal	out-group	members	into	a	self-formed	in-group.	I	propose	that	this	happens	for	two	reasons;	first,	because	individuals	that	are	high	in	independent	self-construal	have	more	relaxed	in-group	boundaries	and	are	more	open	to	admitting	others,	and	second	because	individuals	that	are	high	in	independent	self-construal	will	be	more	likely	to	view	animals	from	a	perspective	of	having	social	connection	potential.			Building	on	from	this,	I	replicate	the	findings	of	study	1	in	an	experimental	lab	setting	(study	2),	and	then	go	on	to	demonstrate	the	moderating	influence	of	social	inclusion	status,	especially	for	individuals	that	are	high	in	independent	self-construal.	Specifically,	study	2	provides	evidence	that	an	affirmation	of	social	inclusion	may	result	in	the	consolidation	of	in-group	boundaries,	and	a	decrease	in	the	ratings	of	anthropomorphism	(assignment	of	human	attributes)	given	to	animals,	for	individuals	that	are	high	in	independent	self-construal.	Furthermore,	study	2	also	offers	evidence	to	support	the	claim	that	a	high	interdependent	self-	 6	construal	orientation	may	provide	a	buffer,	both	against	threats	of	social	exclusion	and	promises	of	social	inclusion.			Following	this,	I	then	extend	these	findings	(study	3),	to	an	important	new	dependent	variable,	donation	intentions	for	an	out-group	animal	charity,	and	offer	evidence	for	the	role	that	both	social	inclusion	status,	as	well	as	self-construal	orientation,	may	play	in	this	setting.	Specifically,	I	provide	evidence	that	donation	intentions	are	higher	under	normal	conditions,	but	are	reduced	following	assurance	of	social	inclusion,	for	individuals	that	are	high	in	independent	self-construal.	However,	I	show	that	the	same	effect	is	not	seen	for	individuals	that	are	high	in	interdependent	self-construal.	I	propose	that	this	happens	as	a	result	of	individuals	who	are	high	in	independent	self-construal,	giving	to	an	out-group	animal	charity,	in	order	to	bolster	belonging	needs,	and	I	argue	that	this	behaviour	should	reduce	if	belonging	needs	have	already	been	met	through	an	assurance	of	social	inclusion.	In	study	3,	I	also	seek	to	demonstrate	that	this	behaviour	is	not	as	a	result	of	general	positive	affect	following	social	inclusion,	but	is	specific	to	an	increase	of	belonging	that	happens	following	affirmation	of	inclusion.	See	figure	4	for	a	diagram	of	the	projected	prosocial	behaviour	path	as	a	consequence	of	self-construal	and	belonging.			In	study	4,	I	then	demonstrate	the	effect	that	a	perception	of	similarity	(vs.	difference)	to	the	out-group	charity	may	have	on	donation	support,	as	evidenced	with	both	donation	intentions	and	actual	cash	donation.	Once	more,	I	demonstrate		 7	the	moderating	influence	that	both	social	inclusion	and	independent	self-construal	orientation	have	on	the	process,	as	well	as	the	mediating	role	of	connection	to	the	cause.	See	figure	5	for	a	diagram	of	the	projected	prosocial	behaviour	path	as	a	consequence	of	perception	of	similarity	(vs.	difference)	to	the	out-group	charity.	I	also	examine	(study	4)	the	influence	that	participant	variables,	such	as	gender	and	pet	ownership,	may	have	on	the	mechanisms	involved.			Taken	together,	the	results	across	the	first	four	studies	suggest	that,	following	assurances	of	belonging,	individuals	high	in	independent	self-construal	will	behave	in	a	less	inclusive	fashion,	and	will	donate	less	towards	out-group	causes.	I	argue	that	this	happens	as	a	result	of	a	reduction	in	need	for	social	connection—a	need	that	is	especially	active	in	individuals	with	high	independent	self-construal,	and	that	acts	as	a	motivating	mechanism	for	prosocial	behaviour.	Finally,	I	demonstrate,	in	a	brief	experimental	field	study	(study	5),	the	risky,	delicate	and	challenging	nature	of	using	manipulations	and	reminders	of	belonging	on	people’s	sense	of	core	belonging	status,	and	the	potential	for	such	manipulations	to	backfire.	Study	5	also	poses	the	question	as	to	whether	the	current	research	findings	extend	to	public	(vs.	private)	donation	settings.				 		 8	Chapter	2:	When	‘I’	Belong	I	Don’t	Care	about	‘You’:	The	Role	of	Self-Construal	and	Social	Inclusion	in	Prosocial	Behaviour	Directed	at	Animal	Out-group	Recipients.		2.1	Introduction	of	Dissertation	Topic		“Human	beings	are	a	part	of	the	animal	kingdom,	not	apart	from	it.	The	separation	of	"us"	and	"them"	creates	a	false	picture	and	is	responsible	for	much	suffering.	It	is	part	of	the	in-group/out-group	mentality	that	leads	to	human	oppression	of	the	weak	by	the	strong	as	in	ethnic,	religious,	political,	and	social	conflicts.”		Mark	Bekoff	(2007),	Animals	Matter.	“Humans	are	like	other	social	animals	in	that	their	hostility	to	outsiders	is	the	flipside	of	strong	friendliness	towards	their	own	group.	The	distinction	between	friends	and	enemies	is	as	central	to	human	life	as	it	is	to	the	lives	of	wolves,	meerkats	and	chimpanzees.”		Mary	Midgley	(2009),	Hobbes'	Leviathan,	part	7:	His	idea	of	war. 	As	the	above	quotes	highlight,	we	share	many	commonalities	with	other	animal	species.	According	to	Serpell,	in	his	history	of	the	human-animal	relationship	(In	the	Company	of	Animals,	1986),	for	millennia	traditional	hunter-gatherers	regarded	the	other	animals,	which	they	lived	amongst	and	hunted,	as	equals.		However,	a	fundamental	shift	of	mindset	occurred	around	12,000	years	ago,	with	the	advent	of	farming	and	domestication,	from	which	the	human-animal	relationship	has	never	fully	recovered.	As	a	result	of	this	shift,	the	previously	more	integrated	attitude		 9	human	beings	harboured	towards	the	animal	world	has	been	riven	into	one	of	“them”	and	“us”	and,	in	its	wake,	a	more	ambivalent	and	fickle	relationship	with	the	animal	world	has	emerged.		For	the	modern	day	domestic	animal,	says	Serpell,	the	human	is	its	overlord	and	master,	and	the	animals	are	reduced	to	his	servants	and	slaves.			Despite	this	shift	of	power,	many	humans	continue	to	seek	out	the	company	of	animals,	integrating	them	into	their	daily	lives	as	treasured	pets	and	even	casting	them	in	the	role	of	surrogate	family	members.		However,	the	acceptance	of	a	modern	day	domestic	animal	into	the	human	world	has	become	a	highly	complex	equation.	It	is	an	equation	in	which	associated	costs	may	be	due,	and	‘honorary	human’	status	(Midgley,	1983;	Serpell,	1986)	may	be	conferred	as	a	means	based	mechanism	for	enabling	their	exploitation	as	potential	social	resources,	only	to	be	withdrawn	when	the	need	abates.			This	research	examines	the	modern	day	human-animal	relationship,	but,	more	specifically,	it	examines	the	inclusive	behaviour	of	humans	towards	animals,	and	the	prosocial	actions	that	individuals	direct	towards	domestic	animal	charities.	The	research	takes	a	novel	approach	to	the	topic	of	human	animal	interaction	(HAI),	by	using	the	literatures	on	intergroup	social	psychology	and	prosocial	marketing	to	make	novel	predictions	regarding	when	animals	will	be	more	likely	to	be	regarded	as	more	or	less	in-group	(vs.	out-group)	members,	and	when	individuals	will	be	more	likely	to	support	animal-based	charities.	In	doing	so,	the	current	research		 10	delves	into	the	research	on	self-construal	(Aaker	&	Lee,	2001;	Block,	2005;	Cross,	Hardin,	&	Gercek-Swing,	2011;	Gardner,	Gabriel,	&	Lee,	1999;	Markus	&	Kitayama,	1991;	2010;	Singelis,	1994),	intergroup	behaviour	(Turner,	Hogg,	Oakes,	Reicher,	&	Wetherell,	1987;	Yuki	&	Takemura,	2013),	social	inclusion	and	exclusion	(Baumeister,	DeWall,	Ciarocco,	&	Twenge,	2005;	Nezlek,	Kowalski,	Leary,	Blevins,	&	Holgate,	1997;	Williams	&	Sommer,	1997;	Williams,	2009),	human	animal	interactions	and	anthropomorphism	(Amiot	&	Bastian,	2017;	Epley,	Akalis,	Waytz,	&	Cacioppo,	2008),	as	well	as	motives	for	prosocial	behaviour	(Batson,	Ahmad,	&	Stocks,	2011;	Duclos	&	Barasch,	2014;	Sturmer	&	Snyder,	2010).				In	my	research,	I	argue	that	humans	may	construct	more	inclusive	in-group	categories,	which	include	animals	within	them,	in	order	to	fulfill	social	belonging	needs.	However,	when	belonging	needs	are	satiated,	boundaries	may	tighten,	in-group	categories	may	become	more	exclusive,	and	animals	may	once	more	be	categorized	as	out-group	members.	In	addition,	I	argue	that	self-construal	orientation	will	moderate	this	behaviour	since,	according	to	Yuki	and	Takemura	(2013),	self-construal	orientation	fundamentally	impacts	the	lens	through	which	in-groups	are	perceived,	and	the	mechanisms	by	which	members	are	categorized.	Furthermore,	I	suggest	that	the	effect	will	be	mediated	by	feelings	of	connection	to	the	cause,	and	moderated	by	the	perceived	similarity	between	human	and	animal	targets.	Additionally,	I	propose	that	the	effect	is	linked	to	the	mechanisms	that	give	rise	to	anthropomorphism	directed	towards	non-humans,	as	well	as	prejudice	towards,	and	dehumanisation	of,	human	out-group	members.	Finally,	I	predict	that		 11	these	patterns	will	extend	into	prosocial	donation	support	behaviour,	directed	towards	distant	others	and,	more	specifically	in	my	research,	domestic	animal	charities.		2.2	Conceptual	Background	and	Hypothesis	Development	2.2.1	Social	identity	and	self-construal:	Our	relationships	with	others.		Tell	me	with	whom	you	associate,	and	I	will	tell	you	who	you	are.	Johann	Wolfgang	Von	Goethe	(1906),		The	Maxims	and	Reflections	of	Goethe		Man	is	by	nature	a	social	animal;	an	individual	who	is	unsocial	naturally	and	not	accidentally	is	either	beneath	our	notice	or	more	than	human.	Society	is	something	that	precedes	the	individual.	Anyone	who	either	cannot	lead	the	common	life	or	is	so	self-sufficient	as	not	to	need	to,	and	therefore	does	not	partake	of	society,	is	either	a	beast	or	a	god.	Aristotle	(1252a-1342b),		Politics		As	Goethe	and	Aristotle	highlight,	humans	are	not	only	intensely	social	creatures	that	need	to	exist	within	a	social	framework	in	order	to	function	properly,	but	are	also	more	or	less	defined	by	the	social	frameworks	they	find	themselves	inhabiting.		A	number	of	theories	have	been	put	forward	to	explain	the	mechanisms	governing	how	we	understand,	and	define,	our	sense	of	self,	as	well	as	how	our	sense	of	self	contributes	to	our	social	identity	and	impacts	the	social	connections	in	our	lives.				 12	Social	identity.	Social	identity	theory,	(SIT;	Tajfel	&	Turner,	1979),	defines	social	identity	as	a	person's	knowledge	that	he	or	she	belongs	to	one	or	more	social	categories	or	groups	(Hogg	&	Abrams,	1988).	A	social	group	is	a	set	of	individuals	who	hold	a	common	social	identification,	or	view	themselves	as	members	of	the	same	social	category.	Through	a	social	comparison	process,	persons	who	are	similar	to	the	self	are	categorized	with	the	self,	and	are	labeled	as	the	in-group,	whereas	persons	who	differ	from	the	self	are	categorized	as	the	out-group.	There	are	two	important	processes	involved	in	social	identity	formation;	namely,	self-categorization	and	social	comparison,	each	of	which	produce	different	consequences	(Hogg	&	Abrams,	1988).	The	consequence	of	self-categorization	is	an	accentuation	of	the	perceived	similarities	between	the	self	and	other	in-group	members,	and	an	accentuation	of	the	perceived	differences	between	the	self	and	out-group	members.	The	consequence	of	the	social	comparison	process	is	the	selective	application	of	the	accentuation	effect,	primarily	to	those	dimensions	that	will	result	in	self-enhancing	outcomes	for	the	self.		Most	importantly,	in	social	identity	theory,	one's	self-esteem	is	enhanced	by	evaluating	the	in-group	and	the	out-group	on	dimensions	that	lead	the	in-group	to	be	judged	positively,	and	the	out-group	to	be	judged	negatively	(Stets	&	Burke,	2000).			According	to	Hogg	and	Abrams	(1988),	the	social	categories	in	which	individuals	place	themselves	are	parts	of	a	structured	society,	and	exist	only	in	relation	to	other	contrasting	categories	(i.e.	male	vs.	female;	human	vs.	animal),	with	each	category		 13	possessing	more	or	less	power,	prestige,	status,	and	so	on.	Individuals	derive	their	identity,	or	sense	of	self,	largely	from	the	social	categories	to	which	they	belong,	with	each	person	maintaining	memberships	in	a	unique	combination	of	different	social	categories.	In	such	a	way,	the	make	up	each	person's	self-concept	is	unique.	One	of	the	key	assumptions	within	social	identity	theory	is	that	individuals	are	intrinsically	motivated	to	achieve	positive	distinctiveness,	through	the	striving	for	a	positive	self-concept.	However,	at	the	same	time,	people	are	also	motivated	to	strive	for	a	positive	social	identity,	through	membership	with	groups	that	are	positively	distinct	from	relevant	out-groups.			Much	of	social	identity	theory	deals	with	intergroup	relations.	Intergroup	relations	may	be	understood	as	the	manner	by	which	individuals	come	to	see	themselves	as	members	of	one	group,	or	category	(the	in-group),	in	comparison	with	another	group	or	category	(the	out-group).		According	to	social	identity	theory,	one	consequence	of	this	categorization	may	be	intergroup	conflict	and	ethnocentrism	(Turner,	Hogg,	Oakes,	Reicher,	&	Wetherell,	1987).	Furthermore,	evidence	has	shown	that	stereotyped	perceptions	of	in-group	members	and	out-group	members	may	be	enhanced	and	made	more	homogeneous	by	identification	with	the	in-group	(Haslam,	Oakes,	McGarty,	Turner,	Reynolds,	&	Eggins,	1996).	In	sum,	based	on	the	extant	research,	we	can	conclude	that	having	a	particular	social	identity	means	feeling	a	part	of	a	certain	group,	being	like	others	in	the	group,	and	seeing	things	from	the	group's	perspective.	All	of	this	entails,	according	to	Turner	and	colleagues		 14	(1987),	“a	shift	towards	the	perception	of	self	as	an	interchangeable	exemplar	of	some	social	category,	and	away	from	the	perception	of	self	as	a	unique	person.”			Two	other	related	theories	have	more	recently	emerged	from	social	identity	theory;	self-categorization	theory	(SCT;	Turner,	Hogg,	Oakes,	Reicher,	&	Wetherell,	1987;	Turner,	Oakes,	Haslam,	&	McGarty,	1994),	and	optimal	distinctiveness	theory	(ODT;	Brewer,	1991;	Leonardelli,	Pickett,	&	Brewer,	2010).	According	to	self-categorization	theory,	rather	than	seeing	interpersonal	and	intergroup	dynamics	as	being	at	either	end	of	a	bipolar	spectrum,	identity	should	be	categorized	and	seen	as	operating	at	different	levels	of	inclusiveness.	These	levels	of	inclusiveness	vary	from	the	self	as	a	human	being,	to	the	self	as	a	part	of	a	group,	and/or	compared	to	other	groups.	Amiot	and	Bastian	(2017)	argue	we	may	usefully	approach	our	understanding	of	human-animal	relations	from	this	type	of	intergroup	perspective.		According	to	Brewer	(1991),	and	Brewer	and	Gardner	(1996),	seen	this	way,	the	self	may	be	viewed	as	existing	within	a	framework	of	concentric	circles,	with	personal	identity	at	the	core,	and	various	social	identities	surrounding	the	self	(see	figure	1).		Optimal	distinctiveness	theory,	on	the	other	hand,	purports	to	fill	a	gap	that	has	been	missed	by	both	social	identity	theory	and	self-categorization	theory.	This	gap	is	centered	around	what	drives	identification	with	in-groups,	particularly	in	the	chronic	long-term	(Brewer,	1999).	Optimal	distinctiveness	theory	suggests	that	individuals	are	faced	with	two	basic	competing	human	needs:	to	be	part	of	social	group,	but	also	to	satisfy	one’s	own	individual	needs.	Accordingly,	a	tension	exists	in		 15	which	the	two	opposing	identity	needs,	or	forces,	compete.	On	one	side	is	individual	uniqueness	(differentiation/distinctiveness),	and	on	the	other	side	is	a	need	for	validation	and	similarity	that	can	only	be	found	through	in-group	homogeneity	(inclusion/assimilation).	Individuals	are	compelled	to	form	an	identity	that	satisfies	both	needs,	through	the	striking	of	an	‘optimal’	balance.	This	‘optimal	balance’	resides	somewhere	between	the	desire	for	distinctiveness	(i.e.	feeling	unique	or	different	from	others),	and	the	desire	for	assimilation	or	belonging	(i.e.	feeling	similar	to	others)	according	to	Brewer	(1991;	see	figure	2).			Self-construal.	An	alternate	way	of	thinking	about	the	self,	in	relation	to	others,	is	from	the	perspective	of	self-construal.	As	previously	discussed,	self-construal	is	best	defined	as	what	people	“believe	about	the	relationship	between	the	self	and	others,	and	especially,	the	degree	to	which	they	see	themselves	as	separate	from	others,	or	as	connected	with	others”;	with	a	crucial	distinction	hinging	on	the	role	assigned	to	‘other,’	in	terms	of	the	self-definition	(Markus	&	Kitayama,	1991,	p.	26).	For	those	with	a	more	independent	construal	of	self,	‘others’	are	less	implicated	in	one’s	self-definition,	or	self-identity,	and	there	is	a	strong	sense	of	individuality.	This	includes	a	feeling	of	oneself	as	an	agent	in	control,	standing	out	from	the	crowd	and	being	“true	to	one’s	own	internal	structures	of	preference,	rights,	convictions	and	goals”	(Markus	&	Kitayama,	1994;	see	figure	3).		Individuals	with	highly	independent	self-construal	are	therefore	considered	to	be	more	likely	to	behave	according	to	their	own	preferences,	and	for	their	personal	pleasure.				 16	On	the	other	hand,	individuals	with	a	strongly	interdependent	self	are	thought	to	define	their	own	identity	through	their	relationship	with	other	people	in	the	group.	As	a	result,	individuals	with	a	highly	developed	interdependent	self-construal	see	the	connectedness	between	people	as	fundamentally	important	and,	as	a	consequence	of	this,	are	more	likely	to	behave	in	accordance	with	“anticipated	expectations	of	others	and	social	norms”	(Markus	&	Kitayama	1991,	p.	228).	They	value	harmonious	relations,	and	their	actions	are	focused	on	fostering	relatedness	and	connection	to	others	(Chuang,	Xie,	&	Liu,	2016;	Cross,	Hardin,	&	Gercek-Swing,	2011;	Markus	&	Kitayama,	1991;	Singelis,	1994),	as	well	as	working	to	maintain	harmony	(Bagozzi,	Verbeke,	&	Gavino	Jr.,	2003).	For	those	with	a	relatively	more	interdependent	self-construal,	‘others’	are	often	included	within	the	boundaries	of	the	self,	since	the	contexts	of	‘other’	relationships	are	seen	as	defining	features	of	the	self.	In	this	situation	the	individual	self	may	be	seen	as	a	“fraction”	(Lebra,	1976)	of	a	larger	social	unit.	As	a	result,	self-esteem,	for	those	high	in	interdependent	self-construal,	is	largely	born	of	self-restraint	and	accommodation.	This	may	be	contrasted	with	those	individuals	that	are	higher	in	independent	self-construal,	who	are	more	likely	to	derive	self-esteem	from	an	individual	pride	of	self-achievement	and	distinction	(Cross,	Bacon,	&	Morris,	2000).				Self-construal	research	has	been	approached	from	both	an	individual	difference	perspective	(Aaker	&	Lee	2001;	Block	2005;	Singelis	1994)	and	a	cultural	perspective	(Cross	et	al.,	2011;	Gardner,	Gabriel,	&	Lee,	1999;	Graham,	Waytz,	Meindl,	Iyer,	&	Young,	2017;	Markus	&	Kitayama	1991).	Looking	at	self-construal	in		 17	terms	of	individual	differences,	Singelis	(1994)	argues	that	traits	of	both	independent	and	interdependent	self-construal	may	co-exist	within	one	individual.	From	a	cultural	perspective,	on	the	other	hand,	the	independent	(or	idiocentric)	self	is	assumed	to	be	stronger	in	western	cultures,	and	the	interdependent	(or	allocentric)	self	more	prevalent	in	non-western	cultures	(Cross,	1995;	Cross	et	al.,	2011;	Markus	&	Kitayama,	1991,	2010).			More	recent	research	has	embraced	both	social	identity	theory	and	self-categorization	theory	alongside	self-construal,	to	further	explore	how	different	culturally	embedded	self-construals	might	interact,	in	terms	of	group	processes	(Yuki	&	Takemura,	2013).		Yuki	and	Takemura	propose	a	socio-ecological	approach	that	focuses	on	the	differences	in	group	processes	across	cultures,	and	specifically	between	collectivist	cultures,	containing	more	individuals	with	higher	levels	of	interdependent	self-construal,	and	individualist	cultures,	comprised	of	individuals	with	higher	levels	of	independent	self-construal.	Yuki	and	Takemura	argue	that	there	exist	two	distinct,	and	differing,	concepts	of	how	in-group	membership	is	understood,	which	are	dependent	on	self-construal	orientation.	Collectivist	non-Western	cultures,	according	to	the	socio-ecological	perspective,	are	more	likely	to	view	their	in-group	as	a	non-permeable	long-term	structure,	containing	inter-connective	networks	founded	on	shared	connections	that	must	be	kept	in	harmony.		On	the	other	hand,	individualist	Western	cultures	are	likely	to	see	in-groups	as	more	dynamic	in	nature,	based	on	shared	categories	within	the	group,	which	are	paramount	to	social	identity.	In	these	latter	in-groups	members	are	interchangeable,		 18	and	differ	in	their	prototype-based	position	in	the	group.		Societies	dominated	by	individuals	with	high	independent	self-construal,	tend	to	have	higher	interpersonal	and	intergroup	mobility,	also	termed	‘relational	mobility’	(Yuki	&	Takemura,	2013).	Relational	mobility	is	defined	as	the	degree	to	which	there	is	an	availability	of	social	identity	options,	such	as	opportunities	to	acquire	new,	maintain	current,	or	sever	old	relationships	(Yuki,	Maddux,	&	Masuda,	2007;	Yuki	&	Shug,	2012).	Where	there	is	low	relational	mobility	it	is	critical	for	individuals	to	work	hard	to	maintain	long-standing	relationships	and	groups.	This	is	because	with	low	relational	mobility,	group	memberships	are	long	standing,	ascribed	and	predetermined,	and	it	is	hard	for	individuals	to	leave	their	groups,	even	if	they	are	found	to	be	unsatisfactory.	Where	there	is	high	relational	mobility,	on	the	other	hand,	there	exists	an	abundance	of	opportunities	to	meet	strangers	and	create	new	relationships	(Yuki	&	Takemura,	2013).				High	relational	mobility	affords	those	experiencing	it,	typically	individuals	with	a	more	independent	self-construal,	the	powerful	freedom	to	choose	which	groups	they	belong	to	at	any	given	moment;	allowing	them	to	select	according	to	personal	goals	and	social	categories.	On	the	other	hand,	high	relational	mobility	also	brings	a	“Sword	of	Damocles”1	situation,	in	which	individuals	must	constantly	monitor	intergroup	status	differences,	in	order	to	associate	with	groups	of	the	highest	status,																																																									1	1	The	Sword	of	Damocles	originates	in	a	text	of	the	orator	Cicero,	and	refers	to	a	great	sword	that	hung	over	the	throne	of	King	Dionysius,	to	denote	the	belief	that	with	great	power	and	fortune	also	comes	great	danger	and	fear.			 19	and	best	category	fit.	Moreover,	individuals	must	also	take	care	to	monitor	their	own	personal	belonging	status,	and	take	action	to	maintain	an	optimal	level	of	inclusion	(Pickett	&	Gardner,	2005).		Conversely,	societies	dominated	by	low	relational	mobility,	more	frequently	seen	in	collectivist	societies	and	primarily	populated	by	individuals	with	high	interdependent	self-construal,	experience	little	opportunity	to	change	group	membership.	Yamagishi	and	Kosugi	(1999)	suggest	we	consider	this	type	of	situation	as	a	“collectivist’s	collectivism”	whereas,	according	to	socio-ecological	perspective,	North	America,	filled	primarily	with	individuals	high	in	independent	self-construal,	may	be	better	termed	an	“individualist’s	collectivism”	(Yuki	&	Takemura,	2013).	The	result	of	this	difference,	argue	Yuki	and	Takemura,	is	that	people	high	in	interdependent	self-construal,	or	those	living	in	collectivist	cultures,	not	only	have	a	strong	motivation	to	maintain	harmonious	intragroup	relations,	but	also	consider	their	memberships	within	their	in-groups	to	be	chronically	guaranteed.			These	findings	are	supported	by	the	research	of	Mandel	(2003),	demonstrating	that	when	people	were	primed	with	interdependent	self-construal	they	were	able	to	identify	more	individuals	on	whom	they	could	depend	than	people	primed	with	independent	self-construal.	According	to	Mandel	(2003)	this	creates	a	cushion	effect,	which	in	turn	encourages	people	primed	with	interdependent	self-construal	to	be	more	likely	to	take	on	certain	types	of	risks.		For	individuals	with	a	higher		 20	independent	self-construal,	however,	group	membership	is	more	precarious,	due	to	the	dynamic	nature	of	in-group	formation,	and	the	group’s	more	porous	and	fluid	boundaries.			Yuki	and	Takemura’s	argument	is	supported	by	evidence,	suggesting	that	people	from	individualist	cultures,	and	people	with	a	more	independent	self-construal,	do	indeed	have	distinctive	group	mechanics,	compared	to	people	from	more	collectivist	cultures	or	those	with	a	more	interdependent	self-construal	(Buchan,	Croson,	&	Dawes,	2002;	Buchan,	Johnson,	&	Croson,	2006;	Yamagishi,	Makimura,	Foddy,	Matsuda,	Kiyonari,	&	Platow,	2005).	One	example	of	this	is	that	people	with	a	more	independent	self-construal	often	show	more	in-group	bias,	in	terms	of	trust,	compared	to	people	with	a	more	interdependent	self-construal.		While,	at	first	glance,	this	finding	may	appear	counterintuitive,	it	can	be	better	understood	if	we	accept	that	there	are	two	bases	for	trust	of	strangers	within	in-groups.	The	first	is	based	on	a	social	identity	theory	approach	of	shared	social	categories	leading	to	equitable	resource	distribution	(Brewer	&	Kramer,	1986;	Foddy,	Platow,	&	Yamagishi,	2009);	the	second	is	based	on	a	more	personalized	trust,	which	stems	from	shared	connections	and	interpersonal	ties	(Coleman,	1990).		Evidence	(Yuki,	Maddox,	Brewer,	&	Takimura,	2005)	supports	the	proposal	that	collectivists	(Japan)	may	be	more	likely	to	show	trust	for	an	unknown	out-group	member	who	has	shared	connections,	in	the	form	of	indirect	interpersonal	ties,	than	individualists	(USA)	will	be.			 21	Finally,	one	last	area	of	interest,	regarding	self-construal	and	intergroup	behaviour,	relates	to	the	proclivity	for	demonstrating	in-group	bias.	Research	has	shown	that	evaluative	in-group	bias–the	act	of	favouring	members	of	one’s	own	in-group	over	out-group	members–may	be	greater	in	those	with	a	more	individualist	orientation,	if	the	targets	of	evaluation	are	category	based	and	therefore	evaluation	is	related	to	shared	categories	(Bond	&	Hewstone,	1988;	Heine	&	Lehman,	1997;	Rose,	1985;	Snibbe,	Kitayama,	Markus,	&	Suzuki,	2003).	Conversely,	if	groups	are	instead	defined	by	relational	shared	connections,	then	individuals	higher	in	interdependent	self-construal	may	be	found	to	be	more	likely	engage	in	evaluative	in-group	bias	(Endo,	Heine,	&	Lehman,	2000).			In	sum,	research	evidence	supports	the	claim	by	Yuki	and	Takemura	(2013)	that	social	identity	differences	between	independent	and	interdependent	self-construals	primarily	exist	in	terms	of	the	kinds,	rather	than	the	levels,	of	group	orientations.	As	a	result	of	the	differences,	individuals	that	are	more	individualist,	or	have	a	higher	independent	self-construal,	will	be	more	driven	to	define	social	groups	in	terms	of	an	inter-group	comparison.	In	this	situation,	the	focus	will	be	on	the	shared	categories	that	group	members	enjoy,	with	the	self	viewed	as	a	prototypical	group	member,	and	with	individuals	motivated	to	achieve	higher	intergroup	status.	This	may	be	contrasted	with	individuals	that	are	more	collectivist,	or	have	a	higher	interdependent	self-construal,	who	may	be	more	likely	to	utilize	intra-group	comparison,	in	which	groups	are	defined	in	terms	of	shared	connections	with	bounded	interpersonal	networks,	and	members	are	motivated	to	maintain		 22	harmonious	reciprocal	relationships	between	members	of	the	same	group.	Simply	put,	according	to	this	conceptualization,	the	sense	of	self,	seen	as	typical	of	independent	self-construal,	is	not	a	form	of	self	that	is	alienated	or	in	social	isolation	from	society,	as	has	been	previously	argued	by	some	(Cushman,	1990),	but	is	rather	a	form	of	social	orientation	in	which	social	identity	is	more	fluid	and	category	driven.				As	a	consequence,	I	chose	with	the	current	research	to	build	off	two	perspectives	regarding	the	self.	On	the	one	hand,	I	work	from	an	optimal	distinctiveness	theory	approach,	which	argues	that	the	self	is	a	combination	of	personal	identity	and	social	identity,	existing	in	a	state	of	tension,	with	a	level	of	fluidity	that	allows	for	some	movement	between	both.	In	addition	to	this,	however,	my	research	also	embraces	specific	aspects	of	Yuki	and	Takemura’s	(2013)	socio-ecological	perspective,	which	argues	that	distinct	differences	exist,	not	just	in	how	the	self	is	constructed,	but	also	in	how	the	self	interacts	with	its	closest	groups,	in	terms	of	social	connection.	Rather	than	seeing	interdependent	self-construal	from	a	narrow	perspective,	where	the	self	is	defined	as	a	fraction	of	a	greater	social	whole,	and	independent	self-construal	from	an	equally	narrow	perspective,	in	which	an	individual	exists	in	isolation	from	the	group,	I	follow	in	the	tradition	of	the	socio-ecological	perspective.	As	a	result,	I	argue	that	individuals	with	high	independent	self-construal,	and	individuals	with	high	interdependent	self-construal,	are	likely	to	be	equally	concerned	with	their	social	identity.	However,	I	argue	that	because	social	identity	is	experienced	in	a	fundamentally	different	way	for	individuals	with	high	independent	self-construal	(in		 23	which	social	identity	is	precarious	but	fluid),	compared	to	individuals	with	high	interdependent	self-construal	(in	which	social	identity	is	assured	but	rigid),	obtaining	an	optimal	social	identity	status	will	require	different	maintenance	strategies	for	these	people.	Moreover,	I	argue	that	even	more	social	monitoring	will	be	required	by	individuals	with	high	independent	self-construal,	in	order	to	maintain	an	optimal	balance	of	self	and	social	identity.		Additionally,	I	propose,	that	one	of	the	reasons	why	people	with	high	independent	self-construal	have	previously	been	found	to	express	a	more	inclusive	and	benevolent	attitude	towards	out-group	targets	(Duclos	&	Barasch,	2014;	Leeuwen	&	Tauber,	2012;	Sturmer	&	Snyder,	2010)	is	because	people	with	a	high	independent	self-construal	are	more	open	to	categorizing	out-group	members	as	in-group,	and	including	them	within	a	self	formed	in-group.	I	argue	that	this	is	as	a	result	of	the	more	flexible,	and	porous,	nature	of	their	in-group	boundaries;	along	with	the	propensity	of	individuals	that	are	high	in	independent	self-construal	to	form	in-groups	primarily	based	on	shared	categories,	rather	than	based	on	the	shared	connections	that	are	more	typical	of	interdependent	self-construal	grouping	criteria.		As	a	result	of	this	theorizing,	I	have	generated	the	following	hypothesis:	Hypothesis	1:		Out-group	members	will	form	a	higher	percentage	of	a	self-formed	in-group,	the	higher	a	person	measures	on	independent	self-construal.		Additionally,	I	predict	that,	since	individuals	with	a	high	independent	self-construal		 24	experience	a	lower	baseline	state	of	social	inclusion	than	do	individuals	with	a	high	interdependent	self-construal,	they	will	therefore	also	be	more	vigilant	for	cues	of	potential	for	social	connection	coming	from	out-group	members	that	they	may	come	in	contact	with.	This	prediction	directly	builds	from	the	work	of	Pickett,	Gardner,	and	Knowles	(2004),	who	show	that	the	need	to	belong	enhances	sensitivity	to	social	cues.	I	therefore	propose	a	second	hypothesis:	Hypothesis	2:		The	higher	an	individual’s	measure	in	independent	self-construal,	the	greater	the	number	of	items	that	have	high	social	connection	potential	will	be	put	in	a	self-formed	in-group.			I	do	not	expect	to	see	the	same	pattern	in	individuals	with	high	interdependent	self-construal.		2.2.2	Exclusion	and	belonging:	Impacts	of	inclusion	on	sense	of	self.		The	Loneliness	whose	worst	alarm,	Is	lest	itself	should	see—And	perish	from	before	itself,	For	just	a	scrutiny—	The	Horror	not	to	be	surveyed—But	skirted	in	the	Dark,	with	Consciousness	suspended,	And	Being	under	Lock.																																																																																																																															Emily	Dickenson	(c.1860)	The	Loneliness	One	Dare	Not	Sound			After	all,	what	are	any	of	us	after	but	the	conviction	of	belonging?		Wallace	Stegner	(c.1988)	On	Teaching	and	Writing	Fiction				 25	If	the	in-groups,	and	others	with	whom	we	interact,	are	such	an	important	part	of	our	self-construal,	then	it	stands	to	reason	that	our	social	inclusion	status,	and	a	sense	that	we	belong,	will	be	vital	to	personal	wellbeing.	Past	research	has	noted	that	humans	are	intensely	social	creatures,	with	a	deep	urge	to	seek	company	and	to	meaningfully	interact	with	fellow	humans	on	a	regular	basis	(see	Baumeister	&	Leary,	1995,	for	a	comprehensive	review).	The	negative	consequences	from	failed	inclusion	can	be	powerful	and	immediate	(Baumeister,	DeWall,	Ciarocco,	&	Twenge,	2005;	Nezlek,	Kowalski,	Leary,	Blevins,	&	Holgate,	1997;	Williams	&	Sommer,	1997;	Williams,	2009).	People	report	not	only	lower	levels	of	belonging	following	ostracism,	but	also	lower	control,	self-esteem,	and	meaningful	existence	(Zadro,	Williams,	&	Richardson,	2004),	even	when	the	source	of	ostracism	is	a	machine	(Williams,	Forgas,	&	Von	Hippel,	2005).		It	should	therefore	hardly	be	surprising	to	find	that	most	people	devote	significant	energies	into	maintaining	social	connections	throughout	their	lives,	and	managing	reasonable	levels	of	what	we	might	term	good	‘social	health’	(Leary,	Tambor,	Terdal,	&	Downs,	1995;	Pickett	&	Gardner,	2005).			Research	into	responses	to	both	in-group	and	out-group	members,	following	threats	to	social	inclusion,	have	generated	mixed	results,	however.	Twenge,	Baumeister,	Tice,	and	Stucke	(2001)	have	found	that	social	exclusion	may	lead	to	an	increase	in	violence	and	aggression.	Twenge,	Baumeister,	DeWall,	Ciarocco,	and	Bartels	(2007)	have	put	forward	evidence	to	show	that	it	reduced	cooperation	and	prosocial	behaviour.		Cuadrado,	Tabernero,	and	Steinel	(2015)	have	noted,	however,	that		 26	excluded	individuals	may	behave	more	prosocially	than	included	individuals,	if	they	have	the	hope	of	reconnecting.	This	echoes	the	findings	of	Maner,	DeWall,	Baumeister,	and	Schaller	(2007),	which	found	that	social	interactions	following	exclusion	crucially	depended	on	a	number	of	factors,	one	of	which	being	whether	there	existed	the	potential	for	social	connection	with	novel	acquaintances.	The	notion	that	individuals	will	respond	to	exclusion	by	engaging	in	positive	behaviours	with	others	has	been	referred	to	as	the	reconnection	hypothesis	(Baumeister	&	Leary,	1995;	Maner,	DeWall,	Baumeister,	&	Schaller,	2007).	Researchers	have	found	evidence	for	the	reconnection	hypothesis	in	various	areas,	including	workplace	interactions	(Bernstein,	Sacco,	Brown,	Young,	&	Claypool,	2010),	ingratiation	behaviour,	(Romero-Canyas,	Downey,	Reddy,	Rodriguez,	Cavannaugh,	&	Pelayo,	2010),	and	discrimination	(Smart	Richman	&	Leary,	2009).			Sometimes	it	is	not	possible,	or	desirable,	to	connect	with	other	people,	however.		Epley,	Akalis,	Waytz,	and	Cacioppo	(2008)	found	that	both	acute	exclusion,	as	well	as	chronic	isolation,	motivated	people	to	look	for,	and	if	necessary	create,	social	connections	with	non-humans.	Building	on	this,	Epley	and	colleagues	have	offered	evidence	that	anthropomorphism	may	be	driven	by	sociality	motivation.	Anthropomorphism	is	the	attribution	of	human	traits,	emotions,	intentions,	behaviours,	or	looks	to	non-human	entities,	and	is	considered	to	be	an	innate	tendency	of	human	psychology,	being	shown	from	a	very	young	age	(see	Epley,	Akalis,	Waytz,	&	Cacioppo,	2008,	for	a	more	detailed	description).	The	finding	that	anthropomorphism	may	be	driven	by	sociality	motivation	follows	on	from	previous		 27	research	in	the	area	of	anthropomorphism,	and	has	been	recently	corroborated	by	the	research	of	Powers,	Worsham,	Freeman,	Wheatley,	and	Heatherton	(2014).	Powers	and	colleagues	(2014)	have	proposed	that	over-attributing	animacy	to	non-humans	is	an	adaptive	strategy—one	that	allows	isolated	individuals	to	cast	a	wide	net	when	looking	for	social	connection,	and	to	maximize	opportunities	to	connect.	Work	in	a	separate,	but	not	unconnected,	stream,	has	come	to	similar	conclusions.	Gardner,	Pickett,	and	Knowles	(2005)	have	proposed	that	the	mechanisms	involved	in	belonging	regulation,	may	be	understood	as	similar	to	those	regulating	hunger,	with	social	snacking	(e.g.,	re-reading	letters/	looking	at	photos	of	friends)	and	social	shielding	(e.g.,	use	of	parasocials	such	as	plants,	pets,	etc.)	taking	place,	when	no	social	interaction	with	human	contacts	appears	to	be	immediately	possible.			Past	research	has	also	explored	the	interactions	between	exclusion	and	self-construal.	Ren,	Wesselmann,	and	Williams	(2013)	found	that	an	interdependent	self-construal	orientation	facilitated	participants’	recovery	from	social	exclusion,	even	if	it	didn't	seem	to	impact	the	feeling	of	initial	pain.	Pfundmair,	Aydin,	Du,	Yeung,	Frey,	and	Graupmann	(2015)	have	suggested	that	an	interdependent	self-construal	orientation	may	help	buffer	a	person	from	the	negative	impacts	of	exclusion,	in	a	way	that	a	high	independent	self-construal	orientation	may	not.	This	buffering	hypothesis	is	also	supported	by	Uskul	and	Over	(2014),	who	have	provided	evidence	that	the	strategic	importance	of	others,	be	they	close	or	distant,	may	mediate	responses	to	social	exclusion,	if	it	does	occur.				 28	Pfundmair,	Graupmann,	Frey,	and	Aydin	(2015)	also	support	these	findings	with	further	research,	demonstrating	that	participants	with	a	collectivist	orientation	show	little	difference	on	a	range	of	measures	(antisocial	intention,	prosocial	intention,	avoiding	intention,	affect),	following	either	social	inclusion,	or	social	exclusion;	when	compared	to	participants	with	a	more	individualist	orientation.	In	their	studies,	Pfundmair	and	colleagues	describe	more	individualist	participants	as	reporting	higher	prosocial	intention	and	affect,	following	inclusion;	but	reporting	more	antisocial	intention	and	avoiding	intention,	following	exclusion.	Pfundmair,	Graupmann,	Du,	Frey,	and	Aydin	(2014)	have	also	previously	found	that	participants	with	a	more	individualist	orientation	experience	re-inclusion	differently	to	continued	inclusion;	with	more	residual	exclusion	feelings	reported,	along	with	reduced	fulfillment	of	basic	psychological	needs.	Collectivists,	in	this	research,	did	not	differentiate	between	re-inclusion	and	continued	inclusion.	This	finding	adds	further	support	for	the	notion	that	a	collectivist	orientation	may	buffer	against	the	negative	impacts	of	exclusion,	and	a	more	individualist	orientation	may	experience	increased	susceptibility	and	threat	from	exclusion,	which	is	harder	to	overcome,	in	terms	of	negative	impacts	to	basic	need	fulfillment.		A	lesser	amount	of	research	has	looked	at	the	positive,	or	negative,	impacts	of	inclusion,	or	at	the	impacts	of	inclusion	following	exclusion.	Results	from	the	studies	that	have	been	carried	out	have	delivered	mixed	findings.	On	the	one	side,	Brewer	(1991)	and	DeWall,	Baumeister,	and	Vohs	(2008)	have	put	forward	evidence	to	support	the	view	that	people	may	experience	belonging	satiation;	and	that	those		 29	who	feel	very	socially	connected	will	be	less	motivated	to	connect	with	others,	or	gain	social	acceptance	elsewhere.	Waytz	and	Epley	(2012)	have	similarly	demonstrated	that	assurances	of	social	connection	can	result	in	the	dehumanization	of	others,	as	well	as	an	increase	of	perceived	distance,	between	oneself	and	distant	others.		Conversely,	a	number	of	studies	have	found	affiliation	motivation	to	be	a	positive	predictor	of	prosocial	behaviour	(Baumeister	&	Leary,	1995;	DeWall	&	Richman,	2011;	Pavey,	Greitmeyer,	&	Sparks,	2011;	Zaskodna,	Simek,	&	Mlcak,	2013).	Specifically,	Ng	(2015)	has	shown	that	social	inclusion	may	help	recover	depleted	self-control,	and	Thompson	(2015)	found	that	priming	participants	with	relevant	social	affiliation	had	an	impact	on	self-perceived	morality.		Thompson	also	found	that	a	prime	of	social	affiliation	affected	allegiances	to	a	person’s	own	social	group	and	the	community,	with	some	individuals	more	likely	to	report	greater	concern	for	others	wellbeing	and	a	wider	moral	concern	for	others	beyond	their	own	group.	Mikulincer	and	Shaver	(2001)	found	a	sense	of	being	loved	made	participants	less	negative	towards	out-groups,	and	Cuadrado,	Tabernero,	and	Steinel	(2015)	similarly	found	that	included	individuals	reported	higher	levels	of	positive	affect	and	trust	following	inclusion,	as	well	as	exhibiting	more	prosocial	behaviour.		Building	on	from	this	body	of	past	work,	I	have	opted	in	my	research,	rather	than	to	concentrate	solely	on	social	exclusion–a	negative	and	unpleasant	experience	by	all	accounts–to	instead	focus	my	investigation	primarily	on	its	lesser	researched	counter,	social	inclusion	and	sense	of	belongingness.	Specifically,	I	wish	to		 30	investigate	whether	social	inclusion	status	might	impact	prosocial	behaviour	in	the	form	of	less	benevolent	and	receptive	intentions	expressed	towards	out-group	members.	Furthermore,	I	wish	to	investigate	whether	the	impact	of	social	inclusion	status	on	prosocial	behaviour	will	vary	as	a	function	of	self-construal	orientation.			Following	previous	findings	(Duclos	&	Barasch,	2014;	Ren,	Wesselmann,	&	Williams,	2013)	demonstrating	that	interdependent	self-construal	appears	to	offer	some	protection	against	exclusion	threats,	I	anticipate	that	people	with	a	high	interdependent	self-construal	orientation	will	react	less,	to	both	inclusion	and	exclusion	manipulations,	when	compared	to	people	with	an	independent	self-construal	orientation.	I	propose	that	this	is	because	they	are	essentially	buffered	against	these	conditions	by	their	social	belonging	status,	which,	as	previously	detailed	by	Yamagishi	and	Kosugi	(1999),	is	assured	as	a	result	of	the	lack	of	potential	for	intergroup	social	mobility.	On	the	other	hand,	I	predict	that	people	with	an	independent	self-construal	orientation	will	respond	to	an	inclusion	manipulation	by	showing	less	interest	in	interacting	with	non-humans,	and	therefore	including	less	animals	in	their	self	formed	in-group.	Moreover,	I	anticipate	that	people	with	an	independent	self-construal	orientation	will	respond	to	an	exclusion	manipulation	by	showing	more	interest	in	interacting	with	others	that	are	not	associated	with	the	act	of	exclusion–non-humans	in	this	instance–and	therefore	including	more	animals	in	a	self	formed	in-group.				 31	I	propose	that	this	is,	in	part,	because	people	with	high	independent	self-construal	are	likely	to	have	a	more	flexible	group	perspective,	which	allows	for	increased	group	mobility	and	more	porous	group	boundaries	and,	furthermore,	are	more	likely	to	define	in-group	identity	by	category	than	by	interconnected	ties	(Yamagishi	&	Kosugi,	1999;	Yuki	&	Takemura,	2013).	Drawing	on	these	insights,	I	propose	that	people	with	high	independent	self-construal	will	be	able	to	re-categorize	others	and	themselves	fairly	easily	into	groups–as	opportunity	or	need	presents–and	will	seek	to	do	so	in	order	to	maintain	belonging	needs.	One	of	the	key	elements	specific	to	my	research	is	that	this	ability–to	flexibly	re-categorize	people	according	to	need–which	may	be	seen	in	individuals	with	high	independent	self-construal,	should	likewise	facilitate	categorizing	animals	into	sources	of	social	potential,	and	awarding	animal	parasocials	an	honorary	human	status	more	easily	(see	Midgley,	1983;	Serpell,	1986).			There	are	solid	foundations	to	the	notion	that	people	will	be	willing	to	award	human	attributes	to	non-human	animals	when	it	suits	their	needs.	Epley,	Waytz,	Akalis,	and	Cacioppo	(2008)	have	demonstrated	that	people	will	attribute	minds	to	animals	when	in	need	of	companionship,	and	Bastian,	Loughnan,	Haslam,	and	Radke	(2012)	have	detailed,	conversely,	how	people	will	deny	mental	capacities	to	animals	if	they	wish	to	eat	them,	in	order	to	enable	moral	disengagement.		Put	simply,	humans	are	pretty	adept	at	re-categorizing	non-human	others	as	honorary	humans,	when	it	suits	their	needs;	and	are	prepared	to	dehumanize	humans,	and	deny	minds	to	others,	to	justify	ill	treatment	(Kozak,	March,	&	Wegner,	2006)	or	lack	of	assistance		 32	(Cuddy,	Rock	&	Norton,	2007).		Consequently,	I	propose	that,	as	a	result	of	the	manner	in	which	individuals	with	high	independent	self-construal	construct	their	sense	of	social	self–and	therefore	function	in	group	settings–individuals	with	high	independent	self-construal	may	be	viewed	as	more	‘socially	opportunistic’	than	individuals	with	high	interdependent	self-construal.			All	individuals	are	considered	to	monitor	social	cues	in	order	to	maintain	optimal	belongingness,	and	this	behaviour	is	considered	to	increase	in	the	face	of	belonging	threats	(Pickett	&	Gardner,	2005).	I	argue,	however,	that,	due	to	the	precariousness	of	their	social	inclusion	status,	individuals	high	in	independent	self-construal	should	be	especially	vigilant	for	cues	of	social	potential.		As	a	result	of	this,	I	propose	that	individuals	with	high	independent	self-construal	will	also	be	more	sensitive,	and	adept,	at	monitoring	for	cues	of	social	potential	in	non-human	others.	Furthermore,	I	propose	that	when	it	comes	to	fulfilling	their	needs	for	social	connection,	people	with	high	independent	self-construal	orientation	will	be	more	willing	to	conscript	social	resources,	or	‘friends’,	from	a	variety	of	other	outside	sources,	in	order	to	fulfill	personal	needs,	compared	to	those	with	a	more	interdependent	self-construal.	Consequently,	I	expect	individuals	with	high	independent	self-construal,	under	normal	conditions	(control),	to	not	only	anthropomorphize	more	than	individuals	with	high	interdependent	self-construal,	but	also	to	anthropomorphize	more	following	an	exclusion	manipulation,	and	less	following	an	inclusion	manipulation.				 33	This	expectation	is	because,	as	previously	discussed,	individuals	with	high	independent	self-construal	are	more	prone	to	construct	in-groups	to	suit	current	needs.	Furthermore,	if	needs	are	satiated	in	one	domain	(for	example	social	connection),	they	will	be	more	likely	to	move	on	to	look	for	greener	pastures	in	the	form	of	better	social	status	opportunities.	As	a	result	I	have	developed	four	more	hypotheses,	as	follows:	Hypothesis	3:		People	with	higher	independent	self-construal	will	put	a	higher	percentage	of	animals	into	a	self-formed	in-group,	following	an	exclusion	manipulation,	than	they	will	do	normally	(control	condition).		 Hypothesis	4:		People	with	higher	independent	self-construal	will	give	higher	ratings	of	anthropomorphism	for	a	non-human	subject,	following	an	exclusion	manipulation,	than	they	will	do	normally	(control	condition).		Hypothesis	5:		People	with	higher	independent	self-construal	will	put	a	lower	percentage	of	animals	into	a	self-formed	in-group,	following	an	inclusion	manipulation,	than	they	will	do	normally	(control	condition).		 Hypothesis	6:		People	with	higher	independent	self-construal	will	give	lower	ratings	of	anthropomorphism	for	a	non-human	subject,	following	an	inclusion	manipulation,	than	they	will	do	normally	(control	condition).			 34	2.2.3	Prosocial	behaviour:	Mechanisms	and	motivations	for	donation	behaviour	and	inclusive	attitudes	explored.		We	make	a	living	by	what	we	get,	but	we	make	a	life	by	what	we	give.		Winston	Churchill		Giving	to	others,	in	the	form	of	money	or	time,	is	widely	cherished	across	many	cultures	as	a	moral	good,	if	not	a	moral	imperative.	But	why	do	we	give?	A	common	assumption,	founded	in	moral	and	religious	history,	and	evidenced	in	the	writings	of	Thomas	Aquinas,	David	Hume,	Adam	Smith,	Charles	Darwin,	Herbert	Spencer,	William	McDougall,	and	more	recently,	in	contemporary	psychology	(Hoffman,	1981;	Krebs,	1975;	Batson,	1987),	is	that	we	help	out	of	an	altruistic	desire	to	do	good	to	others;	a	desire	often	fanned	by	empathetic	feelings.	However,	a	variety	of	motivations	to	behave	in	a	prosocial	manner	exist,	beyond	pure	altruism.			Simpson	and	Willer	(2015)	highlight	the	critical	roles	that	social	mechanisms	such	as	normative	rules,	social	reputations	and	group	relations	play	in	our	prosocial	behaviour,	at	times	obscuring	the	true	underlying	individual	motivations.		Sturmer	and	Snyder	(2010),	approaching	prosocial	behaviour	from	a	group-level	theory,	argue	that	people	may	behave	prosocially	out	of	a	motivation	to	gain	social	or	material	rewards;	to	gain	satisfaction	at	sticking	or	demonstrating	personal	values;	to	avoid	feelings	of	distress,	shame	or	guilt	at	others’	suffering;	or	to	demonstrate	empathy	to	others.	In	the	latter	case	they	argue	that	empathy	concerns	may	be		 35	especially	motivational,	if	the	donor	feels	similar	or	close	to	the	cause.	Batson,	Ahmad,	and	Stocks	(2011)	adopt	a	similar,	but	benefit	directed,	approach,	and	separate	prosocial	motivation	into	four	forms	of	benefits,	only	one	of	which	involves	altruism.	These	benefits	are	egoism	(benefit	another	to	benefit	oneself),	altruism	(benefit	another	as	an	ultimate	goal),	collectivism	(benefit	another	to	benefit	the	group),	and	principlism	(benefit	another	to	uphold	a	moral	principle).	Duclos	and	Barasch	(2014)	also	argue	that	charitable	behaviour	towards	others	is	most	frequently	be	driven	by	hopes	of	a	self-benefit,	such	as	that	of	personal	happiness.	In	line	with	this	prior	research	I	propose	that	most	of	the	prosocial	behaviour	that	is	in	evidence	in	my	research	will	take	place	as	the	result	of	either	a	‘self	benefit’	or	a	‘self	benefit	from	group	benefit’	motivation,	in	which	the	good	of	the	group	is	seen	as	overlapping	with	the	good	of	the	self.			Not	surprisingly,	motivations	to	behave	prosocially	are	also	usually	contingent	on	the	complex	social	relationships	that	exist	between	the	giver	and	receiver,	and	their	respective	groups.	In	many	situations	prosocial	behaviour	is	directed	towards	close	others,	family,	and	in-groups,	with	distant	others	are	viewed	as	competitors	(van	Vugt	&	Park,	2009).			Indeed,	Sturmer	and	Snyder	(2010)	found	that	when	people	behaved	prosocially	to	demonstrate	empathy	to	others,	it	appeared	to	be	especially	motivational	if	the	donor	felt	similar	or	close	to	the	cause.	This	makes	intuitive	sense	to	most	people,	as	may	be	evidenced	by	the	common	use	of	the	proverb	‘charity	begins	at	home.’	Offering	evidence	for	the	human	propensity	towards	in-group	focused	giving,	previous	research	has	demonstrated	that,	in	life	and	death		 36	situations,	people	are	more	likely	to	help	kin	than	non-kin	(Burnstein,	Crandall,	&	Kitayama,	1994).		Supporting	this	finding,	other	research	has	demonstrated	that	primes	of	mortality	salience,	in	North	Americans,	leads	to	an	increase	in	donations	that	are	specifically	directed	to	in-group	charities	(Jonas,	Schimel,	Greenberg	&	Pyszczynski,	2002).	Do	thoughts	of	impending	death	prompt	us	to	draw	up	the	hatches	on	a	castle	under	siege,	as	suggested	by	the	findings	of	Castano	(2004),	who	demonstrated	that	people	under	a	mortality	salience	threat	are	more	likely	to	tightly	define	an	in-group,	and	strictly	regulate	who	is	admissible?		Despite	the	remonstrations	of	Singer	(2009)	that,	from	a	utilitarian	perspective,	there	is	nothing	remotely	rational	about	giving	to	our	nearest	and	dearest	instead	of	to	distant	others,	it	seems	that	we	may	be	hard	wired	to	prioritize	‘close	others.’		This	appears	to	hold	true	even	when	the	categorization	‘close	others’	makes	little	sense.	Brewer	(1979)	and	Tajfel	&	Turner	(1986)	have	previously	both	found	that,	even	when	groups	were	divided	by	trivial	criteria	(e.g.	the	minimal	group	paradigm),	people	preferred	to	give	time	or	money	to	in-groups.	Likewise,	Kuchenbrandt,	Eyssel,	Bobinger,	and	Neufeld	(2013)	have	found	that	when	robots	were	categorized	as	in-group	members	(vs.	out-group	members),	participants	were	more	likely	to	anthropomorphize	them,	interact	with	them,	and	increase	their	evaluations	of	them.	As	Cialdini	(1984)	famously	noted,	the	influential	pull	of	‘similarity’	is	strong,	even	when	its	supposition	is	false.			Nonetheless,	some	notable	examples	of	out-group	helping	do	exist	(Van	Leeuwen,		 37	2007;	Van	Leeuwen,	&	Täuber,	2010;	Van	Leeuwen,	&	Täuber,	2012),	which	poses	the	question	as	to	what	motivates	us	to	help	distant	others?	Hopkins,	Reicher,	Harrison,	Cassidy,	Bull	and	Levine	(2007)	have	approached	this	question	from	a	strategic	angle,	demonstrating	that	group	members	may	be	motivated	to	help	out-groups,	but	for	very	different	reasons	to	those	motivating	in-group	helping.	Motives	for	out-group	helping	thus	include	the	wish	to	disconfirm	negative	stereotypes	of	their	group;	for	example	to	counter	the	supposed	meanness	of	Scottish	people	(Hopkins,	Reicher,	Harrison,	Cassidy,	Bull,	&	Levine,	2007).	Leeuwen	and	Täuber	(2012)	have	furthered	this	body	of	research	into	the	strategic	side	of	out-group	helping	(see	Leeuwen	&	Täuber,	2010,	for	an	over	view)	by	examining	instances	in	which	out-group	helping	may	also	act	as	a	tool	to	communicate	in-group	warmth,	or	to	bolster	collective	pride	(Leeuwen,	Dijk,	&	Kaynak,	2013).			Some	out-group	helping	may	have	more	sinister	self-orientated	motives,	however.	Although	some	dispute	the	necessity	of	an	out-group	(Brewer,	1999),	according	to	social	identity	theory	(Tajfel,	1959;	Tajfel	&	Turner,	1979)	the	denigration	of	out-groups,	and	the	bolstering	of	in-groups,	is	a	necessary	part	of	intergroup	behaviour,	with	an	out-group	being	a	necessary	requirement	for	in-group	definition	(Yuki	&	Takemura,	2013).	This	intergroup	comparison	behaviour	serves	as	an	effective	way	in	which	to	maintain	self-esteem;	serving	both	self-protective	and	self-enhancing	functions,	especially	in	times	of	identity	threat	(Branscombe,	Ellemers,	Spears,	&	Doosje,	1999).	Nadler	(2002;	2010)	and	Nadler,	Harpaz-Gorodeisky,	and	Ben-David	(2009)	have	described	cases	of	‘defensive	helping,’	in	which	out-group	assistance	is		 38	focused	on	low	status	dependent	groups.	In	this	manner,	defensive	helping	acts	as	a	way	for	an	in-group	to	maintain	social	dominance	and	power	over	an	out-group,	especially	in	the	face	of	threatened	in-group	stability.			This	approach	to	intergroup	helping	follows	a	social	identity	perspective	(Turner	&	Reynolds,	2001),	basing	itself	on	self-categorization	theory,	in	order	to	build	an	‘Intergroup	Helping	as	Status	Relations’	(IHSR)	model.	The	‘Intergroup	Helping	as	Status	Relations’	model	argues	that	dominant	groups	help	others,	in	order	to	signal	their	status	to	other	groups.	In	summary	therefore,	we	can	see	that	much	intergroup	prosocial	behaviour	is	motivated	by	self	benefit,	be	it	in	the	form	of	in-group	helping	or	out-group	helping	(Sturmer	&	Snyder,	2010;	Duclos	&	Barasch,	2014).			In	addition	to	the	membership	status	of	the	recipient	and	donor,	relative	to	each	other,	the	self-construal	of	the	donor	also	has	the	potential	to	impact	prosocial	behaviour.	Self-construal	orientation	appears	to	influence	attitudes	towards	a	range	of	societal	goods,	including	environmental	conservation,	sustainability	and	prosocial	behaviour	(Arnocky,	Stroink	and	DeCicco,	2007).	Past	research	has	suggested	that	individuals	from	individualistic	cultures	may	be	less	cooperative	than	individuals	from	collectivist	cultures,	because	the	former	mainly	focus	on	their	own	outcomes	and	less	on	the	welfare	of	others	(Hemesath	&	Pomponio,	1998;	Kagan	&	Knight,	1979;	McClintock,	1974;	Parks	&	Vu,	1994;	Probst,	Carnevale,	&	Triandis,	1999).	However,	as	previously	discussed,	according	to	Yuki	and	Takemura	(2013),	this	argument	runs	the	risk	of	grossly	oversimplifying	the	mechanisms	involved	in		 39	cultural	self-definition	and	social	identity.			Duclos	and	Barasch	(2014)	have	put	forward	evidence	for	a	more	nuanced	perspective,	arguing	that	a	more	collectivist,	or	interdependently	orientated	self-construal	orientation	is	specifically	connected	to	benevolent	and	prosocial	behaviour	towards	in-group	recipients.	This	finding	is	supported	by	the	research	of	Kemmelmeier,	Jambor	and	Letner,	(2006),	who	found	that	individualistic	societies	in	the	USA	were	more	disposed	to	favour	out-group	targets,	than	were	more	collectivist	societies.	This	may,	in	part,	be	because,	in	individualistic	societies,	personal	virtues	such	as	self-determination	and	self-identity	are	typically	valued	apart	from	group	membership.	Kemmelmeier,	Jambor	and	Letner	(2006),	in	the	same	research,	also	note	a	tendency	to	use	personal	choices,	such	as	voluntary	giving,	as	a	means	by	which	to	signal	personal	characteristics	and	interests	to	others,	which	provides	additional	support	for	the	proposition.			If	the	arguments	of	Duclos	and	Barasch	(2014)	are	correct,	and	much	charitable	behaviour	is	indeed	fueled	by	a	motivation	to	increase	personal	happiness,	it	appears	reasonable	to	expect	that	this	will	be	fulfilled	in	different	ways	by	those	high	in	interdependent	self-construal,	compared	to	those	high	in	independent	self-construal.	After	all,	if,	according	to	Yuki	and	Takemura	(2013),	people	high	in	interdependent	self-construal	regard	their	in-group	as	a	rigid	long-term	entity;	with	non-porous	boundaries	and	a	membership	that	is	fixed	for	life,	a	good	deed	done	to	an	in-group	member	is	likely	to	have	higher	potential	to	bring	personal	happiness		 40	longer	term	than	any	good	deed	done	to	an	out-group	member.	Conversely,	for	people	high	in	independent	self-construal,	in-group	boundaries	are	more	flexible	and	porous,	and	members	may	come	and	go.	As	a	result,	an	out-group	member	today	might	possibly	be	an	in-group	member	tomorrow.	The	consequence	of	this	is	that	people,	with	a	high	independent	self-construal	orientation,	might	be	expected	to	be	as	likely	to	provide	beneficence	to	an	out-group	member	as	an	in-group	member.	Indeed,	previous	research	has	demonstrated	that	out-group	recipients	may	always	fare	better	seeking	support	from	independent	self-construal	orientated	donors	(Duclos	&	Barasch,	2014).		In	an	ideal	world	there	would	be	no	prejudice	towards	others,	we	would	exist	in	harmony	without	suffering	and	pain,	and	all	humanity	would	adhere	to	the	highest	moral	codes	of	conduct	in	all	relationships.		Philosophers,	psychologists,	sociologists	and	policymakers,	however,	are	left	to	grapple	with	the	practicalities	of	poor	day-to-day	behaviour,	and	aggressive	forms	of	prejudice	and	discrimination	towards	others.	Much	interest	has	been	directed	at	discovering	methods	to	combat	prejudice	and	promote	more	inclusive	attitudes	towards	distant	others.			One	area	for	potential	research	is	grounded	in	the	topic	of	moral	identity,	and	explores	the	options	for	expanding	our	moral	circle	of	concern	to	more	distant	others,	as	a	method	for	improving	intergroup	conflict,	fostering	benevolence,	and	promoting	inclusive	behaviour	to	out-groups	(Aquino,	&	Reed,	2002;	Reed	&	Aquino,	2003).	The	‘moral	circle’	is	the	boundary	drawn	around	those	entities	in	the		 41	world	deemed	worthy	of	moral	consideration,	and	may	be	influenced	by	inclusion	and	exclusion	framing	effects	(Laham,	2009),	as	well	as	similarity	and	difference	frames	(Bastian,	Costello,	Loughnan,	&	Hodson,	2011).		Gaertner,	Mann,	Murrell,	and	Dovidio	(1989),	in	a	similar	vein,	have	suggested	that	the	re-categorization	of	individual	members	representations	of	an	aggregate,	has	potential	to	act	as	a	method	for	reducing	both	negative	bias	towards	out-group	members,	as	well	as	positive	bias	towards	in-group	members.	Levine,	Prosser,	Evans,	and	Reicher	(2005),	likewise,	offer	evidence	that	expanding	social	categorization,	via	a	shared	group	membership	(rival	soccer	teams	vs.	all	soccer	fans),	may	be	a	successful	manner	in	which	to	increase	helping	behaviour	towards	recipients	in	need,	and	one	that	allows	for	a	transgression	of	traditional	group	boundaries.			Focusing	on	the	commonalities	that	we	share	with	others	has	been	found	to	improve	our	view	of	them	(Dovidio,	Gaertner,	Isen,	&	Lowrance,	1995;	Gaertner,	Dovidio,	Banker,	Houlette,	Johnson,	&	McGlynn,	2000).	Indeed,	Bastian,	Costello,	Loughnan,	and	Hodson	(2011)	found	that	comparing	animals	to	humans	had	the	result	of	not	only	expanding	moral	concern	towards	animals,	but	also	towards	human	out-groups;	reducing	both	speciesism	and	racial	prejudice.		Recent	research,	looking	at	human-animal	relations	from	an	intergroup	perspective,	has	likewise	found	animal	attitudes	are	higher	when	a	more	encompassing	frame	that	includes	‘other’	within	self	is	applied	(Amiot	&	Bastian,	2017).	Batt	(2009)	likewise	has	found	that	attitudes	towards	animals	improve	in	accordance	with	how	much	participants	perceive	the	animal	species	to	be	similar	to	themselves.		Tam,	Lee,	and	Chao	(2013)		 42	have	demonstrated	that	when	nature	is	anthropomorphized,	people	are	both	more	likely	to	feel	connected	to	it,	and	more	likely	to	engage	in	protective	conservation	behaviour.	I,	therefore,	predict	that	primes	of	similarity	to	a	cause	will	increase	an	individual’s	feelings	of	connectedness	to	it,	as	well	as	increasing	donation	support	for	it.	However,	once	more,	I	predict	that	self-construal	and	inclusion	status	will	moderate	these	results.		In	summary,	past	research	has	delivered	mixed	evidence	in	the	field	of	prosocial	behaviour	towards	others,	following	threats	or	affirmations	of	inclusion.	While	some	research	has	suggested	that	social	exclusion	may	negatively	predict	prosocial	behaviour	(Beest	&	Williams,	2011;	Catanese	&	Tice,	2005;	Twenge,	Baumeister,	DeWall,	Ciarocco,	&	Bartels,	2007),	other	research	has	suggested	that	it	may	positively	predict	it	(Lee	&	Shrum,	2012).	My	proposition	is	that	the	discrepant	findings	might	be	due,	at	least	in	part,	to	a	failure	to	account	for	the	role	that	self-construal	orientation	may	play	in	determining	responses	to	exclusion	and	inclusion	threats,	in	the	context	of	charitable	giving.	Moreover	I	propose,	additionally,	that	the	influence	that	self-construal	orientation	may	have,	on	perceptions	of	similarity	(vs.	difference)	to	target	cause,	should	also	be	accounted	for.	I	posit	that,	just	as	with	my	predictions	regarding	the	group	formation	task,	people	with	an	independent	self-construal	orientation	will	express	greater	donation	intentions	for	an	out-group	charity,	and	more	feelings	of	psychological	closeness	and	connection	to	the	cause,	than	people	with	an	interdependent	self-construal	orientation.	However,	I	expect	this	effect	to	be	mitigated	as	a	result	of	an	affirmation	of	social	belonging	(i.e.,	as	in		 43	the	inclusion	condition),	as	well	as	their	perceptions	of	how	similar,	or	different,	they	are	to	the	target	cause.			In	the	current	research,	I	endeavor	to	build	upon	these	two	areas	of	prosocial	research–work	on	charitable	giving	and	work	on	inclusive	behaviour	expressed	towards	others–by	examining	the	moderating	roles	that	self-construal	orientation,	and	inclusion	status,	may	play.	In	line	with	past	research	(Duclos	&	Barasch,	2014),	I	predict	that	individuals	high	in	independent	self-construal	will	normally	behave	more	prosocially	towards	out-group	members	than	will	individuals	high	in	interdependent	self-construal.	I	argue	that	this	is	as	a	result	of	two	factors:	first	because	individuals	high	in	independent	self-construal	form	more	flexible	dynamic	in-groups	based	on	shared	categories	than	do	individuals	high	in	interdependent	self-construal,	and	second	because	individuals	high	in	independent	self-construal	are	more	vigilant	to	cues	of	social	connection	potential	in	out-group	members,	and	will	behave	in	a	more	socially	opportunistic	manner	if	a	self-benefit	to	do	so	is	apparent.			As	a	result,	when	we	compare	high	independent	self-construal	individuals	to	high	interdependent	self-construal	individuals,	the	former	may	appear	‘fair-weather	friends,’	adding	and	dropping	others	from	flexible	in-groups	in	order	to	fulfill	social	resource	needs.	On	the	other	hand,	the	latter	may	appear	to	be	more	socially	rigid	and	loyal,	tightly	bound	to	pre-determined	in-groups,	with	little	opportunity	for	intergroup	movement	or	admission	of	outsiders.	Following	an	assurance	of	social		 44	inclusion,	however,	I	propose	that	this	pattern	will	not	persist.	I	predict	that,	under	circumstances	when	social	inclusion	is	assured,	high	independent	self-construal	individuals	will	re-assess	their	grouping	priorities	and	switch	to	a	more	in-group	defensive	strategy.	This	strategy,	I	propose,	will	be	more	akin	to	that	seen	in	individuals	high	in	interdependent	self-construal,	in	which	in-group	similarity	and	distinctiveness	is	important,	in-group	boundaries	must	be	defended,	and	out-groups	must	be	strongly	differentiated	from.			I	also	predict	that	individuals	high	in	independent	self-construal	will	be	highly	responsive	to	manipulations	that	re-categorize	out-group	members	as	in-group	members.	Specifically,	I	argue	that,	under	normal	(control)	conditions,	individuals	high	in	independent	self-construal	will	be	more	likely	to	feel	higher	levels	of	connection	to	an	out-group	cause	framed	as	being	similar,	and	express	higher	prosocial	/donation	behaviour	towards	an	out-group	cause	framed	as	being	similar,	than	to	an	out-group	cause	framed	as	being	different.	However,	following	an	affirmation	of	social	inclusion,	I	predict	this	pattern	will	reverse,	and	that	individuals	high	in	independent	self-construal	will	prefer	out-group	causes	that	are	different	from	the	in-group,	since	they	offer	less	threat	to	in-group	distinctiveness.			As	a	result,	I	predict	that	framing	an	out-group	cause	as	similar	(vs.	different)	to	an	individual	high	in	independent	self-construal	will	moderate	donation	support.	Moreover,	I	predict	that	the	interaction	between	self-construal	and	inclusion	status	will	moderate	this	moderation.	Finally,	I	predict	that	feelings	of	connection	to	the	cause	will	mediate	donation	intentions,	for	individuals	high	in	independent	self-	 45	construal.	As	a	result,	I	have	generated	the	following	remaining	hypotheses:		 Hypothesis	7:			Under	normal	conditions	(control	group)	people	with	a	higher	independent	self-construal	orientation	will	express	increased	donation	intentions	for	an	out-group	animal	charity	than	people	with	a	low	independent	self-construal	orientation.			Hypothesis	8:			Under	normal	conditions	(control	group)	people	with	a	higher	independent	self-construal	orientation	will	express	increased	donation	intentions	for	an	out-group	animal	charity	that	is	framed	as	similar	(vs.	different)	to	them,	since	it	offers	increased	connection	potential.			 Hypothesis	9:		Affirmation	of	social	inclusion	(inclusion	manipulation)	will	result	in	reduced	donation	intentions	expressed	for	an	out-group	animal	charity	for	people	with	a	higher	independent	self-construal	orientation,	compared	to	under	normal	conditions.			Hypothesis	10:		Affirmation	of	social	inclusion	(inclusion	manipulation)	will	result	in	increased	donation	intentions	for	an	out-group	animal	charity	that	is	framed	as	different	(vs.	similar)	to	the	donor	for	individuals	with	a	higher	independent	self-construal	orientation,	since	it	offers	less	of	a	threat	to	in-group	distinctiveness.					 46	Hypothesis	11:		Feelings	of	connection	to	the	cause	will	mediate	donation	support	for	individuals	with	high	independent	self-construal	orientation.			Hypothesis	12:		People	with	an	interdependent	self-construal	orientation	will	not	express	higher	or	lower	donation	intentions	as	a	result	of	social	inclusion	(delivered	via	an	inclusion	manipulation).				 		 47	Chapter	3:	Empirical	Investigation			3.1	Pre-Test	(Social	Potential)	Study	3.1.1	Overview.	My	initial	study	(study	1)	had	two	goals.	The	first	goal	was	to	explore	the	role	that	self-construal	might	play	in	group	formation	and	inclusive	behaviour	towards	others.	The	second	goal	was	to	explore	whether	affiliation	motivation–the	desire	to	socially	connect	with	others–might	also	play	a	role	in	group	formation,	and	specifically	the	decision	to	include	traditional	outsiders,	such	as	animals,	into	a	self-formed	in-group.	However,	I	expected	this	to	be	contingent	on	whether	the	animals	offered	an	opportunity	for	social	interaction.	Support	for	this	pattern	of	behaviour	comes	from	previous	research,	which	has	demonstrated	that	people	are	chronically	motivated	to	create	and	look	for	social	connections	and,	if	no	suitable	human	options	are	available,	they	will	be	open	to	attributing	human	characteristics	onto	animals,	or	other	objects	(Epley,	Akalis,	Waytz,	&	Cacioppo,	2008;	Gardner,	Pickett,	&	Knowles,	2005;	Powers,	Worsham,	Freeman,	Wheatley,	&	Heatherton,	2014),	in	an	attempt	to	re-frame	them	as	social	resources,	capable	of	gratifying	social	needs.	This	behaviour	has	been	termed	social	anthropomorphism	(Epley,	Akalis,	Waytz,	&	Cacioppo,	2008),	or	social	shielding	(Gardner,	Pickett,	&	Knowles,	2005).		Not	all	non-humans,	however,	are	likely	to	offer	the	same	degree	of	potential	for	social	interaction.	For	example,	a	fish	might	be	considered	to	offer	less	“social	interaction”	potential,	than	a	friendly	dog,	for	most	people.	Similarly,	with	humans,		 48	there	is	likely	to	be	a	variation	in	“social	interaction”	potential,	with	a	stranger	perhaps	offering	less	potential	as	a	meaningful	social	resource	than	a	friend	or	valued	family	member.		With	this	in	mind,	this	pre-test	(social	potential)	study	was	designed	with	the	specific	purpose	of	building	a	list	of	human	and	animal	items	that	would	represent	a	range	of	potential	for	social	connection	and	use	as	a	social	resource.			3.1.2	Procedure.	Materials.	The	pre-test	study	was	administered	in	the	marketing	research	lab	of	a	Canadian	University,	using	47	undergraduate	student	participants	(49%	female,	Median	age	=	21-25years	old),	who	participated	in	the	study	in	exchange	for	course	credit.	Participants	completed	the	study	on	Qualtrics	Survey	Software,	at	individual	computer	terminals	in	groups	of	10-18	per	session,	and	were	told	that	they	would	be	completing	a	single	task	with	no	further	explanation	given.	Participants	were	asked	to	rate	twenty-four	animal	items	and	twenty-four	human	items,	on	a	ten	point	scale,	for	their	potential	to	offer	a	social	interaction	likely	to	provide	an	opportunity	for	one,	or	more,	of	the	following	elements:	nurturing,	physical	or	emotional	intimacy,	love,	trust,	companionship,	communication,	comfort,	validation,	affection,	support,	and/or	a	feeling	of	belonging,	being	needed	or	being	understood	by	another.				 49	Animal	items	were	chosen	to	include	a	wide	range	of	domestic	and	wild	animals	that	were	likely	to	be	regarded	as	having	high	vs.	low	social	connection	potential,	and	included	items	such	as	“butterfly”,	“horse”,	“chick”,	“small	monkey”,	“cat”,	and	“dolphin”	on	the	high	social	potential	side	and	“snail”,	“goldfish”,	“tuna”,	“turkey”,	“alligator”,	“stick	insect”	on	the	low	social	potential	side.		Human	items	were	also	chosen	to	include	a	range	of	items	that	were	likely	to	be	regarded	as	having	high	vs.	low	social	connection	potential,	and	included	items	such	as	“nurse”,	“friend”,	“manicurist”,	”partner”,	“advice-columnist”,	“mother”	on	the	high	social	potential	side	and	“dentist”,	“stranger”,	“tele-marketer”,	“dead	person”,	“border-control	officer”,	“sleeping	person”	on	the	low	social	potential	side.		The	rating	section	of	the	study	took	less	than	ten	minutes	to	complete.		Following	this	task,	basic	demographics	were	collected	and	participants	were	thanked	and	debriefed.		3.1.3	Results.	Demographics.	Demographics	collected	showed	that	participants	in	the	sample	identified	as	64.4%	Chinese,	8.9%	South	Asian,	8.9%	Korean,	6.7%	South	East	Asian,	2.2%	Caucasian,	2.1%	Japanese,	and	6.7%	other.	Results	also	showed	that	13.3%	of	participants	reported	having	lived	in	the	USA	or	Canada	for	over	21	years,	35.6%	for	11-21	years,	24.4%	for	6-10	years,	and	26.7%	for	under	6	years.	Participants	identified	as	51.1%	female.	Age	demographics	(collected	as	a	categorical	variable)	were	as		 50	follows:	35.6%	under	21	years	old;	53.3%	21-25	years	old;	8.9%	26-30	years	old	and	2.2%	over	30	years	old.			Mean	ratings	were	collected	for	all	animal	items,	which	ranged	from	a	high	of	M=8.91	(SD	=	2.09)	for	a	Labrador	Dog	to	a	low	of	M=3.27	(SD=2.42)	for	a	Skunk.	The	average	for	animal	ratings	was	M=5.74	(SD=1.56).	Mean	ratings	were	collected	for	all	human	items	which	ranged	from	a	high	of	M=9.84	(SD	=	1.83)	for	Mother	to	a	low	of	M=4.18	(SD=2.92)	for	a	dead	person.	The	average	for	human	ratings	was	M=7.16	(SD=1.24).	See	appendix	A.1	for	detailed	analysis	of	results.			Following	the	analysis	five	low	social	animal	items	and	five	high	social	animal	items	were	chosen,	with	preference	given	to	the	highest	and	lowest	examples,	as	measured	by	their	means.	As	a	result	a	final	animal	list	was	created	which	composed	of	the	following	items:	(high	social)	Labrador	Dog	(M=8.91,	SD=2.09);	Horse	(M=7.93,	SD=2.31);	Dolphin	(M=7.82,	SD=2.58);	Cat	(M=7.58,	SD=2.28);	Rabbit	(M=7.07,	SD=2.32)	and	(low	social)	Skunk	(M=3.27,	SD=2.37);	Snail	(M=3.39,	SD=2.42);	Stick	Insect	(M=3.59,	SD=2.84);	Tuna	(M=4.07,	SD=2.70);	Alligator	(M=4.24,	SD=2.94).	Similarly,	a	list	of	five	low	social	human	items	and	five	high	social	human	items	were	chosen,	with	preferences	given	to	the	highest	and	lowest	examples,	as	measured	by	their	means.	As	a	result,	a	final	human	list	was	created	which	composed	of	the	following	items:	(high	social)	Mother	(M=9.84,	SD=1.83);	Partner	(M=9.63,	SD=2.10);	Friend	(M=9.50,	SD=1.72);	Co-worker	(M=8.19,	SD=1.78);	Nurse	(M=7.65,	SD=1.58)	and	(low	social)	Tele-marketer	(M=4.78,		 51	SD=2.76);	Sleeping	Person	(M=5.13,	SD=2.69);	Border	Control	Officer	(M=5.46,	SD=2.55);	Mall-Security	Guard	(M=5.64,	SD=2.00);	Stranger	(M=5.80,	SD=2.02).		3.1.4	Discussion.	The	results	of	the	pre-test	enabled	a	final	list	of	twenty	items	to	be	created	that	was	balanced	between	items	that	ranged	in	terms	of	social	connection	potential,	and	included	a	balance	of	human	and	animal	items.	Following	on	from	the	pre-test,	the	first	study	was	constructed	to	examine	if	a	relationship	existed	between	self-construal	and	the	types	of	items	that	were	selected	in	a	group	formation	task,	and	to	explore	whether	affiliation	motivation	towards	out-group	members	might	play	a	role	in	decisions	regarding	in-group	membership.			3.2	Study	1		3.2.1	Overview.	Past	research	has	found	that	people	high	in	independent	self-construal	may	be	more	disposed	to	extend	favourable	attitudes	and	behaviours	towards	prototypical	out-group	members	than	people	with	high	interdependent	self-construal	(Kemmelmeier,	Jambor	&	Letner,	2006;	Duclos	&	Barasch,	2014).	One	of	the	reasons	for	this	finding	may	lie	in	the	differing	perceptions	of	what	potential	exists	for	intergroup	mobility.	As	previously	noted,	Yuki	&	Takemura	(2013)	propose	that	individuals	with	a	more	interdependent	self-construal	orientation	are	more	likely	to	view	their	in-group	as	a	non-permeable	long-term	structure,	within	which	inter-connective	networks	must	be	kept	in	harmony	and	group	membership	is	fixed.			 52	Conversely,	individuals	with	a	more	independent	self-construal	orientation	are	more	likely	to	see	in-groups	as	more	dynamic	and	fluid	in	nature,	with	the	group	category,	not	the	membership,	being	paramount	to	identity.			While	more	a	more	dynamically	defined	and	relaxed	approach	to	in-group	membership	and	boundaries	may	have	advantages,	one	potential	negative	consequence	is	that	belonging	may	not	be	as	assured,	or	guaranteed,	since	social	inclusion	is	experienced	as	a	more	precarious	state	(Triandis,	1989).	Cushman	(1990)	has	gone	as	far	as	to	argue	that	a	more	independent	outlook,	such	as	that	found	in	a	person	high	in	independent	self-construal,	may	render	a	person	under	constant	threat	of	becoming	socially	isolated,	and	lacking	in	social	sustenance.	This	may	be	contrasted	with	the	chronically	assured	belonging	status,	experienced	by	individuals	with	higher	interdependent	self-construal,	who	are	confident	that	they	will	remain	as	members	of	their	more	rigid	fixed	groups,	bound	as	they	are	by	interconnected	ties	that	buffer	and	protect	them	against	threats	to	belonging	(Gardner,	Pickett	&	Knowles,	2005;	Powers,	Worsham,	Freeman,	Wheatley,	&	Heatherton,	2014;	Uskul	&	Over,	2014).		With	this	in	mind,	one	task	of	study	1	was	to	explore	whether	people	high	in	interdependent	self-construal	would	be	more	rigid	in	terms	of	in-group	definitions	and	less	likely	to	include	animals.		Another	task	of	study	1	was	to	explore	whether	affiliation	motivation–the	desire	to	form	social	connections–might	play	a	role	in	the	decision	to	include	traditional	outsiders,	such	as	animals,	into	an	in-group.	As	previously	indicated,	support	for	this		 53	motivation	is	supplied	by	research	demonstrating	that	people	are	motivated	to	create	and	look	for	social	connections,	and,	if	no	suitable	human	options	are	available,	they	will	be	more	open	to	include	non-humans	in	their	social	framework	(Epley,	Akalis,	Waytz	&	Cacioppo,	2008).		Belonging	status	is	vital	to	monitor	and	maintain.		Past	research	has	put	forward	a	Sociometer	or	social	monitoring	system	(SMS)	concept,	in	which	attention	to	social	aspects	of	the	environment	are	heightened,	whenever	a	threat	is	perceived	(Leary,	Tambor,	Terdal,	&	Downs,	1995;	Leary	1999;	Pickett,	&	Gardner,	2005).	While	most	research	has	typically	measured	the	effect	of	the	social	monitoring	system	in	situations	of	social	rejection	(Leary,	Kelly,	&	Schreindorfer,	2001),	Gardner,	Pickett,	and	Knowles	(2005)	have	clearly	stated	that	the	purpose	of	the	social	monitoring	system	is	to	attune	individuals	to	information	that	will	help	them	navigate	the	social	environment	more	successfully,	whatever	the	belonging	state	they	find	themselves	in.	In	light	of	this	argument,	I	propose	that	it	is	reasonable	to	expect	that	people	high	in	independent	self-construal	will	be	more	likely	to	actively	monitor	their	state	of	belonging,	due	to	the	more	precarious	nature	of	their	social	bonds,	and	will	therefore	be	especially	attuned,	vigilant,	and	responsive	to	signals	of	social	interaction	potential	in	others.			With	this	past	research	in	mind,	study	1	was	designed	to	examine	whether	self-construal	effects	(independent	vs.	interdependent	self-construal)	would	manifest	in	a	group	selection	task,	in	which	participants	were	asked	to	identify	in-group		 54	members	from	the	pre-tested	list	of	human	and	animal	targets.	I	predicted	that	individuals	with	a	higher	independent	self-construal	orientation	would	be	more	open	to	selecting	animals	to	put	in	an	in-group	than	individuals	with	a	more	interdependent	self-construal	orientation,	for	two	reasons.	First,	I	expected	this	increased	openness	among	individuals	with	a	higher	independent	self-construal	because	individuals	with	a	higher	independent	self-construal	orientation	have	a	more	flexible	view	of	what	may	be	categorized	as	an	in-group	member,	along	with	a	greater	expectation	for	relational	mobility	(Yuki	&	Takemura,	2013).	In	contrast,	I	predicted	that	individuals	with	a	more	interdependent	self-construal	orientation	not	only	believe	that	there	is	less	potential	for	intergroup	mobility,	and	regard	group	boundaries	as	being	less	porous,	but	also	will	be	more	likely	to	regard	animals	as	rigidly	out-group,	and	therefore	will	be	inclined	to	include	less	animals	in	a	self-formed	in-group	category.			Second,	I	anticipated	that	individuals	with	a	more	independent	self-construal	orientation	would	be	more	motivated	to	group	according	to	social	connection	potential,	since	their	belonging	needs	are	less	guaranteed,	as	a	result	of	the	more	precarious	and	dynamic	nature	of	their	existing	in-groups.	On	the	other	hand,	individuals	with	a	more	interdependent	self-construal	would	be	more	motivated	to	group	according	to	existing	relational	connections,	along	with	more	rigid	and	formal	group	rules.	In	summary,	my	hypotheses	(H1)	were	that	animals	would	form	a	higher	percentage	of	a	self-formed	in-group	the	higher	a	person	measures	on	independent	self-construal,	and	a	lower	percentage	of	a	self-formed	in-group	the		 55	higher	a	person	measures	on	interdependent	self-construal.	Furthermore,	I	predicted	(H2)	that	social	affiliation	motivation	would	have	a	greater	influence	the	higher	a	person	measured	on	independent	self-construal.		3.2.2	Procedure.	A	sample	of	95	American	participants	was	recruited	using	Mechanical	Turk,	and	completed	the	study	online	using	Qualtrics	Survey	Software,	in	exchange	for	monetary	compensation.		Sample	size	was	based	on	practical	considerations	regarding	recruitment	and	budgetary	constraints,	but	I	aimed	to	get	100	people	for	this	first	correlational	study.	The	study	was	a	correlational	design	that	included	a	grouping	task,	as	well	as	a	measure	of	trait	self-construal	orientation,	using	a	previously	validated	24-item	scale	(Singelis,	1994).			Materials.	Participants	were	told	that	they	would	be	completing	two	unconnected	tasks.	The	first	task	required	them	to	complete	a	series	of	measures,	which	began	with	a	condensed	10-item	Individual	Differences	in	Anthropomorphism	Questionnaire	of	which	5-items	were	intended	to	measure	anthropomorphism	(condensed	IDAQ;	α	=	.778).	This	condensed	10-item	IDAQ	contained	only	animal	related	questions	and	was	adapted	from	a	longer	30-item	measure	devised	by	Waytz,	Cacioppo	and	Epley	(2010)	which	also	contained	machine	and	nature	related	questions.		The	condensed	IDAQ	was	administered	for	exploratory	reasons,	since	it	has	previously	been	shown	to	be	predictive	of	increased	moral	care	and	concern	being	expressed	towards		 56	animals.		The	condensed	IDAQ	contained	questions	such	as	“	To	what	extent	does	a	cheetah	experience	emotions?”	and	“	To	what	extent	does	the	average	fish	have	free	will?”.		After	the	condensed	IDAQ	participants	completed	the	24-item	Self-construal	scale	(Singelis,	1994;	independent	self-construal	α	=	.817,	interdependent	self-construal	α	=	.758).	This	scale	includes	twelve	questions	designed	to	measure	trait	independent	self-construal,	such	as	“My	personal	identity,	independent	of	others,	is	very	important	to	me	”	and	“I	enjoy	being	unique	and	different	from	others	in	many	respects”,	as	well	as	twelve	questions	designed	to	measure	trait	interdependent	self-construal,	such	as	“I	have	respect	for	the	authority	figures	with	whom	I	interact”	and	“Even	when	I	strongly	disagree	with	group	members,	I	avoid	an	argument”.		This	was	followed	by	a	12-item	Fear	of	Negative	Evaluation	(FNE)	scale	(Leary,	1983;	α	=	.753),	designed	to	measure	people’s	sensitivity	to	criticism	from	others,	and	also	used	for	exploratory	purposes.	The	FNE	contains	questions	such	as	“	I	am	afraid	that	people	will	find	fault	with	me”	and	“	Other	peoples	opinions	of	me	do	not	bother	me”.		Participants	then	completed	a	3	item	loneliness	scale	(Hughes,	Waite,	Hawkley,	&	Cacioppo,	2004;	α	=	.910)	that	consisted	of	the	following	questions:	“Do	you	feel	that	you	lack	companionship?”,	“Do	you	feel	isolated	from	others?”	and	“Do	you	feel	left	out?”			The	IDAQ,	and	FNE	scales	were	administered	as	a	cover	story	for	the	study	being	offered	as	two	unconnected	tasks.	However,	I	also	wished	to	measure	participant’s	trait	measures	on	IDAQ	and	FNE,	and	check	for	correlations	with	self-construal	orientation,	loneliness,	and	grouping	of	humans	and	animals.	Anthropomorphism,		 57	as	previously	detailed,	has	been	shown	to	increase	with	a	desire	for	social	connection	(Epley,	Akalis,	Waytz,	&	Cacciopo,	2008).		FNE	has	been	shown	to	impact	impression	management	and	social	comparison	behaviour,	in	intergroup	settings,	and	has	been	linked	to	stereotyping	of	out-group	members	and	polarization	of	evaluations	of	both	in-and	out-group	members	(Stephen	&	Stephen,	1985).	The	3-item	loneliness	scale	was	administered	to	measure	loneliness,	which	has	previously	been	shown	to	impact	people’s	attitudes	towards	animals	(Epley,	Akalis,	Waytz,	&	Cacciopo,	2008).				Following	these	preliminary	measures,	participants	were	told	they	were	now	to	complete	second	unrelated	task,	which	was	designed	to	investigate	how	people	categorize	information.	They	were	asked	to	create	two	groups,	from	a	list	of	humans	and	animal	items	that	included	human	items	such	as	“your	mother”,	“a	friend”,	“a	telemarketer”,	“a	sleeping	person”	and	animal	items	such	as	“a	dog”,	“a	cat”,	“a	snake”,	“a	snail”.	See	appendix	A.4	for	full	list	and	details.	The	groups	were	depicted	as	empty	boxes,	with	one	box	already	containing	a	“yourself”	item	at	the	start.	Participants	were	requested	to	click	and	drag	items	into	a	relevant	box;	to	leave	no	item	unselected;	and	were	told	that	each	group	must	have	at	least	8	items	within	it.		Following	the	grouping	task,	participants	were	asked	a	number	of	demographic	questions.	Finally,	participants	were	probed	for	hypothesis	guessing,	in	an	open-ended	response	format,	and	thanked.					 58	3.2.3	Results.	On	inspection	of	the	open-ended	responses,	no	participants	appeared	to	guess	the	study	1	hypotheses.	My	exclusion	policy	was	to	exclude	participants	if	they	failed	all	attention	checks,	or	if	they	accurately	guessed	the	key	study	hypotheses.	Ten	participants	failed	all	attention	checks	and	were	therefore	removed	from	the	study	(leaving	n=85	remaining).			Power.	A	post	hoc	power	analysis,	using	G*Power	3.1,	indicated	that	with	the	sample	size	remaining	(n=85)	there	was	sufficient	power	(>.80)	to	detect	correlations	of	.30	and	greater,	with	a	critical	r=0.2133	at	95%	confidence.			Demographics.	Demographics	collected	showed	that	participants	in	the	sample	identified	as	82.4%	Caucasian,	14.1%	Black,	and	3.6%	mixed	other.	Results	also	showed	that	95.3%	of	participants	reported	having	lived	in	the	USA	for	over	21	years.	Participants	identified	as	44.7%	female.	Age	demographics	were	as	follows:	3.5%	under	21	years	old;	34.1%	21-30	years	old;	29.4%	31-40	years	old;	22.4%	41-50	years	old;	9.4%	51-60	years	old	and	1.2%	over	60	years	old.			A	combined	self-construal	measure	was	computed,	which	has	been	used	in	past	research	(Aaker	&	Williams,	1998;	Singelis,	1994),	and	is	reported	as	having	a	high	consistency	score	of	alpha	=.86	(Zhang,	Feick,	&	Price,	2006).	It	is	calculated	by		 59	adding	the	reversed	independent	self-construal	score	to	the	standard	interdependent	self-construal	score.	The	results	of	Study	1	showed	a	significant	negative	correlation	(r(83)=-.294,	p=.006)	between	self-construal	(measured	using	a	combined	self-construal	scale,	Zhang,	Feick,	&	Price,	2006;	α	=.724)	and	the	percentage	of	animals	that	were	included	in	the	in-group;	indicating	that	less	animals	were	included	the	higher	participants	measured	in	interdependent	self-construal	(and	more	animals	included	the	higher	participants	measured	in	independent	self-construal).	In	order	to	probe	whether	the	effect	was	driven	more	by	one	self-construal	orientation,	than	the	other,	I	also	separated	self-construal	into	its	two	components	to	run	tests.	These	subsequent	analyses	revealed	that	that	being	higher	in	independent	self-construal	was	significantly	positively	correlated	with	a	higher	percentage	of	animals	being	included	in	a	self	designated	in-group	(r(83)=.220,	p=.04).	The	reverse	was	true	with	interdependent	self-construal,	in	that	higher	interdependent	self-construal	was	positively	correlated	with	a	lower	percentage	of	animals	being	included	in	a	self	designated	in-group;	although	not	significantly	when	measured	alone	(r(83)=-.173,	p=.114).	Thus,	the	tendency	to	include	out-group	members	in	the	in-group	was	more	driven	by	being	higher	in	independence,	than	by	being	lower	in	interdependence.			Interestingly,	the	results	of	study	1	also	revealed	a	positive	correlation	(r(83)=.305,	p=.005)	between	independent	self-construal	orientation,	and	how	many	“high	social”	items	were	put	in	the	in-group.	Put	another	way,	the	higher	participants	rated	in	independent	self-construal	the	more	items	they	put	in	their	in-group	that		 60	were	rated	as	“high	social	potential”.	The	“high	social”	items	included	both	human	and	animal	items.		There	was	no	significant	correlation	between	the	number	of	“high	social”	or	“low	social”	items	put	in	the	in-group	for	participants	and	high	interdependent	self-construal	orientation	(r(83)=.055,	p=.614).		In	terms	of	trait	measures	the	results	of	study	1	also	revealed	a	significant	positive	correlation	(r(83)=.309,	p=.004)		between	anthropomorphism	(measured	in	the	IDAQ)	and	independent	self-construal.	Loneliness	(measured	using	the	3-item	loneliness	measure)	and	independent	self-construal	were	negatively	correlated		(r(83)=-.359,	p=.001).		Fear	of	negative	evaluation	(FNE)	and	Loneliness	were	positively	correlated	(r(83)=.321,	p=.003).		Study	1	also	showed	a	significant	positive	correlation	between	FNE	and	interdependent	self-construal	(r(83)=.245,	p=.024).		The	results	showed	that	neither	trait	anthropomorphism	(IDA),	(r(83)=.094,	p=.392),	nor	the	Loneliness	measure,	(r(83)=-.035,	p=.753),	were	significantly	correlated	with	the	percentage	of	animals	being	included	in	a	self-formed	in-group.	An	independent	samples	t-test	showed	no	significant	differences	between	genders	on	measures	of	any	of	the	variables	(FNE,	IDAQ,	self-construal,	percentage	of	animals	in	in-group).			3.2.4	Discussion.	As	predicted,	the	results	of	study	1	provided	support	for	H1;	that	the	higher	a	person	measures	on	independent	self-construal,	the	higher	the	percentage	of	a	self-formed	in-group	will	be	animal	out-group	members.	As	a	result,	study	1	provided		 61	some	non-causal	support	for	the	argument	of	Kemmelmeier,	Jambor	and	Letner	(2006);	that	interdependent	self-construal	is	associated	with	more	in-group	favouritism.	Additionally,	study	1	supported	the	proposition	of	Yuki	and	Takemura	(2013);	that	people	with	a	more	independent	self-construal	orientation	may	have	more	permeable	boundaries	to	their	in-groups	and,	if	sufficiently	motivated,	may	be	more	willing	to	allow	traditional	out-group	members	in.			The	results	of	study	1	also	show	that	independent	self-construal	and	trait	anthropomorphism	(measured	through	the	IDAQ)	appeared	to	be	positively	correlated,	offering	support	for	the	argument	that	people	with	a	more	independent	self-construal	orientation	may	be	more	open	to	categorizing	animals	as	potential	“honorary	humans.”		Trait	anthropomorphism	on	its	own,	however,	did	not	appear	to	be	correlated	with	inclusion	of	animals.				Study	1	also	revealed	a	positive	correlation	between	how	many	“high	social”	items	were	put	in	the	in-group	box,	and	high	independent	self-construal,	and	thereby	offered	support	for	H2:	that	the	higher	individuals	measure	in	independent	self-construal,	the	greater	the	number	of	items	that	have	high	social	connection	potential	will	be	put	in	a	self-formed	in-group.	There	was	no	correlation	between	the	number	of	“high	social”	potential	items	in	the	in-group	and	high	interdependent	self-construal.	This	result	provides	evidence	that,	while	participants	with	high	independent	self-construal	are	likely	to	include	more	“high	social”	humans	and		 62	animals	in	a	self-formed	in-group,	the	same	relationship	is	not	in	evidence	for	participants	with	high	interdependent	self-construal.			Taken	together,	the	results	of	study	1	therefore	offer	support	for	the	suggestion	that	at	least	some	of	the	motivation	for	the	choices	made	in	the	group	selection	task,	in	people	with	an	independent	self-construal	orientation,	results	from	a	desire	to	maximise	social	connection	potential.				The	next	task	was	to	investigate,	in	more	detail,	the	motivations	and	mechanisms	that	enable	traditional	outsiders,	such	as	animals,	to	be	admitted	into	an	in-group.	Study	1	had	offered	evidence	for	one	potential	motivation,	the	desire	to	socially	connect.	Study	2	therefore	was	designed	to	manipulate	social	connection.		Limitations.	A	limitation	of	study	1	may	be	noted	from	the	power	analysis,	which	showed	only	sufficient	power	to	detect	a	correlation	of	.30	and	higher.	While	some	of	the	correlations	were	over	.30,	others	were	in	the	region	of	.20-.30.	To	be	sure	of	the	robustness	of	these	results	it	would	therefore	be	advisable	to	replicate	this	study	in	future	research.	A	further	note	regarding	study	1	is	that	the	demographic	descriptives,	particularly	in	the	areas	of	ethnicity	and	age,	differed	considerably	from	the	pre-test	study	demographic	descriptives.						 63	3.3	Study	2	3.3.1	Overview.	Social	exclusion,	in	almost	any	form,	is	considered	to	have	a	significant	impact	on	individuals,	and	the	negative	consequences	of	failed	inclusion	can	be	powerful	and	immediate	(Williams,	2009).	However	the	detailed	impacts	of	exclusion	and	inclusion	on	subsequent	social	interactions	with	others–including	in	the	realms	of	prosocial	behaviour–appear	to	be	complex	and	nuanced.	As	previously	noted,	Epley,	Akalis,	Waytz,	and	Cacioppo	(2008)	argue	that	both	induced	exclusion,	and	chronic	isolation,	can	motivate	people	to	create	and	look	for	social	connections	with	others,	even	including	non-humans.	Gardner,	Pickett,	and	Knowles	(2005)	have	suggested	that	when	social	connection	with	humans	is	not	an	option,	seekers	may	instead	look	to	the	use	of	parasocials,	such	as	plants	or	pets—a	phenomenon	they	call	social	shielding.	While	less	research	has	examined	the	impacts	of	social	inclusion,	both	Brewer	(1991),	and	DeWall,	Baumeister,	and	Vohs,	(2008)	have	proposed	that	not	only	may	people	experience	belonging	satiation,	but	also	that	those	who	feel	socially	connected	may	be	less	motivated	to	connect	with	others	or	gain	social	acceptance	elsewhere.	Waytz	and	Epley	(2012)	have	also	demonstrated	that	social	connection	may	enable	dehumanization	of	others,	and	increase	perceived	distance	between	ourselves	and	less	close	others.			A	variety	of	research	has	explored	the	impact	of	inclusion	and	exclusion	on	different	self-construal	orientations.	Pfundmair,	Aydin,	Du,	Yeung,	Frey,	and	Graupmann	(2015)	argue	that	possession	of	a	more	interdependent	self-construal	orientation		 64	may	buffer	a	person	from	the	negative	impacts	of	exclusion,	offering	protection	in	a	form	that	people	with	a	more	independent	self-construal	do	not	have	access	to.	Pfundmair	and	colleagues	also	found	that	participants	with	a	collectivist	orientation	showed	little	difference	on	a	range	of	measures	(antisocial	intention,	prosocial	intention,	avoiding	intention,	affect)	following	either	social	inclusion	or	social	exclusion,	as	compared	to	participants	with	a	more	individualist	orientation.		Considering	this	prior	research,	study	2	included	both	an	inclusion	and	an	exclusion	manipulation,	in	order	to	examine	how	social	exclusion	and	social	inclusion	might	impact	a	grouping	task,	and	to	further	probe	the	role	that	social	connection	might	be	playing	in	the	selection	of	in-group	members,	dependent	on	self-construal	orientation.		Based	on	the	previous	findings,	I	expected	people	with	a	more	independent	self-construal	to	respond	to	an	inclusion	manipulation,	in	which	they	were	assured	of	their	future	belongingness	status	with	other	humans,	by	showing	less	interest	in	interacting	with	non-humans,	and	therefore	including	fewer	animals	in	their	self	formed	in-group.		On	the	other	hand,	I	expected	people	with	an	independent	self-construal	orientation	to	respond	to	the	exclusion	manipulation–in	which	they	were	told	that	they	would	be	more	socially	isolated	over	their	lifespan–by	showing	more	interest	in	interacting	with	non-humans,	and	therefore	including	more	animals	in	their	self	formed	in-group.	In	addition,	I	predicted	that	participants	measuring	high	in	interdependent	self-construal	might	demonstrate	less	of	a	reaction	to	both	the	inclusion	and	exclusion	manipulations,	being	somewhat	buffered	to	these	conditions.			 65	In	summary,	I	hypothesized	(H5)	that	people	with	a	higher	independent	self-construal	orientation	would	respond	to	the	inclusion	manipulation	by	putting	a	lower	percentage	of	animals	into	a	self-formed	in-group.	On	the	other	hand,	I	predicted	(H3)	that	following	an	exclusion	manipulation,	people	with	a	higher	independent	self-construal	orientation	would	put	a	higher	percentage	of	animals	into	a	self-formed	in-group.	Furthermore,	I	predicted	that	people	with	a	higher	interdependent	self-construal	would	produce	less	of	a	response	to	either	manipulation.		Study	2	also	included	an	anthropomorphism	task	as	a	dependent	variable,	in	order	to	probe	whether	anthropomorphism	varied	with	self-construal	and	the	inclusion	manipulation.	Based	on	past	research,	along	with	the	results	of	Study	1,	I	hypothesised	(H6)	that	under	normal	conditions	(control),	participants	with	high	independent	self-construal	orientation	would	be	more	likely	to	anthropomorphize	animals,	especially	on	factors	relating	to	social	connection,	but	that	following	an	inclusion	manipulation	this	effect	would	reduce,	or	disappear.		I	expected	(H4)	this	effect	to	reverse	following	exclusion,	and	that	individuals	with	higher	independent	self-construal	will	give	higher	ratings	of	anthropomorphism	for	a	non-human	subject.			3.3.2	Procedure.	Study	2	took	the	form	of	a	single	factor	3-level	(control,	inclusion	and	exclusion	manipulation)	mixed	experimental	design,	with	self-construal	measured	as	a		 66	moderator	variable.	The	study	was	administered	in	the	marketing	research	lab	of	a	Canadian	University,	using	120	undergraduate	student	participants	(54%	female,	Mage	=	24),	who	participated	in	the	study	in	exchange	for	course	credit.	The	sample	was	a	convenience	sample	and	condition	was	assigned	randomly,	using	the	randomizing	embedded	coding	system	in	Qualtrics	Survey	Software. The	sample	size	was	based	on	practical	considerations,	regarding	recruitment	and	budgetary	constraints,	with	a	rule	of	thumb	approach	regarding	power,	based	on	previous,\	similar	studies.  	Participants	completed	the	study	at	individual	computer	terminals,	in	groups	of	10-18	per	session,	and	were	told	that	they	would	be	taking	part	in	two	unrelated	studies.		‘Study	One’	was	described	as	a	study	looking	at	personality	characteristics,	in	which	I	was	interested	in	how	general	life	views	and	personal	sense	of	self	related	to	future	life	events.	‘Study	Two’	was	described	as	a	study	looking	at	categorization	behaviour.			After	filling	in	a	number	of	trait	scales,	participants	were	randomly	assigned	to	one	of	three	conditions:	control,	inclusion,	and	exclusion.	Dependent	on	condition	participants	either	received	no	prediction	at	all	(control),	a	future	life	prediction	that	they	would	have	plenty	of	friends	(inclusion),	or	a	future	life	prediction	that	they	would	have	few	friends	(exclusion).	Following	the	future	life	prediction,	participants	were	told	they	would	now	be	starting	the	second	study.				 67	Participants	were	told	the	next	study	was	looking	at	how	people	categorized	information.	The	first	task	they	completed,	was	the	same	grouping	task	as	was	used	in	Study	1.	In	addition,	they	were	asked	to	rate	their	self-formed	in-group,	in	terms	of	entitativity	(group	cohesiveness).	The	prediction	for	this	was	that	people	with	a	higher	independent	self-construal	would	rate	the	self-formed	in-group	higher	in	entitativity,	since	people	with	a	higher	independent	self-construal	not	only	form	new	groups	more	easily,	but	also	derive	meaning	specifically	from	category	based	groups	that	they	have	chosen	to	belong	to.	Conversely,	people	with	a	higher	interdependent	self-construal	consider	themselves	tied	to	long	term	group	structures;	which	makes	it	not	only	harder	to	form	meaningful	new	groups,	but	also	makes	it	more	likely	that	any	meaningful	new	groups	formed,	will	be	based	on	pre-existing	relational	ties,	rather	than	categories.	Therefore,	I	expected	that	people	with	a	higher	interdependent	self-construal	would	feel	less	attached	to	their	new	self-formed	group,	and	would	therefore	rate	it	lower	in	entitativity.	Furthermore,	I	predicted	that	the	manipulation	condition	would	have	an	impact	on	perceptions	of	entitativity	of	the	self-formed	in-group.	Specifically,	following	an	exclusion	manipulation	I	expected	people	to	rate	the	self-formed	in-group	higher	in	entitativity,	as	a	form	of	defensive	behaviour;	whereas	following	a	reassurance	of	chronic	inclusion,	a	self-formed	in-group	would	not	seem	so	important	to	personal	happiness,	and	would	be	evaluated	lower	in	entitativity.			Participants	then	completed	an	anthropomorphism	photo	task,	designed	to	measure	how	much	they	were	anthropomorphizing,	especially	in	the	social	realm.	Following		 68	this,	participants	responded	to	a	manipulation	check,	a	hypothesis	probe	and	some	demographic	questions.	Participants	were	then	debriefed,	offered	candy	as	a	mood	elevator,	and	thanked.				Materials.	Participants	were	asked	to	complete	the	same	24	item	self-construal	scale	(12	items	independent	self-construal,	12	items	interdependent	self-construal)	as	used	in	study	1.	Participants	were	also	asked	to	complete	the	same	amended	Individual	Differences	in	Anthropomorphism	Questionnaire	(IDAQ;	α	=.872	)	previously	used	by	Waytz,	Cacioppo,	and	Epley,	(2014)	to	monitor	attitudes	to	animals,	and	the	same	12	item	Fear	of	Negative	Evaluation	(FNE;	α	=.774)	scale	(Leary,	1983).	In	addition	a	10-item	Need	To	Belong	(NTB;	α	=.665)	scale	(Leary,	Kelly,	&	Schreindorfer,	2001)	was	administered	(see	appendix	A.5).	These	personality	scales,	taken	all	together,	served	to	support	the	cover	story;	namely	that	I	was	interested	in	how	personality	traits	influenced	future	life	behaviour.	The	specific	scales	were	selected	based	on	past	findings,	demonstrating	their	potential	to	interact	with	Human	Animal	Interactions	(HAI),	social	exclusion	threats,	and	intergroup	evaluations	and	behaviour.	The	NTB	scale	was	developed	to	measure	need-to-belong,	the	desire	to	form	meaningful	social	attachments.	Need-to-belong	is	arguably	a	fundamental	human	motivation,	and	is	considered	to	conform	to	motivational	patterns	of	satiation	and	substitutions	(Baumeister	&	Leary	1995).			Prior	to	receiving	one	of	the	three	experimental	conditions,	participants	also		 69	completed	a	3-Item	Loneliness	Scale	(Hughes,	Waite,	Hawkley,	&	Cacioppo,	2004;	α	=.863).	This	was	administered	in	study	2	in	accordance	with	Behavioural	Research	Ethics	Board	requirements,	as	a	precautionary	screening	test	for	vulnerable	participants,	who	might	experience	a	more	pronounced	reaction	to	the	Future	Life	Prediction.			The	Future	Life	Prediction	manipulation	has	been	successfully	used	on	many	occasions	(Epley,	Akalis,	Waytz	and	Cacioppo,	2008	and	Twenge,	Baumeister,	DeWall,	Ciarocco	and	Bartels,	2007),	to	evoke	feelings	of	loneliness/	exclusion	or	feelings	of	belongingness/	inclusion	in	participants,	and	takes	one	of	three	forms	(exclusion,	inclusion,	and	control).		With	the	exclusion	manipulation,	participants	were	told	that	their	answers	to	the	prior	questions	had	enabled	a	‘Future	Life	Prediction’	to	be	constructed,	which	predicted	that	they	would	lack	long-term	friendships	and	relationships	and	be	lonely	through	much	of	their	life.		With	the	inclusion	manipulation,	participants	were	told	that	their	answers	to	the	previous	questions	had	enabled	a	‘Future	Life	Prediction’	to	be	constructed,	which	informed	them	that	they	would	always	have	friends	and	loved	ones	in	their	life.	In	the	control	condition,	participants	were	not	told	they	would	be	given	a	future	life	prediction,	and	did	not	receive	one.	See	appendix	A.6	for	the	full	wording	of	the	‘Future	Life	Prediction’.	Participants	were	then	moved	onto	what	they	were	told	was	the	second	study.		In	this	next	section,	participants	were	asked	to	complete	the	same	grouping		 70	selection	that	was	completed	in	study	1,	and	to	rate	the	self	created	in-group	on	a	5-item	entitativity	scale	developed	for	the	study	(α=.718).	This	measure	asked	participants	to	respond	to	a	series	of	statements,	as	follows	“I	am	proud	to	think	of	myself	as	a	member	of	this	group”,	“this	group	is	meaningful	and	cohesive”,	“this	group	is	important	to	its	members”,	“members	of	this	group	are	similar”,	and	“members	of	this	group	share	common	goals”,	on	a	scale	of	1-7;	where	1	=	not	at	all,	and	7	=	extremely.	Following	this	task,	participants	were	asked	to	rate	a	photograph	of	a	black	Labrador	dog	on	a	variety	of	factors	relating	to	anthropomorphism,	and	especially	social	anthropomorphism	(α=.785);	using	an	amended	version	of	a	scale	originated	and	previously	used	by	Epley,	Waytz,	and	Cacioppo	(2007)	and	Epley,	Akalis,	Waytz	and	Cacioppo	(2008).	Social	anthropomorphism	has	been	previously	identified	by	Epley,	Waytz,	and	Cacioppo	(2007)	as	a	tendency	to	anthropomorphize	based	on	specific	values	that	may	offer	potential	for	social	connection,	which,	for	example,	are	expressed	in	the	statement	“This	dog	can	be	sympathetic”.	This	may	be	differentiated	from	more	general	anthropomorphism	values,	such	as	can	be	seen	expressed	in	the	statement	“This	dog	can	suffer	from	embarrassment	at	times”.			The	black	Labrador	dog	was	photographed	in	a	lying	position,	looking	towards	the	camera.		See	appendix	A.7,	for	both	the	photograph	and	the	amended	anthropomorphism	scale.	Following	this,	demographics	were	collected,	along	with	a	manipulation	check	asking	them	to	recall	which	future	life	prediction	they	received,	and	to	rate	their	mood	following	the	prediction,	using	an	8-item	scale	(positive:	happy,	positive,	content,	supported;	negative:	left	out,	isolated,	negative,	lonely).	At		 71	the	end,	participants	were	asked	a	hypothesis	probe	question	to	check	hypothesis	guessing.	Participants	were	then	debriefed	and	thanked.		3.3.3	Results.	Two	participants,	rated	moderate	on	the	3-Item	Loneliness	Scale	(Hughes,	Waite,	Hawkley,	&	Cacioppo,	2004),	and	were	categorized	as	vulnerable	in	accordance	with	Behavioural	Research	Ethics	Board	requirements.	They	were	therefore	screened	out	of	the	experimental	study,	before	administration	of	the	manipulation,	as	a	precautionary	measure	(leaving	n=118	remaining).	The	study	exclusion	policy	stipulated	that	participants	would	be	screened	out	if	they	failed	all	attention	checks,	or	if	they	accurately	guessed	the	key	study	hypotheses;	which	no	participants	did.			Power.	A	post	hoc	power	analysis	was	conducted,	using	G*Power	3.1.	With	a	sample	size	of	81	in	a	multiple	regression	model,	with	3	predictors	(2	predictors	and	1	interaction	term)	and	an	observed	R2	of	0.12,	ρ2	was	calculated	at	0.127,	which	provided	a	power	estimation	of	0.71,	at	95%	confidence	level.		A	post	hoc	power	test	conducted,	using	G*Power	3.1,	for	an	independent	sample	t-test	revealed	that	with	a	sample	size	of	80	(40	per	condition)	an	observed	d=0.65,	with	critical	t	of 1.991,	was	needed	for	a	power	estimation	of	>0.80	at	95%	confidence	level.	Furthermore,	a	post	hoc	power	test	for	correlations	revealed	that	with	a	sample	size	of	118,	an	observed	correlation	of	.26,	was	needed	for	a	power	estimation	of	>0.80	at	95%	confidence	level.		 72	Manipulation	check.		To	assess	whether	participants	responded	as	expected	to	the	manipulation,	they	were	asked	to	recall	the	prediction	they	received	in	a	multiple-choice	question	at	the	end	of	the	study,	after	demographics	were	taken.	They	were	also	asked	to	answer	some	post	prediction	mood,	and	manipulation	probe	questions.	When	asked	to	recall	their	prediction,	89.4%	participants	were	able	to	do	so	with	accuracy,	3.3%	inaccurately	remembered	getting	a	future	belonging	(inclusion)	prediction	when	in	fact	they	were	in	the	control	group	(no	prediction),	and	7.3%	could	not	remember	their	prediction.				Post	manipulation	measurements	of	positive	mood	were	also	tested	across	conditions	(inclusion,	exclusion,	control),	at	the	end	of	the	study	following	the	manipulation	probe.	A	one-way	between	groups	ANOVA	revealed	a	significant	main	effect	of	condition	on	post	prediction	mood,	F(2,116)=36.57,	p<.001,	η=.39.	Post	hoc	comparisons,	using	Tukey’s	HSD	test,	indicated	that	the	post	prediction	mood	scores	in	the	exclusion	condition	(M=	3.58,	SD=1.31)	were	significantly	lower	than	the	post	prediction	mood	scores	in	both	the	control	condition	(M=	4.69,	SD=1.26),	(p<.001),	95%	CI	[-1.74,-0.48]	and	the	inclusion	condition	(M=	5.77,	SD=0.85),	(p<.001),	95%	CI	[-2.80,-1.58].	Furthermore,	post	hoc	comparisons	indicated	that	that	the	post	prediction	mood	scores	in	the	inclusion	condition	were	significantly	higher,	than	the	post	prediction	mood	scores	in	the	control	condition	(p<.001),	95%	CI	[0.49,1.68].	It	must	be	noted,	however,	that	a	Levene’s	test	of	equality	of	error	variances	was		 73	significant	F(2,116)=4.36,	p=.02	and	therefore	the	assumption	of	homogeneity	of	variance	was	violated	in	the	test.			Demographics.	Demographics	collected	showed	that	participants	in	the	sample	identified	as	82.1%	Asian,	6.3%	Caucasian,	3.6%	East	Indian,	1.8%	Black,	and	6.3%	other.	Results	also	showed	that	5.7%	reported	living	in	Canada	less	than	2	years,	17.1%	reported	living	in	Canada	2-5	years,	19.5%	reported	living	in	Canada	6-10	years,	30.9%	reported	living	in	Canada	11-20	years	and	26.8%	reported	living	in	Canada	21	years	or	over.	Participants	identified	as	53.9%	female.	The	mean	age	was	24.1	years	old.	See	appendix	A.8	for	detailed	demographic	and	trait	analyses.			Means	comparisons.		Independent	samples	t-tests	were	carried	out	between	the	control	group	and	respective	conditions	(exclusion;	inclusion),	as	separate	analyses.	Tests	showed	no	significant	differences	between	conditions,	on	the	either	the	grouping	task,	or	the	anthropomorphism	task.	See	appendix	A.8	for	detailed	results.			Regression	tests	for	moderation.	Moderated	multiple	regressions	were	carried	out	in	SPSS,	using	process	v.2	(Hayes,	2013)	to	examine	a)	whether	the	manipulation	(inclusion	vs.	exclusion	condition)	influenced	participant	intentions	and	evaluations,	including	the	percentage	of	animals	that	were	included	in	a	self-formed	in-group	and	b)	whether	these	effects		 74	were	moderated	by	independent	and	interdependent	self-construal	orientation.	Following	Cohen	and	Cohen	(1983)	and	Wendorf	(2004),	I	dummy	coded	the	condition	so	that	the	inclusion	manipulation	(vs.	control),	and	the	exclusion	manipulation	(vs.	control),	were	run	as	separate	tests,	and	the	results	reported	independently.	Independent	self-construal	and	interdependent	self-construal	scales	were	kept	separate	for	the	analysis,	but	were	mean	centered	in	accordance	with	guidelines,	stipulating	that	this	practice	renders	subsequent	tests	of	hypotheses,	and	regression	coefficients	for	X	and	M,	more	meaningful	and	substantially	interpretable,	as	well	as	to	reduce	the	likelihood	of	errors	in	interpretation	(Hayes,	2013).	All	dependent	measures	were	subjected	to	separate	moderated	multiple	regressions,	with	each	manipulation	condition	and	the	control,	independent	and	interdependent	self-construal	scores,	and	their	interactions,	simultaneously	entered	as	predictor	variables.			3.3.3.1	Role	of	exclusion	manipulation	in	grouping,	and	anthropomorphism	tasks.	Summary.	Against	expectations,	the	exclusion	manipulation	appeared	to	have	no	significant	impact	on	either	of	the	dependent	variables	(grouping	task	and	anthropomorphism	task),	in	terms	of	a	main	effect.	This	was	despite	appearing	to	be	successful	as	a	manipulation,	and	despite	being	accurately	recalled	by	participants.	Furthermore,	a	regression	model	showed	no	significant	interaction	of	self-construal	(independent;		 75	interdependent)	with	the	exclusion	condition.	See	detailed	analyses	below,	and	appendix	A.8	for	covariate	analyses.			Grouping	task	&	independent	self-construal.	The	percentage	of	animals	included	in	a	self-formed	in-group	was	regressed	on	to	the	dummy	(exclusion	condition)	and	moderator	variables.		The	social	exclusion	manipulation,	independent	self-construal,	and	the	interaction	together	were	not	found	to	account	for	any	significant	proportion	of	variability	in	the	inclusion	of	animals	into	an	in-group	in	the	overall	model,	R2=.07,	F(3,70)=1.85,	p=.146.		Independent	self-construal	did	not	significantly	predict	inclusion	of	animals	B=0.006,	SE=0.005,	t(74)=1.318,	p=.192,	95%	CI	[-0.003,0.015]	nor	did	the	social	exclusion	manipulation,	B=-0.099,	SE	=0.053,	t(74)=-1.873,	p=.065,	95%	CI	[-0.20,0.01],	although	it	appeared	to	result	in	a	very	minor	reduction.	In	sum	however	there	was	no	conditional	effect	found	for	either	the	manipulation	or	self-construal.		Neither	was	a	significant	two-way	interaction	between	the	social	exclusion	condition	(control	vs.	exclusion),	and	independent	self-construal	found	B=-0.005,	SE	=	0.007,	t(74)=-0.756,	p=	.452,	95%	CI	[-0.02,0.01].				Grouping	task	&	interdependent	self-construal.	The	percentage	of	animals	included	in	a	self-formed	in-group	was	regressed	on	to	the	dummy	(exclusion	condition)	and	moderator	variables.		The	social	exclusion	manipulation,	interdependent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	a	significant	proportion	of	variability	in	the		 76	inclusion	of	animals	into	an	in-group	in	the	overall	model,	R2=.117,	F(3,	0)=3.09,	p=.032.		Interdependent	self-construal	significantly	predicted	less	inclusion	of	animals	B=-0.008,	SE=0.004,	t(74)=-2.311,	p=.024,	95%	CI	[-0.02,-0.001],	but	the	social	exclusion	manipulation	did	not,	B=-0.095,	SE	=	0.051,	t(74)=-1.891,	p=.067,	95%	CI	[-0.20,0.01]	.	In	sum	there	was	no	conditional	effect	found	for	the	manipulation.		Neither	was	a	significant	two-way	interaction	between	the	social	exclusion	condition	(control	vs.	exclusion),	and	interdependent	self-construal	orientation	found	B=0.008,	SE	=	0.006,	t(74)=1.49,	p=	.141,	95%	CI	[-0.003,0.02].				Anthropomorphism	of	dog	picture	&	independent	self-construal.	Ratings	of	social	anthropomorphism	were	regressed	on	to	the	dummy	and	moderator	variables	for	the	dog	picture.	For	the	picture	the	social	exclusion	manipulation,	independent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	for	a	significant	proportion	of	variability	in	ratings	of	social	anthropomorphism	in	the	overall	model,	R2=.15,	F(3,70)=3.98,	p=.011.	Independent	self-construal	significantly	predicted	ratings,	B=-0.090,	SE	=0	.040,	t(74)=2.502,	p=.02,	95%	CI	[0.02,0.16],	but	the	social	exclusion	manipulation	did	not,	B=-0.560,	SE	=	0.410,	t(74)=-1.36,	p=.180,	95%	CI	[-1.38,0.26]	.	There	was	no	significant	two-way	interaction	between	the	social	exclusion	condition	(control	vs.	exclusion),	and	independent	self-construal	orientation,	B=-0.028,	SE	=	0.05,	t(74)=-0.542,	p=.59,	95%	CI	[-0.13,0.07].				 77	Anthropomorphism	of	dog	picture	&	interdependent	self-construal.	Ratings	of	social	anthropomorphism	were	regressed	on	to	the	dummy	and	moderator	variables	for	the	dog	picture.	For	the	picture	the	social	exclusion	manipulation,	interdependent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	for	a	significant	proportion	of	variability	in	ratings	of	social	anthropomorphism	in	the	overall	model,	R2=.11,	F(3,70)=2.96,	p=.038.	Interdependent	self-construal	did	not	significantly	predict	ratings,	B=0.001,	SE	=	0.030,	t(74)=0.042,	p=.97,	95%	CI	[-0.06,0.06],	neither	did	the	social	exclusion	manipulation,	B=-0.570,	SE	=	0.420,	t(74)=-1.36,	p=.18,	95%	CI	[-1.40,0.26].	There	was	no	significant	two-way	interaction	between	the	social	exclusion	condition	(control	vs.	exclusion),	and	interdependent	self-construal	orientation,	B=0.088,	SE	=	0.050,	t(74)=1.89,	p=.06,	95%	CI	[-0.01,0.18].		Controlling	for	gender,	age	and	ethnicity	in	the	above	analyses	revealed	no	significant	results,	as	detailed	in	the	appendix	A.8.	Controlling	for	the	trait	measures	(FNE,	IDAQ,	NTB,	short	loneliness	scale)	also	revealed	no	significant	findings.	See	appendix	A.8	for	detailed	results.	In	sum,	study	2	failed	to	support	the	hypothesis	(H3),	that	following	an	exclusion	manipulation	people	with	a	higher	independent	self-construal	orientation	will	put	a	higher	percentage	of	animals	into	a	self-formed	in-group.	Furthermore	study	2	failed	to	support	the	hypothesis	(H4),	that	following	an	exclusion	manipulation	people	with	a	higher	independent	self-construal	orientation	will	anthropomorphize	more.	Subsequently,	I	focused	the	remainder	on	my	analysis	on	the	comparison	between	the	inclusion	and	the	control	condition.		 78	3.3.3.2	Role	of	inclusion	manipulation	in	grouping	and	anthropomorphism	tasks.	Grouping	task	&	independent	self-construal.	The	percentage	of	animals	included	in	a	self-formed	in-group	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	social	inclusion	manipulation,	independent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	for	a	significant	proportion	of	variability	in	the	inclusion	of	animals	into	an	in-group	in	the	overall	model,	R2=.12,	F(3,77)=3.43,	p=.02.	While	independent	self-construal	alone	did	not	significantly	predict	inclusion	of	animals,	the	social	inclusion	manipulation	did,	B=-0.110,	SE	=	0.050,	t(81)=-2.253,	p=.03,	95%	CI	[-0.21,-0.01]	.	Importantly	a	significant	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	independent	self-construal	was	found	B=-0.010,	SE	=	0.010,	t(81)=-2.24,	p=	.03,	95%	CI	[-0.03,-.002].	2			Grouping	task	&	interdependent	self-construal.	The	percentage	of	animals	included	in	a	self-formed	in-group	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.	The	social	inclusion	manipulation,	interdependent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	a	significant	proportion	of	variability	in	the	inclusion	of	animals	into	an	in-group	in	the	overall	model,	R2=.190,	F(3,77)=6.00,	p=.001.		Interdependent	self-construal	significantly	predicted	inclusion	of	animals																																																									2	Note	I	report	the	unstandardized	coefficient	(B),	rather	than	betas.		Because	betas	are	not	properly	standardized	in	interaction	terms	they	are	not	interpretable,	whereas	B	represents	the	difference	between	the	un-weighted	means	of	the	groups	involved	(see	Cohen,	Cohen,	West,		&	Aiken,	2003).		 79	B=-.008,	SE=.004,	t(81)=-2.333,	p=.02,	95%	CI	[-0.02,-0.001],	as	did	the	social	inclusion	manipulation,	B=-0.114,	SE	=	0.048,	t(81)=-2.364,	p=.021,	95%	CI	[-0.21,-0.02]	.	There	was	also	a	significant	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	interdependent	self-construal	B=0.019,	SE	=	0.005,	t(81)=3.55,	p=	.001,	95%	CI	[-0.01,-0.03].				Controlling	for	demographics	(ethnicity,	age,	gender)	revealed	no	significant	findings.	Controlling	for	the	trait	measures	(FNE,	IDAQ,	NTB,	short	loneliness	scale)	also	revealed	no	significant	findings	(see	appendix	A.8	for	results).		To	both	probe	and	visualize	the	nature	of	the	interaction,	a	floodlight	analysis	(Spiller	et	al,	2013),	also	known	as	the	Johnson-Neyman	(JN)	technique,	was	performed	using	Process	(Hayes,	2013).	A	floodlight	analysis	was	chosen,	over	a	spotlight	or	simple	slopes	analysis	(Rogosa,	1980;	Bauer	&	Curran,	2005),	to	avoid	the	necessity	of	making	an	arbitrary	choice	for	values	of	M,	and	to	allow	for	the	results	to	be	of	use	beyond	sample	specific	circumstances	(Hayes,	2013,	p.	239).			As	illustrated	in	figure	11,	the	analysis	showed	that	that	under	normal	circumstances	(control	condition),	participants	that	measured	higher	in	independent	self-construal	appeared	to	include	significantly	more	animals	in	a	self	designated	in-group,	than	participants	measuring	higher	in	interdependent	self-construal.	However,	for	participants	that	received	an	inclusion	manipulation,	prior	to	the	grouping	task,	telling	them	that	a	future	life	prediction	suggested	they	would	always	be	strongly	included	by	other	humans,	this	pattern	significantly	reversed.	In		 80	the	inclusion	condition,	participants	showing	a	higher	independent	self-construal	orientation	were	now	less	likely	to	include	animals	in	their	self-designated	in-group,	than	were	participants	measuring	higher	in	interdependent	self-construal.		See	figure	11,	for	the	regression	lines,	with	a	visualization	of	the	regions	of	significance,	as	derived	from	the	Johnson-Neyman	technique,	as	well	as	figure	12,	for	a	floodlight	analysis	for	each	interaction,	as	a	plot	of	0XàY	as	a	function	of	M	along	with	confidence	bands	(Bauer	&	Curran,	2005;	Preacher,	Curran,	&	Bauer,	2006;	Rogosa,	1980;	Spiller	et	al.	2013).		Entitativity	&	independent	self-construal.	Participants	were	also	asked	to	rate	the	entitativity	of	their	self-formed	in-group,	on	a	number	of	items,	following	the	grouping	task.	The	entitativity	score	was	regressed	on	to	the	dummy	and	moderator	variables.		The	social	inclusion	manipulation,	independent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	for	a	significant	proportion	of	variability	in	the	entitativity	rating	of	the	in-group	in	the	overall	model,	R2=.12,	F(3,77)=3.47,	p=.02.	While	the	social	inclusion	manipulation	alone	did	not	significantly	predict	the	entitativity	rating	of	the	in-group,	independent	self-construal	did,	B=0.390,	SE	=	0.130,	t(81)=3.058,	p=.003,	95%	CI	[0.14,	0.64]	.	Furthermore	a	marginal	trend	was	found,	albeit	not	significant,	for	the	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	independent	self-construal	orientation	B=-0.280,	SE	=	0.170,	t(81)=-1.679,	p=	.10,	95%	CI	[-0.62,0.05],	which	shows	that	while	normally	(control)	there	appears	to	be	a	strong	positive	correlation	between	independent	self-construal		 81	orientation	and	subsequent	rating	of	the	entitativity	of	a	self-formed	in-group,	this	relationship	disappears	with	an	inclusion	condition.			Entitativity	&	interdependent	self-construal.	Regression	analysis	revealed	no	significant	results,	in	any	variables	when	interdependent	self-construal	was	investigated.	See	appendix	A.8	for	detailed	results.			See	figure	13	for	a	visual	comparison	of	the	trends	in	entitativity	ratings,	for	both	independent	and	interdependent	self-construal.		Anthropomorphism	of	dog	picture.	Ratings	of	social	anthropomorphism	were	regressed	on	to	the	dummy	and	moderator	variables,	for	the	dog	picture.	For	the	picture,	the	social	inclusion	manipulation,	independent	self-construal	orientation,	and	the	interaction	together	were	found	to	account	for	a	significant	proportion	of	variability	in	ratings	of	social	anthropomorphism	in	the	overall	model,	R2=.130,	F(3,77)=3.72,	p=.015.	Independent	self-construal	significantly	predicted	ratings,	B=-0.090,	SE	=	0.040,	t(81)=2.39,	p=.02,	95%	CI	[0.02,0.17],	as	did	the	social	inclusion	manipulation,	B=-0.867,	SE	=	0.410,	t(81)=-2.14,	p=.04,	95%	CI	[-1.66,-0.06]	.	There	was	no	significant	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	independent	self-construal	orientation,	B=-0.061,	SE	=	0.050,	t(81)=-1.22,	p=.23,	95%	CI	[-0.16,0.04].		 82	While	there	was	no	consistent	significant	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	independent	self-construal	orientation,	there	was	significance	in	the	area	of	high	independent	self-construal.	I	therefore	opted	to	visualize	the	interaction	and	the	region	of	significance	using	the	Johnson-Neyman	technique.	A	floodlight	analysis	for	the	interaction,	as	a	plot	of	0XàY	as	a	function	of	W,	along	with	confidence	bands,	offered	some	tentative	support	for	a	hypothesis	that,	under	normal	conditions,	participants	with	high	independent	self-construal	are	more	prone	to	anthropomorphizing	animals	on	factors	specifically	relating	to	social	connection,	but	that	following	an	inclusion	manipulation,	this	effect	reduces.	Note	however	the	lack	of	statistical	significance	for	the	overall	model,	requires	caution	to	be	taken	over	any	statistical	inferences	that	may	be	made	from	the	result	(Hayes,	2013).	See	figure	14	for	a	visual	representation	of	the	results,	and	figure	15	for	a	floodlight	analysis	of	the	interaction,	as	a	plot	of	0XàY	as	a	function	of	W	along	with	confidence	bands	including	the	region	of	significance.			Anthropomorphism	of	dog	picture	&	interdependent	self-construal.	Ratings	of	social	anthropomorphism	were	regressed	on	to	the	dummy	and	moderator	variables	for	the	dog	picture.	For	the	picture	the	social	inclusion	manipulation,	interdependent	self-construal	orientation,	and	the	interaction	together	were	not	found	to	account	for	a	significant	proportion	of	variability	in	ratings	of	social	anthropomorphism	in	the	overall	model,	R2=.07,	F(3,77)=1.98,	p=.12.	Interdependent	self-construal	did	not	significantly	predict	ratings,	B=0.001,	SE	=	0.030,	t(81)=0.04,	p=.97,	95%	CI	[-0.06,0.06],	but	the	social	inclusion		 83	manipulation	did,	B=-0.910,	SE	=	0.420,	t(81)=-2.17,	p=.03,	95%	CI	[-1.75,-0.07]	.	There	was	no	significant	two-way	interaction	between	the	social	inclusion	condition	(control	vs.	inclusion),	and	interdependent	self-construal,	B=0.040,	SE	=	0.050,	t(81)=0.88,	p=.38,	95%	CI	[-0.05,0.13].		3.3.4	Discussion.	Study	2	replicated	the	findings	of	study	1,	and	provided	an	extension	to	the	findings	when	participants	were	manipulated	to	feel	social	inclusion.	Specifically,	in	study	2	I	was	able	to	provide	further	support	for	H1;	that	when	participants	were	asked	to	make	in-groups	and	out-groups	under	normal	conditions	(control),	participants	with	higher	independent	self-construal	appeared	to	put	a	relatively	high	percentage	of	animals	into	their	in-group.			Study	2	did	not	provide	explicit	support	for	my	prediction	that	participants	measuring	high	in	interdependent	self-construal	would	react	less	to	both	the	inclusion	and	exclusion	manipulations,	being	somewhat	buffered	to	these	conditions.	However,	study	2	did	provide	evidence	that	participants	who	measured	high	in	interdependent	self-construal,	would	react	in	a	different	direction	to	those	high	in	independent	self-construal.	Furthermore,	it	is	perhaps	indicative	that	the	area	of	significance	in	the	regression	model	with	interdependent	self-construal	was	in	the	low	regions,	rather	than	the	higher	regions;	in	other	words	the	result	was	seen	for	those	low	in	interdependent	self-construal.	This	finding	suggests	that	the		 84	pattern	was	being	driven	primarily	by	high	independent	self-construal,	which	is	in	line	with	previous	research	and	the	predictions.				The	ratings	of	entitativity	for	the	self-formed	in-group	showed	a	(non-significant)	trend	for	higher	ratings	with	people	who	were	higher	in	independent	self-construal.		This	effect	disappeared	with	an	inclusion	manipulation.	Although	not	significant,	this	pattern	may	be	compared	with	the	results	measuring	high	interdependent	self-construal,	where	there	appeared	to	be	no	association	between	entitativity	and	interdependent	self-construal.	This	result	provides	some	very	tentative	support	for	an	argument	that	under	normal	conditions	individuals	with	high	independent	self-construal	may	rate	a	self-formed	in-group	as	more	cohesive	and	perhaps	more	meaningful,	than	they	do	after	an	affirmation	of	inclusion.			In	terms	of	the	social	exclusion	condition,	study	2	did	not	support	hypothesis	H3,	which	predicted	that	following	an	exclusion	manipulation	people	with	higher	independent	self-construal	would	put	a	higher	percentage	of	animals	into	a	self-formed	in-group	than	in	the	control.	Furthermore,	it	did	not	offer	support	for	H4;	that	people	with	higher	independent	self-construal	will	give	higher	ratings	of	anthropomorphism	for	a	non-human	subject	following	an	exclusion	manipulation	than	they	will	normally	(control	condition).	I	have	no	conclusive	explanation	to	offer	for	this	result,	since	I	could	find	no	relationship	with	any	of	my	measured	variables	thought	to	moderate	exclusion	threat	responses,	such	as	fear	of	negative	evaluation	(Maner,	De	Wall,	Baumeister,	&	Schaller,	2007).	I	can	at	this	stage	only	speculate		 85	regarding	other	reasons	for	the	null	result.	Perhaps	the	younger	age	of	my	participants	may	have	acted	as	a	boundary	condition,	or	perhaps	my	chosen	dependent	variables	failed	to	draw	out	a	response	from	excluded	participants.	For	example,	perhaps	my	participants	did	not	perceive	an	opportunity	to	reconnect	would	be	manifested	in	the	outcome	variables.	As	previously	detailed	by	Maner,	De	Wall,	Baumeister,	and	Schaller	(2007)	exclusion	manipulations	can	at	times	deliver	complex	results.			In	terms	of	the	social	inclusion	manipulation,	the	results	of	study	2	provided	substantial	evidence	for	H5,	that	following	an	inclusion	manipulation	people	with	higher	independent	self-construal	will	put	a	lower	percentage	of	animals	into	a	self-formed	in-group	than	normally	(control	condition).	I	found	a	significant	two-way	interaction	between	the	social	connection	condition	(control	vs.	social	inclusion),	and	measured	self-construal	(high-low	levels	of	independent	and	interdependent	self-construal).		Repeating	the	findings	of	study	1,	this	demonstrates	that,	under	normal	circumstances	(control	condition),	participants	that	measure	high	in	independent	self-construal	are	more	likely	to	include	more	animals	in	a	self-formed	in-group.	However,	if	participants	receive	an	inclusion	manipulation	prior	to	the	grouping	task,	telling	them	that	a	future	life	prediction	suggests	that	they	will	always	be	strongly	included	by	other	humans,	this	pattern	significantly	reverses.	In	this	social	inclusion	condition,	participants	showing	higher	independent	self-construal	were	significantly	less	likely	to	include	animals	in	their	self-designated	in-group	than	in	the	control	condition.			 86	This	finding	has	interesting	theoretical	implications,	in	that	it	offers	an	original	new	explanation	for	what	may	be	driving	the	effects	of	generosity	and	inclusivity	towards	distant	others,	that	are	seen	in	people	with	high	independent	self-construal.		Inclusivity	towards	distant	others	was	lower	in	study	2,	for	people	with	high	independent	self-construal,	that	received	an	inclusion	manipulation,	in	which	they	were	assured	of	their	chronic	future	belonging.		Study	2,	therefore,	offers	some	support	for	the	argument,	that	the	desire	for	belonging	or	social	inclusion	may	be	the	motivation	fuelling	more	inclusive	and	generous	behaviour	towards	distant	others	in	this	group.		Regarding	the	anthropomorphism	findings,	although	the	interaction	between	inclusion	measure	and	independent	self-construal	was	not	significant	across	the	entire	range,	I	believe	the	fact	that	the	floodlight	analysis	showed	a	large	region	of	significance	in	the	medium	to	upper	region	of	independent	self-construal,	offers	some	tentative	evidence	in	support	of	H6;	that	people	with	higher	independent	self-construal	will	give	lower	ratings	of	anthropomorphism	for	a	non-human	subject,	following	an	inclusion	manipulation,	than	they	will	normally	(control	condition).	Furthermore,	it	offers	some	support	for	the	hypothesis	that	increased	anthropomorphism	may	be	related	to	a	desire	for	social	connection,	in	those	higher	in	independent	self-construal.			The	motive	for	an	increase	in	anthropomorphism	behaviour	is	that,	by	adopting	a	hyper	vigilant	approach	to	cues	of	social	connection	potential,	and	by	relaxing		 87	boundaries	of	what	is	considered	capable	of	meeting	those	needs,	an	individual	may	maximise	social	connection	possibilities	in	situations	when	humans	are	either	not	desirable	or	not	available.	This	finding	is	in	line	with	the	social	monitoring	system	(SMS)	concept,	the	primary	purpose	of	which	is	to	attune	individuals	to	information	that	will	help	them	navigate	the	social	environment	more	successfully,	whatever	the	belonging	state	they	find	themselves	in	(Gardner,	Pickett	and	Knowles,	2005).		Furthermore,	by	visualizing	the	interaction	and	the	region	of	significance,	using	the	Johnson-Neyman	technique,	I	show	that	under	normal	conditions	participants	with	high	independent	self-construal	are	more	prone	to	anthropomorphizing	animals	on	factors	specifically	relating	to	social	connection,	but	that	following	an	inclusion	manipulation	this	effect	reduces.	Whilst	not	a	conclusive	finding,	the	result	offers	support	for	the	findings	of	past	research	demonstrating	that	anthropomorphism	is	often	highest	in	people	that	lack	human	social	contact	(Epley,	Akalis,	Waytz	&	Cacioppo,	2008).	The	fact	that	no	significant	effect	was	found	in	participants	with	a	high	interdependent	self-construal	is	unsurprising,	since	we	would	expect	participants	with	a	more	interdependently	orientated	self-construal	to	be	less	open	to	using	animals	as	parasocials,	as	well	as	less	open	to	re-categorizing	others	for	the	purposes	of	social	connection,	for	reasons	previously	outlined.		In	summary,	the	results	of	study	2	follow	in	the	direction	of	prior	research,	demonstrating	that	social	inclusion	may	indeed	negatively	impact	inclusive	attitudes	towards	distant	others	and	out-group	members	(DeWall,	Baumeister	&		 88	Vohs,	2008;	Waytz	&	Epley,	2007;	2012).	The	results	of	study	2,	when	taken	together,	also	offer	some	support	for	the	argument	that	animals	may,	at	times,	be	used	to	play	a	compensatory	role	for	lack	of	human-to-human	relations,	but	that	when	human-to-human	relations	are	assured,	and	this	compensatory	role	is	no	longer	required,	then	the	use	of	animals	as	parasocial	support	mechanisms	may	reduce.		This	re-prioritization	approach	builds	on	past	research	of	Brown,	Young,	Sacco,	Bernstein	and	Claypool	(2009),	who	put	forward	a	reprioritization	hypothesis,	which	proposed	that	following	social	acceptance;	a	proximal	cue	that	indicates	group	based	survival	needs	have	been	satisfied,	individuals	may	shift	their	motives	towards	other	activities	and	interests.			That	the	pattern	of	response	to	an	inclusion	manipulation;	as	described	above,	is	more	apparent	in	people	with	an	independent	self-construal	orientation,	also	builds	on	past	research	by	Pfundmair,	Graupmann,	Frey	and	Aydin	(2015),	as	well	as	by	Ren,	Wesselmann	and	Williams	(2013).	Both	of	these	research	groups	have	offered	up	evidence	that	people	with	a	more	interdependent	self-construal	orientation	may	not	be	as	susceptible	to	promises	of,	or	threats	to,	social	inclusion.	Moreover,	the	results	of	study	2	offer	further	support	for	the	argument	of	Yuki	and	Takemura	(2013)	that	people	with	a	more	independent	self-construal	orientation	should	regard	in-group	boundaries	as	more	permeable,	than	people	with	a	more	interdependent	orientation.				 89	Finally,	study	2	offers	up	some	evidence	to	support	and	potentially	build	on	the	work	of	Epley,	Akalis,	Waytz	&	Cacioppo	(2008);	that	people	will	anthropomorphize	more	when	they	are	in	need	of	social	connection,	by	demonstrating	a	trend	for	this	behaviour	that	is	specific	to	people	high	in	independent	self-construal,	and	not	for	people	high	in	interdependent	self-construal.	This	builds	on	the	results	of	study	1,	which	found	that	there	was	a	significant	positive	correlation	between	trait	anthropomorphism	(measured	in	the	IDAQ)	and	independent	self-construal.	Taken	together,	the	results	of	study	1	and	study	2	further	support	and	build	upon	the	past	research	of	Epley,	Akalis,	Waytz	&	Cacioppo	(2008),	by	providing	evidence	that	following	a	reassurance	of	chronic	social	inclusion,	anthropomorphism	ratings	specifically	reduce	for	people	high	in	independent	self-construal,	offering	support	for	the	notion	that	anthropomorphism	may	be	a	mechanism	that	has	evolved	to	facilitate	a	sense	of	social	inclusion	in	situations	where	human	company	may	not	be	available;	also	that	this	is	more	of	a	threat	for	people	high	in	independent	self-construal,	than	for	people	high	in	interdependent	self-construal.			Study	1	and	2	used	an	inclusion-grouping	task	as	the	main	dependent	variable.	However	inclusion	of	out-group	members	into	an	in-group,	whilst	demonstrating	a	high	level	of	inclusive	behaviour,	does	not	demonstrate	some	of	the	more	tangible	prosocial	behaviours	that	may	be	directed	towards	out-group	members,	such	as	charitable	giving.	In	study	3,	therefore,	I	wished	to	investigate	whether	the	behaviour	observed	in	study	1	and	2	might	extend	into	something	with	financial	benefits,	such	as	donation	intentions	for	an	out-group	charity.	I	also	wished	to	test		 90	whether	feeling	either	more	empathy,	or	more	connection,	to	an	out-group	cause	would	influence	donation	support	for	it,	and	specifically	whether	this	would	decrease	following	a	belonging	manipulation.	Previous	research	has	found	empathy	to	be	associated	with	increased	prosocial	behaviour	(Batson,	1991;	Batson	&	Shaw,	1991;	Davis,	2018;	Dovidio,	Piliavin,	Schroeder,	&	Penner,	2006;	Eisenberg	&	Miller,	1987;	Penner,	Dovidio,	Piliavin,	&	Schroeder,	2005;	Stocks,	Lishner	&	Decker,	2009).	With	this	in	mind,	I	also	elected	to	measure	connection	and	empathy	to	cause	as	additional	dependent	variables.	I	predicted	that	connection	to	cause	would	be	predicted	by	independent	self-construal	and	would	be	moderated	by	the	inclusion	manipulation.	For	empathy	I	was	less	certain	as	to	my	prediction,	since	past	findings	regarding	the	connection	between	empathy	and	prosocial	behaviour	have	been	mixed,	with	some	finding	a	relationship	and	others	not	(see	Eisenberg	&	Miller,	1987;	Underwood	&	Moore,	1982)	for	a	meta-analysis	review).			Finally	in	study	3	I	wished	to	eliminate	the	possibility	that	good	mood	or	positive	affect	(derived	from	an	inclusion	manipulation)	may	be	driving	the	mechanisms.	Specifically,	I	wished	to	eliminate	the	possibility	that	the	reduction	in	prosocial	inclusive	behaviour	for	people	high	in	independent	self-construal,	was	as	a	result	of	positive	affect,	rather	than	it	being	as	a	direct	result	of	an	affirmation	of	social	inclusion.		In	order	to	examine	this,	study	3	included	a	positive	affect	condition,	alongside	the	inclusion	and	control	conditions.	3																																																										3	Note:	A	limitation	of	study	2	may	be	noted	from	the	power	analysis,	which	showed	sub-optimal	power	(0.71)	to	detect	an	effect	size	of	.127	and	higher	in	regressions.		Again,	as	per	study	1,	to	be	sure	of	the	robustness	of	these	results	it	would	be	advisable	to	replicate	this	result	in	future	research.		 91	3.4	Study	3	3.4.1	Overview.	Study	3	had	a	number	of	goals.	First,	I	wished	to	rule	out	the	impact	that	positive	affect	might	be	having	on	the	results.	It	is	possible	that	just	feeling	happy,	as	a	result	of	the	inclusion	manipulation	that	delivered	a	reassurance	of	long-term	social	inclusion,	might	account	for	the	behaviours	so	far	seen.	Past	research	on	positive	affect	has	found	it	to	be	positively	associated	with	prosocial	behaviour	(Aderman,	1972;	Cunningham,	Steinberg,	&	Grev,	1980;	George,	1991;	George	&	Brief,	1992;	Isen,	Clark,	&	Schwartz,	1976;	Isen	&	Levin,	1972;	Isen,	Shalker,	Clarke,	&	Karp,	1978;	Rosenhan,	Salovey,	&	Hargis,	1981;	Rosenhan,	Salovey,	Karylowski,	&	Hargis,	1981).	In	order	to	account	for	this	possibility,	in	study	3	I	included	a	positive	affect	condition,	as	well	as	an	inclusion	condition	and	a	control.			Another	aspect	I	wished	to	account	for	in	study	3	was	the	potential	role	of	participants’	baseline	trust	and	thwarted	belonging.		Twenge,	Baumeister,	DeWall,	Ciarocco,	and	Bartels	(2007),	Twenge,	Ciarocco,	Cuervo,	Bartels,	and	Baumeister	(2005),	as	well	as	Beest	and	Williams	(2011)	and	Catanese	and	Tice	(2005)	have	all	offered	evidence	to	suggest	that	social	exclusion	may	increase	aggression	and	negatively	predict	prosocial	behaviour.	However,	Lee	and	Shrum	(2012)	have	argued	that	being	explicitly	rejected	may	conversely	lead	to	more	prosocial	behaviour.		In	light	of	these	mixed	findings	regarding	the	influence	of	social	exclusion,	we	may	concede	that	trait	personality	differences	may	account	for	the	mixed	results.				 92	In	study	2	I	had	measured	need-to-belong	(NTB),	amongst	other	traits,	with	no	significant	results.	However,	it	is	possible	that	not	only	do	people	differ	in	terms	of	their	need	to	be	socially	included,	but	people	may	also	differ	in	terms	of	how	much	their	need	to	be	socially	included	is	currently	being	met—a	measurement	that	may	have	a	greater	impact.	Thwarted	belonging,	the	extent	to	which	there	may	be	unmet	belonging	needs	or	a	perceived	discrepancy	between	one’s	desired	and	actual	levels	of	belonging,	has	been	found	to	predict	self-defeating	behaviours	which	include	increased	aggression	and	reduced	prosocial	intentions	(Thau,	Aquino,	&	Poortvleit,	2007;	Twenge,	Baumeister,	Tice,	&	Stucke,	2001).	Belongingness	theory	states	that	people	have	a	fundamental	need	to	maintain	high	quality	relationships	with	other	people,	and	that	when	belonging	needs	are	thwarted	people	will	react	adversely	(Baumeister	&	Leary,	1995).	Adverse	reactions	happen	because	the	satisfaction	of	a	fundamental	need	has	been	denied,	causing	individuals	to	suffer	from	ego	depletion	and	identity	threat	(Thau,	Aquino,	&	Poortvleit,	2007).				With	this	in	mind	it	is	reasonable	to	predict	that	people	with	higher	thwarted	belonging	might	respond	differently	to	an	inclusion	manipulation,	than	those	who	have	their	belonging	needs	generally	met.	For	study	3,	I	therefore	decided	to	measure	trait	thwarted	belonging	prior	to	the	manipulation.	I,	additionally,	opted	to	measure	trait	levels	of	trust.	The	reason	for	this	is	that	Baumeister,	Brewer,	Tice	and	Twenge	(2007)	also	suggest	that	that	one	of	the	reasons	why	aggression	increases	and	prosocial	behaviour	decreases	when	belonging	is	thwarted	is	that,	despite	having	a	heightened	interest	in	forming	new	relationships,	people	may	restrain		 93	themselves	because	they	are	also	distrustful	of	others.	One	question	for	study	3,	therefore,	was	whether	high	levels	of	thwarted	belonging,	and	low	levels	of	trust,	might	attenuate	responses	to	the	inclusion	manipulation,	and	if	so	whether	the	effect	would	be	the	same	for	both	self-construal	types.			In	study	3	I	also	wished	to	investigate	whether	a	different	manipulation	of	inclusion–that	of	reliving	a	past	real	life	event	in	which	you	felt	strongly	included	and	writing	about	it–would	serve	to	act	as	an	inclusion	manipulation.	Writing	tasks	such	as	these	have	been	successfully	used	by	a	number	of	studies	to	make	salient	or	prime	emotions	and	mind	states	(e.g.	Arndt,	Greenberg,	&	Cook,	2002,	Knowles	&	Gardner,	2008;	Pickett,	Gardner,	&	Knowles,	2004;	Manner,	DeWall,	Baumeister,	&	Schaller,	2007;	Waytz	&	Epley,	2012).	From	a	practical	perspective,	such	reminders	of	past	events	of	social	inclusion	offer	more	possibilities	for	real-life	applications	than	the	more	onerous	future	life	predictions,	which	take	time	to	set	up	and	may	be	less	plausible	in	a	day-to-day	setting.	Replicating	my	previous	findings,	with	an	alternate	manipulation,	would	therefore	not	only	potentially	increase	the	robustness	of	my	research,	but	also	the	applicability	of	it.			Additionally,	and	perhaps	most	importantly,	in	study	3	I	wished	to	investigate	whether	the	previously	seen	inclusive	behaviour	might	extend	into	another	form	of	prosocial	behaviour,	one	with	direct	financial	benefits.	The	potential	of	increasing	donations	for	out-group	charities	has	direct	applicability	in	a	real	world	setting.	Many	conservation	charities,	as	well	as	charities	aimed	at	benefiting	third	world		 94	causes,	struggle	to	attract	sufficient	donors	and	funding.	For	this	reason,	in	study	3	I	substituted	a	donation	support	dependent	variable,	in	place	of	the	inclusion-group	task.	My	prediction	(H7)	was	that	prosocial	effects	seen	in	studies	1	and	2	would	carry	over,	and	that	under	normal	conditions	(control),	individuals	with	high	independent	self-construal	would	donate	more	to	an	out-group	charity	than	individuals	with	high	interdependent	self-construal.	Conversely,	however,	I	predicted	(H9)	that	under	inclusion	conditions,	individuals	with	high	independent	self-construal	would	behave	more	like	individuals	with	high	interdependent	self-construal,	and	donate	less	than	in	the	control	condition.	My	prediction	(H12)	for	individuals	with	high	interdependent	self-construal	was	that	they	would	not	only	donate	less	for	an	out-group	charity	than	those	high	in	independent	self-construal	in	general,	but	also	that	the	inclusion	manipulation	would	have	little	impact,	based	on	previous	findings	(Pfundmair,	Graupmann,	Frey,	&	Aydin,	2015;	Powers,	Worsham,	Freeman,	Wheatley,	&	Heatherton	2014).		Finally,	I	also	wished	to	test	whether	feeling	either	more	empathy,	or	more	connection,	to	an	out-group	cause	would	influence	donation	support	for	it,	and	specifically	whether	this	would	decrease	following	an	inclusion	manipulation.	My	expectation	here	was	that	feelings	of	connection,	or	empathy,	for	a	cause	would	potentially	mediate	donation	behaviour.		3.4.2	Procedure.	Study	3	took	the	form	of	a	single	factor	3-level	(control,	inclusion	and	positive	affect	manipulation)	mixed	experimental	design,	with	self-construal	measured	as	a	moderator	variable.	A	convenience	sample	of	187	American	participants	(47%		 95	female)	was	recruited	using	Mechanical	Turk,	and	completed	the	study	in	exchange	for	monetary	compensation.		Condition	was	assigned	randomly,	using	randomizing	embedded	coding	within	Qualtrics	survey	software.			The	sample	size	was	based	on	standard	calculations,	using	G*Power	3.1,	for	a-priori	multiple	regression	with	three	predictor	variables	(1	continuous,	1	categorical,	and	the	interaction),	using	random	sampling	at	95%	confidence	level.	Residual	variance	was	estimated	at	1-(R2)	=	0.85.	Variance	explained	by	special	effect	(self-construal	as	a	continuous	variable)	was	estimated	at	0.08.	Calculations	estimated	a	sample	size	of	106	or	greater	would	be	needed	for	each	regression	model,	in	order	to	achieve	a	power	of	>0.80.			Participants	were	told	that	they	would	be	completing	a	series	of	assignments	that	would	include	a	short	writing	skills	task,	as	well	as	a	poster	evaluation	task.	After	filling	in	a	number	of	trait	scales,	ostensibly	to	assess	mood,	participants	were	randomly	assigned	to	one	of	three	conditions:	control,	inclusion,	and	positive	affect.	Dependent	on	condition,	participants	were	asked	to	complete	a	writing	task,	the	subject	of	which	they	were	told	was	randomly	generated.	In	fact	participants	were	either	told	to	“think	about	a	time	in	which	you	went	to	the	grocery	store“	(control	condition),	to	“think	about	a	time	in	which	you	felt	very	happy”	(positive	affect	condition),	or	to	“think	about	a	time	in	which	you	felt	a	strong	sense	of	being	included	or	belonging”	(inclusion	condition).	Participants	were	asked	to	recall	the	event	as	vividly	as	possible,	then	to	write	a	paragraph	describing	the	past	event	in	as		 96	much	detail	as	they	were	able.	This	writing	task	has	previously	been	used	successfully	(e.g.	Knowles	&	Gardner,	2008;	Pickett,	Gardner,	&	Knowles,	2004;	Manner,	DeWall,	Baumeister,	&	Schaller,	2007;	Waytz	&	Epley,	2012),	as	a	way	of	manipulating	a	sense	of	social	inclusion/exclusion.	Following	the	manipulation,	participants	were	thanked	and	immediately	moved	on	to	the	next	task;	which	they	were	told	involved	evaluating	a	charity	poster	appeal	for	a	British	Columbia	Society	for	Prevention	of	Cruelty	to	Animals	(BC	SPCA)	animal	shelter.		They	were	asked	to	express	donation	intentions,	in	a	3-item	measure,	as	well	as	to	state	how	much	empathy,	and	connection,	they	felt	for	the	cause.	Demographics	were	then	collected.	Finally	participants	were	probed	as	to	the	subject	of	the	writing	task	that	they	received,	as	a	manipulation	check,	before	being	debriefed	and	thanked.			Materials.	Participants	were	asked	to	complete	the	same	24-item	self-construal	scale	as	used	in	Study	1	and	2	(scale	reliability:	interdependent	self-construal	12	items α=.81;	scale	reliability: independent	self-construal	12	items α=.83).	Furthermore,	a	9-item	thwarted	belonging	(THWB)	scale	was	administered,	(scale	reliability α=.94), which	was	sourced	from	the	15-item	Interpersonal	Needs	Questionnaire	(INQ;	Van	Orden,	Cukrowicz,	Witte	&	Joiner	Jr.,	2012);	a	scale	designed	to	measure	thwarted	belongingness	and	perceived	burdensomeness.	The	thwarted	belongingness	scale	(THWB)	contains	questions	conceived	to	measure	a	person’s	level	of	unmet	belonging,	as	opposed	to	their	desire	or	need	to	belong	(NTB;	see	appendix	A.9	for	full	scale).	Participants	were	asked	to	respond	on	a	7-point	scale	(1	=	not	very	true		 97	of	me,	7=very	true	of	me)	to	a	series	of	questions:	“	I	feel	like	I	belong”,	“I	am	close	to	other	people”,	“I	have	at	least	one	satisfying	interaction	every	day”,	“other	people	care	about	me”,	“I	am	fortunate	to	have	many	caring	and	supportive	friends”,	“I	feel	that	there	are	people	I	can	turn	to	in	times	of	need”,	as	well	as	the	following	reverse	coded	questions	“I	feel	disconnected	from	other	people”,	“I	rarely	interact	with	people	who	care	about	me”	“I	often	feel	like	an	outsider	in	social	gatherings”.	In	view	of	the	results	from	the	need-to-belong	measure,	administered	in	study	2,	I	anticipated	higher	unmet	belonging	needs	might	heighten	responses	in	the	inclusion	condition.		In	addition	a	2-item	trust	scale	(scale	reliability α=.74)	was	administered	using	the	same	a	7-point	scale	(1	=	not	very	true	of	me,	7=very	true	of	me)	to	respond	to	two	statements	“I	can’t	depend	on	people	to	have	my	best	interests	at	heart”,	”I	feel	that	people	can	be	counted	on	to	help	me”.		Trust	has	been	previously	shown	to	promote	prosocial	behaviour	(Cuadrado,	Tabernero,	&	Steinel,	2015;	Thau,	Aquino,	&	Poortvliet,	2007).		Moreover	trust	has	been	found	to	be	affected	by	empathy,	and	is	potentially	closely	related	(Feng,	Lazar,	&	Preece,	2004;	Ickes,	Stinson,	Bissonnette,	&	Garcia,	1990).	Since	both	empathy	and	prosocial	behaviour,	in	the	shape	of	donation	intentions,	were	to	be	measured	as	dependent	variables,	I	wished	to	measure	trust	as	a	potential	covariate	trait.			The	writing	task	that	followed	these	measures	was	used	in	place	of	the	Future	Life	Prediction,	in	order	to	hopefully	increase	robustness	of	the	findings,	through	the	use	of	a	different	manipulation	for	social	inclusion.	Writing	tasks	have	been	used	by	a	number	of	past	researchers	(e.g.	Knowles	&	Gardner,	2008;	Pickett,	Gardner,	&		 98	Knowles,	2004;	Manner,	DeWall,	Baumeister,	&	Schaller,	2007;	Waytz	&	Epley,	2012),	in	order	to	make	specific	frames	of	mind,	or	emotions,	salient	for	participants.		Participants	were	encouraged	to	take	time	to	conjure	up	a	past	occasion	as	vividly	as	possible	first,	then	to	relive	it	through	the	writing	of	a	detailed	description	of	the	event	that	included	their	feelings	during	the	event.	The	detailed	instructions	are	given	for	each	condition	in	the	appendix	(A.10).		In	the	next	section	of	the	study,	in	place	of	the	grouping	task	used	in	study	1	and	study	2,	participants	were	requested	to	view	a	poster	for	an	animal	charity,	purportedly	to	assess	how	they	think	about	pictorial	images,	and	then	to	rate	the	cause	on	a	number	of	dependent	variables	(see	appendix	A.11	for	poster).	The	dependent	variables	included	a	3-item	donation	intentions	measure:	“how	likely	are	you	to	donate	to	this	cause”,	“how	willing	are	you	to	donate	to	this	cause”	and	“how	inclined	are	you	to	donate	to	this	cause”,	captured	using	a	7-point	scale	(low-high)	(scale	reliability:	α=.957).	Participants	were	then	asked	how	much	empathy	they	felt	for	the	cause,	and	how	connected	they	felt	to	the	cause,	using	a	7-point	scale	(low-high).	Demographics	were	collected.	Participants	were	then	asked	to	recall	their	writing	task	as	a	manipulation	check,	before	being	debriefed	and	thanked.			3.4.3	Results.	My	study	exclusion	policy	stipulated	that	participants	would	be	screened	out	if	they	failed	all	attention	checks,	or	if	they	accurately	guessed	the	key	study	hypotheses;	which	no	participants	did.			 99	Power.	Post-hoc	power	analyses,	using	G*Power,	detected	sufficient	power	(>.80)	to	detect	correlations	of	r(185)=.210	and	higher,	effect	sizes	of	d=0.45	and	higher	for	independent	samples	t-tests	for	condition	(group	sizes:	60,120),	and	effect	sizes	of	d=0.42	and	higher	for	independent	samples	t-tests	for	gender	(group	sizes:	87,100)	at	95%	Confidence.	See	appendix	A.17	for	power	analysis.			Manipulation	check.		To	assess	whether	the	manipulation	had	been	effective,	participants	were	asked	to	recall	the	writing	task	they	received,	in	a	multiple-choice	question	at	the	end	of	the	study,	after	demographics	were	taken.	When	asked	to	recall	their	writing	task	99.5%	of	participants	were	able	to	do	so	with	complete	accuracy.				Demographics.	Demographics	collected	showed	that	participants	in	the	sample	identified	as	77.5%	Caucasian,	8%	Black,	4.3%	Latin	American,	3.7%	South	East	Asian,	3.2%	Asian	and	3.2%	mixed	other	(see	figure	16).	Results	also	showed	that	93%	of	participants	reported	having	lived	in	the	USA	for	over	21	years.	Participants	identified	as	46.5%	female.	Age	demographics	were	as	follows:	0.5%	under	21	years	old;	48.1%	21-30	years	old;	27.3%	31-40	years	old;	14.4%	41-50	years	old;	8.6%	51-60	years	old	and	1.1%	over	60	years	old	(see	figure	17).					100	Demographic	and	trait	correlations.		Results	of	Study	3	revealed	few	significant	or	conclusive	correlations	between	the	demographic	and	trait	participant	variables	measured,	with	one	exception	for	social	desirability,	which	was	significantly	positively	correlated	with	independent	self-construal	(r(187)=.244,	p=.001)		and	significantly	negatively	correlated	with	interdependent	self-construal	(r(187)=-.258,	p<.001).		See	appendix	A.12	for	detailed	non-significant	results.					Means	comparisons.		An	independent	samples	t-test	was	carried	out	between	the	inclusion	group	and	affect	+	control	condition	as	a	separate	analyses.	The	t-test	showed	no	significant	differences	between	conditions	on	the	donation	dependent	variable	(t(185)=-0.449,	p=.65,	nor	for	empathy	to	cause	(t(185)=-0.515,	p=.61),	nor	connection	to	cause	(t(185)=-0.375,	p=.71).		Furthermore	a	separate	independent	samples	t-test	was	carried	out	between	the	control	group	and	affect	+	inclusion	condition.	This	also	showed	no	significant	differences	between	conditions	on	the	donation	dependent	variable	(t(185)=1.45,	p=.15,	nor	for	empathy	to	cause	(t(185)=0.654,	p=.51),	nor	connection	to	cause	(t(185)=1.16,	p=.25).	Finally	a	separate	independent	samples	t-test	was	carried	out	between	the	affect	group	and	control	+	inclusion	condition.	This	also	showed	no	significant	differences	between	conditions	on	the	donation	dependent	variable	(t(185)=-0.981,	p=.33),	nor	for	empathy	to	cause	(t(185)=-0.133,	p=.89),	nor	connection	to	cause	(t(185)=-0.773,	p=.44).									101	An	independent	samples	t-test	was	carried	out	to	test	for	a	difference	between	genders,	as	separate	analysis.	The	t-test	(t(185)=2.215,	p=.03,	d=	0.325)	showed	a	significant	difference	between	genders	on	the	donation	dependent	variable,	with	females	expressing	higher	donation	intentions	(M=14.71,	SD=4.77)	than	males	(M=13.03,	SD=5.51).	No	gender	difference	was	found	regarding	empathy	to	cause	(t(185)=1.579,	p=.17),	or	connection	to	cause	(t(185)=1.503,	p=.13).			Regression	tests	for	moderation.	Moderated	multiple	regressions	were	carried	out	in	SPSS	using	process	v.3	(Hayes,	2017)	to	examine	a)	whether	the	manipulation	(inclusion	vs.	control	vs.	affect	condition)	influenced	participant	donation	intentions,	actual	donation	and	evaluations	of	the	cause,	and	b)	whether	these	effects	were	moderated	by	independent	and	interdependent	self-construal	orientation.	Following	Cohen	and	Cohen	(1983)	and	Wendorf	(2004),	I	dummy	coded	the	condition	so	that	the	inclusion	manipulation	vs.	control,	the	affect	manipulation	vs.	control,	and	the	affect	manipulation	vs.	the	inclusion	manipulation	were	run	as	separate	tests,	and	the	results	reported	independently.	Independent	self-construal	and	interdependent	self-construal	scales	were	kept	separate	for	the	analysis,	but	were	mean	centered	in	accordance	with	guidelines,	stipulating	that	this	practice	renders	subsequent	tests	of	hypotheses	and	regression	coefficients	for	X	and	M	more	meaningful	and	substantially	interpretable,	as	well	as	to	reduce	the	likelihood	of	errors	in	interpretation	(Hayes,	2013).	All	dependent	measures	were	subjected	to	separate	moderated	multiple	regressions	with	each	manipulation	condition	and	the	control,		102	independent	and	interdependent	self-construal	scores,	and	their	interactions	simultaneously	entered	as	predictor	variables.	As	in	the	previous	studies,	unstandardized	coefficients	(B)	are	reported	rather	than	betas.	Because	betas	are	not	properly	standardized	in	interaction	terms	they	are	not	interpretable,	whereas	B	represents	the	difference	between	the	unweighted	means	of	the	groups	involved	(see	Cohen	et	al.	2003).			3.4.3.1	Role	of	positive	affect	manipulation	in	donation	intentions.		Regression	analyses	showed	no	significant	results	for	the	affect	manipulation	when	compared	to	the	control	condition	in	a	series	of	separate	models.	Furthermore,	the	pattern	of	the	non-significant	influence	was	in	the	same	direction	for	both	affect	and	control,	for	all	the	models	with	independent	self-construal.	These	findings	supported	my	prediction	that	the	results	seen	following	an	inclusion	manipulation	were	not	as	a	result	of	positive	affect,	but	were	specifically	as	a	result	of	perceptions	of	social	inclusion.	Although	the	pattern	varied	a	little	in	the	interdependent	models,	in	consideration	of	the	fact	that	in	study	1	and	study	2	the	main	significant	effects	seen	were	all	in	terms	of	people	high	in	independent	self-construal,	and	low	in	interdependent	self-construal,	the	decision	was	made	to	focus	on	the	difference	between	the	inclusion	vs.	control	conditions	and	interactions	between	inclusion	and	independent	self-construal	for	the	remainder	of	the	research.	I	therefore	opted	to	analyze	the	affect	and	control	conditions	combined	together,	in	comparison	with	the	inclusion	condition,	for	the	remainder	of	study	3.	See	appendix	A.12	for	a		103	comparison	of	all	of	the	affect	vs.	control	models;	affect	vs.	inclusion	models;	and	control	(only)	vs.	inclusion	models.		3.4.3.2	Role	of	inclusion	manipulation	(vs.	control	+	affect)	in	donation	intentions.	Following	the	procedures	previously	outlined,	I	regressed	donation	intentions	on	the	dummy	(control	+	affect	vs.	inclusion)	and	moderator	variables.			Donation	intentions	and	independent	self-construal.	Summary	of	results.	In	a	separate	model	with	independent	self-construal	the	inclusion	manipulation	did	not	significantly	predict	donation	intentions,	in	terms	of	a	conditional	effect,	although	independent	self-construal	did.	Of	greater	interest,	however,	was	the	significant	interaction	between	independent	self-construal	and	the	inclusion	manipulation.	As	predicted,	and	in	line	with	study	2,	whether	donation	intentions	were	greater	for	the	inclusion	condition,	over	the	control	+	affect	condition,	depended	on	participant’s	independent	self-construal.		As	illustrated	in	figure	18,	the	higher	an	individual’s	independent	self-construal,	the	higher	the	donation	intentions,	in	the	control	+	affect	condition.	However	this	pattern	was	reversed	with	the	inclusion	manipulation.		See	analysis	results	below.					104	Regression	model	of	donation	intentions	&	independent	self-construal.	Donation	intentions	were	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	inclusion	manipulation,	independent	self-construal	orientation,	and	their	interaction	were	found	to	account	for	any	significant	proportion	of	variability	in	donation	intentions	in	the	overall	model,	R2=.050,	F(3,183)=3.15,	p=.03.		Independent	self-construal	did	significantly	predict	donation	intentions	B=0.111,	SE=0.038,	t(187)=2.908,	p=.004,	[0.04,0.19]	but	the	inclusion	manipulation	did	not,	B=-0.552,	SE	=	0.800,	t(187)=-0.690,	p=.49,	95%	CI	[-2.13,	1.03]	.	In	sum	there	was	no	conditional	effect	found	for	the	manipulation	but	there	was	for	self-construal	orientation.		There	was	a	significant	two-way	interaction	between	the	manipulation	(control	+	affect	vs.	inclusion),	and	independent	self-construal	orientation	found	B=-0.164,	SE	=	0.071,	t(187)=-2.309,	p=	.02,	95%	CI	[-0.30,-0.02].				Regression	model	donation	intentions	&	independent	self-construal	controlling	for	gender.	Donation	intentions	were	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables	with	gender	entered	as	a	covariate.		The	manipulation,	independent	self-construal	orientation,	and	their	interaction	were	found	to	account	for	a	significant	proportion	of	variability	in	donation	intentions	in	the	overall	model,	R2=.075,	F(4,182)=3.701,	p=.006,	when	gender	was	controlled	for.		Independent	self-construal	did	significantly	predict	donation	intentions	on	its	own	B=0.112,	SE=0.038,	t(187)=2.969,	p=.003,	95%	CI	[0.04,0.19],	but	the	inclusion	manipulation		105	did	not,	B=-0.769,	SE	=	0.797,	t(187)=-0.965,	p=.34,	95%	CI	[-2.34,0.80]	.	In	sum	there	was	no	conditional	effect	found	for	the	manipulation.		There	was	a	significant	two-way	interaction	between	the	manipulation	(inclusion	vs.	control	+	affect),	and	independent	self-construal	orientation	found	B=-0.146,	SE	=	0.071,	t(187)=-2.060,	p=	.04,	95%	CI	[-0.29,-0.01]	when	controlling	for	gender.	Gender	did	significantly	predicted	donation	intentions	on	its	own	B=-1.723,	SE=0.760,	t(187)=-2.267,	p=.03,	95%	CI	[-3.22,0.22].	See	figure	18	for	a	visualization	of	this	with	the	Johnson-Neyman	Point	showing	the	region	of	independent	self-construal	values	(filled	area	above	9.1342)	for	which	a	floodlight	test	would	reveal	significant	differences	between	the	two	groups,	and	figure	19	for	a	visualization	of	the	floodlight	analysis.		Controlling	for	demographics	and	trait	measures	(Thwarted	Belonging,	Trust,	Social	Desirability)	revealed	no	different	findings	to	the	above,	although	gender,	trust	and	thwarted	belonging	were	predictive	of	donation	intentions	in	the	models	with	independent	self-	construal,	and	improved	predictive	models.	See	appendix	A.12	for	detailed	covariate	results	including	demographic	covariates.			Donation	intentions	and	interdependent	self-construal.	Summary	of	results.	In	a	separate	model	with	interdependent	self-construal	the	inclusion	manipulation	did	not	significantly	predict	donation	intentions	in	terms	of	a	conditional	effect,	and	did	not	appear	to	differ	significantly	from	the	control	+	affect	condition.	The	regression	model	showed	no	significant	interaction	between	interdependent	self-	106	construal	and	the	affect	manipulation.	Controlling	for	the	trait	measures	(Thwarted	Belonging,	Trust,	Social	Desirability)	revealed	no	significant	or	different	findings	to	the	above.	See	main	analysis	results	below.	See	appendix	A.12	for	detailed	covariate	results	including	demographic	covariates.			Regression	model	of	donation	intentions	&interdependent	self-construal.	Donation	intentions	were	regressed	on	to	the	dummy	(affect	condition)	and	moderator	variables.		The	affect	manipulation,	interdependent	self-construal	orientation,	and	their	interaction	were	not	found	to	account	for	any	significant	proportion	of	variability	in	donation	intentions	in	the	overall	model,	R2=.034,	F(3,183)=2.17,	p=.09.		Interdependent	self-construal	did	not	significantly	predict	donation	intentions	B=0.066,	SE=0.044,	t(187)=1.498,	p=.14,	[-0.02,0.15]	nor	did	the	inclusion	manipulation,	B=-0.416,	SE	=	0.803,	t(187)=-0.518,	p=.61,	95%	CI	[-2.00,1.17]	.	In	sum	there	was	no	conditional	effect	found	for	the	manipulation	or	for	self-construal	orientation.		There	was	no	significant	two-way	interaction	between	the	manipulation	(control	+	affect	vs.	inclusion),	and	interdependent	self-construal	orientation	found	B=0.045,	SE	=	0.070,	t(187)=0.636,	p=	.53,	95%	CI	[-0.09,0.18].				3.4.3.3	Role	of	inclusion	manipulation	(vs.	control	+	affect)	on	empathy	and	connection	to	cause	as	dependent	variables.	Empathy	and	feelings	of	connection	to	a	target	have	been	previously	found	to	influence	prosocial	behaviour	(see	Batson,	1991;	Batson	&	Shaw,	1991	for	further	insights).	In	addition	to	the	donation	intentions	dependent	variable	study	3	also		107	measured	two	other	variables,	empathy	to	cause	and	connection	to	cause,	in	order	to	assess	whether	they	might	be	mediating	the	result.	A	correlation	analysis	showed	empathy	to	be	positively	related	to	donation	intentions	r(185)=.743	p<.001,	and	connection	to	cause	was	also	positively	related	to	donation	intentions	r(185)=.921,	p<.001.	However,	there	were	no	significant	results	found	for	empathy	as	a	dependent	variable	in	any	of	my	regression	models.	There	was	an	approaching	significant	regression	model	with	independent	self-construal	and	connection	to	cause,	which	followed	the	same	pattern	as	donation	intentions	and	produced	a	significant	interaction.	No	significant	results	or	patterns	were	found	with	interdependent	self-construal.	See	detailed	results	analysis	below.			Empathy	to	cause	&	independent	self-construal.	Empathy	to	Cause	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	inclusion	manipulation,	independent	self-construal	orientation	and	their	interaction	together,	were	not	found	to	account	for	a	significant	proportion	of	variability	in	empathy	to	cause	in	the	overall	model,	R2=.012,	F(3,183)=.760,	p=.52.		Independent	self-construal	did	not	significantly	predict	empathy	to	cause	B=0.015,	SE=.012,	t(187)=1.273,	p=.21,	[-0.01,0.04]	nor	did	the	inclusion	manipulation,	B=-0.159,	SE	=	0.252,	t(187)=-0.633,	p=.53,	95%	CI	[-0.66,	0.34]	.	In	sum,	there	was	no	conditional	effect	found	for	the	manipulation,	nor	for	self-construal	orientation.		There	was	no	significant	two-way	interaction	between	the	manipulation	(control	vs.	inclusion),	and	independent	self-construal	orientation	found	B=-0.027,	SE	=	0.022,	t(187)=-1.213,	p=	.23,	95%	CI	[-0.07,0.02].				108	Empathy	to	cause	&	interdependent	self-construal.	Empathy	to	cause	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	inclusion	manipulation,	interdependent	self-construal	orientation,	and	their	interaction	together,	were	not	found	to	account	for	any	significant	proportion	of	variability	in	empathy	to	cause	in	the	overall	model,	R2=.017,	F(3,183)=1.076,	p=.36.		Interdependent	self-construal	did	not	significantly	predict	empathy	to	cause	B=0.024,	SE=0.014,	t(187)=1.720,	p=.09,	[-0.004,0.05]	nor	did	the	inclusion	manipulation,	B=-0.122,	SE	=	0.250,	t(187)=-.489,	p=.63,	95%	CI	[-0.62,	0.37].	In	sum,	there	was	no	conditional	effect	found	for	the	manipulation	or	for	self-construal	orientation.		There	was	no	significant	two-way	interaction	between	the	manipulation	(control	vs.	inclusion),	and	interdependent	self-construal	orientation	found	B=0.023,	SE	=	0.022,	t(187)=-1.023,	p=	.31,	95%	CI	[-0.07,0.02].				Connection	to	cause	&	independent	self-construal.	Connection	to	Cause	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	inclusion	manipulation,	independent	self-construal	orientation,	and	their	interaction	together,		were	found	to	account	for	an	approaching	significant	proportion	of	variability	in	connection	to	cause	in	the	overall	model,	R2=.040,	F(3,183)=2.515,	p=.06.		Independent	self-construal	did	significantly	predict	connection	to	cause	B=0.037,	SE=0.014,	t(187)=2.619,	p=.01,	[0.01,0.07],	but	the	inclusion	manipulation	did	not	B=-0.175,	SE	=	0.299,	t(187)=-0.586,	p=.56,	95%	CI	[-0.77,	0.42].	In	sum	there	was	no	conditional	effect	found	for	the	manipulation	but	there	was	for	self-construal	orientation.		There	was	a		109	significant	two-way	interaction	between	the	manipulation	(control	vs.	inclusion),	and	independent	self-construal	orientation	found	B=-0.054,	SE	=0.027,	t(187)=-2.027,	p=	.04,	95%	CI	[-0.11,-0.01].			Connection	to	cause	&	interdependent	self-construal.	Connection	to	Cause	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	inclusion	manipulation,	interdependent	self-construal	orientation,	and	their	interaction	together,	were	not	found	to	account	for	any	significant	proportion	of	variability	in	connection	to	cause	in	the	overall	model,	R2=.027,	F(3,183)=1.683,	p=.17.		Interdependent	self-construal	did	not	significantly	predict	connection	to	cause	B=0.026,	SE=0.016,	t(187)=1.559,	p=.12,	[-0.01,0.06],	nor	did	the	inclusion	manipulation,	B=-0.127,	SE	=	0.30,	t(187)=-0.422,	p=.67,	95%	CI	[-0.72,	0.47].	In	sum	there	was	no	conditional	effect	found	for	the	manipulation	or	for	self-construal	orientation.		There	was	no	significant	two-way	interaction	between	the	manipulation	(control	vs.	inclusion),	and	interdependent	self-construal	orientation	found	B=0.007,	SE	=	0.026,	t(187)=.253,	p=	.80,	95%	CI	[-0.05,0.06].				See	figure	20	for	a	visual	comparison	of	the	trends	for	both	independent	and	interdependent	self-construal.		3.4.3.4	Tests	of	mediated	moderation.	In	consideration	of	the	previous	result,	my	next	goal	was	to	test	a	mediated	moderation	model,	to	see	if	connection	to	cause	acted	as	a	mediator	for	the		110	mechanisms	involved.	Mediated	moderation	occurs	when	two	predictor	variables	(for	example	Condition	and	Self-Construal)	interactively	affect	a	mediator	(such	as	Connection	to	Cause),	which	in	turn	influences	an	outcome	variable,	such	as	Donation	intentions	(Morgan-Lopez	&	MacKinnon,	2006).	Based	on	my	previous	results,	I	wished	to	investigate	whether	the	mechanism	that	was	driving	donation	intentions	was	mediated	by	connection	to	the	cause.			Connection	to	the	cause	as	a	mediator	(with	inclusion	vs.	control	+	affect	–	model	7).	I	conducted	a	moderated	mediation	model	(model	7;	Hayes	2013)	with	condition	(inclusion	=	1;	control	+	affect	=	0)	as	the	independent	variable,	donation	intentions	as	the	dependent	variable,	connection	to	cause	as	the	mediating	variable,	and	independent	self-construal	and	the	moderating	variable	on	the	link	between	condition	and	connection	to	cause	(see	figure	23	for	conceptual	models	used).			Connection	to	Cause	was	regressed	on	to	the	dummy	(inclusion	condition)	and	moderator	variables.		The	manipulation,	independent	self-construal	orientation,	and	their	interaction	together,	were	found	to	account	for	an	approaching	significant	proportion	of	variability	in	connection	to	cause	in	the	overall	model,	R2=.040,	F(3,183)=2.515,	p=.06.		Independent	self-construal	did	significantly	predict	connection	to	cause	B=0.037,	SE=0.014,	t(187)=2.619,	p=.01,	95%	CI	[0.01,0.07].	However,	there	was	no	significant	conditional	effect	of	the	manipulation	(vs.	control)	on	donation	intentions	B=-0.109,	SE=0.443,	t(187)=-0.245,	p=81,	95%	CI	[-0.98,	0.77].	Nor	was	there	a	significant	conditional	effect	of	the	manipulation	(vs.		111	control)	on	connection	to	cause	B=-0.175,	SE	=	0.299,	t(187)=-0.586,	p=.56,	95%	CI	[-0.77,0.42].	There	was	a	significant	interaction	between	the	manipulation	(vs.	control)	and	independent	self-construal	on	connection	to	cause	B=-0.054,	SE	=0.027,	t(187)=-2.027,	p=	.04,	95%	CI	[-0.11,-0.001].	Finally,	there	was	a	significant	conditional	effect	of	connection	to	cause	on	donation	intentions	B=2.256,	SE	=	0.108,	t(187)=20.936,	p<	.001,	95%	CI	[2.04,2.47].			To	test	this	potential	mediation	effect,	I	followed	the	bootstrapping	method	(with	5000	iterations)	advocated	by	Preacher,	Rucker	and	Hayes	(2007)	and	Hayes	(2013).	This	procedure	tests	the	null	hypothesis	that	the	indirect	path	from	the	interaction	term	to	the	dependent	variable	via	the	mediator	does	not	statistically	differ	from	zero.	If	zero	is	not	contained	in	the	confidence	intervals	computed	by	the	bootstrapping	procedure	then	one	can	conclude	that	the	indirect	effect	is	indeed	significantly	different	from	zero	at	p=.05.	Bootstrap	estimates	generated	a	95%	confidence	interval	around	the	indirect	effect,	with	zero	falling	inside	the	confidence	interval,	indicating	the	mediating	pathway	was	not	significant	at	all	levels	of	W	(independent	self-construal)	Index=-0.1212,	Boot	SE=0.0728	95%	Boot	CI	[-0.2623,	0.0212].	Although	the	trend	was	the	same	as	seen	in	the	moderation	models,	since	there	was	a	zero	in	the	confidence	intervals	I	could	not	conclude	that	connection	to	cause	offered	a	mediating	pathway	to	donation	intentions,	in	this	model.							112	3.4.4	Discussion.	Study	3	was	able	to	repeat	the	main	findings	of	Study	2,	whilst	using	a	different	inclusion	manipulation	(writing	about	a	past	inclusion	event	as	opposed	to	an	affirmation	of	future	inclusion).	Study	3	was	also	able	to	extend	the	pattern	of	these	previous	findings	across	a	new	dependent	variable—that	of	prosocial	behaviour,	in	the	form	of	donation	intentions	directed	towards	an	out-group	animal	charity	cause.			Specifically,	in	study	3	I	was	able	to	provide	strong	support	for	H7,	in	that	under	normal	conditions	(control	group),	people	with	high	independent	self-construal	did	appear	to	express	higher	donation	intentions	for	an	out-group	animal	charity,	when	compared	to	people	with	lower	independent	self-construal,	as	well	when	compared	to	those	with	higher	interdependent	self-construal.	A	word	of	caution	must	be	offered	for	the	latter	of	these	two	observations,	however.		I	cannot	report	the	pattern	between	self-construals	as	statistically	significant,	since	the	observation	is	based	visual	trends	only,	and	no	comparative	tests	were	undertaken	between	the	two	types	of	self-construal;	which	were	measured	as	continuous	variables.		Although	this	result	is	therefore	tentative,	I	believe	it	is	at	least	worth	noting,	since	it	follows	in	line	with	past	research	which	has	demonstrated	that	higher	interdependent	self-construal	is	primarily	connected	to	benevolent	and	prosocial	behaviour	towards	in-group	recipients,	rather	than	out-group	recipients	(Kemmelmeier,	Jambor	and	Letner,	2006),	whilst	individuals	with	high	independent	self-construal	appear	to	show	less	preference	for	the	group	membership	status	of	the	target	(Duclos	&	Barasch,	2014).			113	Study	3	was	also	able	to	provide	strong	support	for	H9:	that	an	affirmation	of	social	inclusion	(inclusion	manipulation)	would	result	in	reduced	donation	intentions,	as	well	as	reduced	connection	to	cause,	expressed	towards	an	out-group	animal	charity,	for	people	with	a	high	independent	self-construal	orientation,	compared	to	under	normal	conditions	(control).	This	was	in	line	with	findings	in	study	1	and	2	for	the	inclusion	of	animals	in	an	in-group	outcome	variable.			Study	3	also	provided	evidence	that	the	influence	of	inclusion	on	donation	intentions	was	not	as	a	result	of	feelings	of	positive	affect.	Study	3	did	this	by	demonstrating	that	a	separate	affect	condition	did	not	produce	similar	results	to	the	inclusion	condition,	but	rather	produced	results	that	were	more	in	line	with	the	control	condition,	for	models	with	independent	self-construal.	Specifically,	study	3	showed	that	the	affect	manipulation	did	not	increase	or	decrease	donation	intentions	for	an	out-group	animal	charity,	for	people	with	an	independent	self-construal	orientation.		Study	3	also	found	evidence	to	support	the	expectation	(H12)	that	for	individuals	with	high	interdependent	self-construal,	the	inclusion	condition	would	have	little	influence	on	donation	intentions	(or	connection	to	cause).	Again	this	finding	is	in	line	with	past	research,	showing	that	an	interdependent	self-construal	may	buffer	an	individual	against	threats	of	exclusion	(Pfundmair,	Graupmann,	Frey	&	Aydin,	2015;	Powers,	Worsham,	Freeman,	Wheatley	and	Heatherton,	2014),	and	extends	the	finding	into	a	new	area;	that	of	promises	for	inclusion.		114	Controlling	for	personality	trait	measures	(thwarted	belonging,	and	trust)	in	study	3	revealed	no	significant	or	different	findings	to	the	above,	although	both	higher	levels	of	trust	and	reduced	levels	of	thwarted	belonging	were	predictive	of	donation	intentions	in	the	models	with	independent	self-construal,	and	therefore	improved	predictive	models.	They	did	not	appear	to	interact	with	any	of	the	other	dependent	variables,	so	my	tentative	prediction–that	thwarted	belonging	might	attenuate	responses	to	the	inclusion	manipulation–was	not	proven	to	be	true.			Study	3	notably	found	a	main	effect	of	gender	on	donation	intentions,	as	evidenced	by	an	independent	samples	t-test,	showing	that	females	expressed	significantly	higher	donation	intentions	than	males.	This	is	in	line	with	much	previous	research,	evidencing	that	females	donate	more	to	charity	appeals	than	do	males	(Einolf,	2011;	Hodgkinson	&	Weitzman,	1992;	1994;	1996;	Kirsh,	Hume,	&	Jalnadoni,	1999;	Mesche,	Rooney,	Steinberg,	&	Denton,	2006;	Winterich,	Mittal,	&	Ross	Jr.,	2009),	and	show	more	positive	attitudes	to	animal	welfare	(Heleski,	Mertig,	&	Zanella,	2006;	Herzog,	Betchart,	&	Pittman,	1991;	Hills,	1993;	Mathews	&	Herzog,	1997;	Schenk,	Templar,	Peters,	&	Schmidt,	1994)	and	therefore	was	somewhat	expected.	Regression	analysis	models	for	gender	in	study	three	revealed	no	significant	interactions	between	gender,	self-construal	and	inclusion	for	the	donation	intentions	task,	although	gender	was	predictive	of	donation	intentions	when	added	as	a	covariate	in	the	models	with	independent	self-construal,	and	therefore	improved	predictive	models.			115	Study	3	did	not	find	strong	evidence	in	support	of	H11,	that	connection	to	the	cause	was	mediating	donation	to	the	cause.	However,	in	the	independent	self-construal	models	tested,	visualizations	showed	that	the	non-significant	results	were	in	the	predicted	direction	(see	A.11).			Study	3	did	not	test	the	influence	of	anthropomorphism	(as	a	trait,	or	as	a	result	of	a	specific	task)	on	donation	intentions,	and	whether	self-construal	and	inclusion	interacted,	since	study	3	did	not	involve	an	inclusion	task.			Finally,	study	3,	while	it	did	find	some	evidence	that	connection	to	the	cause	might	be	predicted	by	inclusion	and	independent	self-construal,	did	not	find	any	evidence	that	empathy	for	the	cause	was	predicted	by	inclusion	or	self-construal.	This	null	finding	may	offer	potentially	interesting	theoretical	insights.		While	some	previous	research	has	suggested	that	empathy	is	strongly	associated	with	increased	prosocial	behaviour	(Batson,	1991;	Batson	&	Shaw,	1991;	Davis,	1996;	Dovidio,	Piliavin,	Schroeder,	&	Penner,	2006;	Eisenberg	&	Miller,	1987;	Penner,	Dovidio,	Piliavin,	&	Schroeder,	2005;	Stocks,	Lishner	&	Decker,	2009),	other	lines	of	research	have	produced	contradictory	results.	Underwood	and	Moore	(1982)	found	no	relationship	between	empathy	and	prosocial	behaviour	in	a	large	meta-analysis	review.	More	recently	Steele,	Schreiber,	Guiltinan,	Nass,	Glynn,	Wright,	Kessler,	Schlumpf,	Tu,	Smith,	and	Garratty,	(2008)	found	no	relationship	between	empathy	and	blood	donation	behaviour,	and	Dickert,	Sagara,	and	Slovic	(2011)	found	that	empathy	did	not	increase	donation	intentions,	but	only	the	amount	actually	given	once	a	decision	was	made.	In	view	of	these	contradictory	findings,	my	null	results	in		116	study	3	for	empathy	as	a	dependent	variable	are	not	a	total	surprise.		My	results	perhaps	beg	the	question	as	to	whether	a	more	nuanced	exploration	of	the	impact	of	empathy	may	offer	potential	for	future	research.			In	conclusion,	the	results	of	study	3,	when	taken	alongside	study	1	and	2,	follow	in	the	direction	of	prior	research,	demonstrating	that	social	inclusion	may	negatively	impact	prosocial	attitudes	towards	distant	others	and	out-group	members	(DeWall,	Baumeister	&	Vohs,	2008;	Waytz	&	Epley,	2012).		That	the	pattern	of	response	to	an	inclusion	manipulation,	as	described	above,	is	primarily	seen	in	people	with	high	independent	self-construal,	also	builds	on	previous	research	by	Pfundmair,	Graupmann,	Frey,	and	Aydin	(2015),	as	well	as	Ren,	Wesselmann,	and	Williams	(2013),	offering	up	evidence	that	people	with	a	more	interdependent	orientation	may	not	be	as	susceptible	to	promises	of,	or	threats	to,	social	inclusion.			Study	3	used	a	donation	intentions	dependent	variable.	However,	intention	to	donate	is	not	the	same	as	actual	donation.	In	study	4	I	wished	to	investigate	whether	the	behaviour	observed	in	study	3	might	extend	into	something	with	real-life	financial	benefits:	in	other	words,	cash	donation	towards	an	animal	charity.	I	also	wished	to	further	explore	the	motivations	for	out-group	giving	in	individuals	high	in	independent	self-construal.	To	do	this	I	planned	to	test	whether	feeling	similar,	or	different,	to	an	out-group	cause	would	influence	donation	support	for	it,	and	whether	similarity	or	difference	would	interact	with	either	the	inclusion	manipulation	or	self-construal.	My	planned	method	for	doing	this	was	to	use	a		117	poster	for	an	animal	charity,	with	a	wording	frame	that	discussed	either	how	different	animals	are	from	humans	“Animals	are	Different,”	or	one	that	discussed	how	similar	animals	are	to	humans	“Animals	are	Similar.”			I	predicted	that	individuals	high	in	interdependent	self-construal	would	preference	in-group	giving,	and	would	therefore	donate	more	to	an	“Animals	are	Similar”	poster	frame.	This	is	because	I	predicted	that	individuals	high	in	interdependent	self-construal	would	potentially	see	the	“Animals	are	Similar”	frame	as	an	in-group	cue,	vs.	the	“Animals	are	Different”	poster	frame,	which	I	anticipated	that	they	would	see	as	a	clear	out-group	cue.	I	did	not	anticipate	that	individuals	high	in	interdependent	self-construal	would	show	any	interaction	with	the	inclusion	manipulation,	as	previously	detailed.				Furthermore,	I	predicted	that	individuals	high	in	independent	self-construal	would	also	preference	the	“Animals	are	Similar”	poster	frame	normally	(control	condition),	yet	for	somewhat	different	reasons.	Study	3	demonstrated	that	individuals	high	in	independent	self-construal	donated	more	to	an	out-group	(animal)	charity	than	individuals	high	in	interdependent	self-construal,	when	the	charity	wasn't	framed	as	either	in-group	or	out-group.		I	argue	that	this	was	because	in	study	3	all	participants	would	have	seen	the	charity,	by	default,	as	an	out-group	charity.			We	have	seen	that	individuals	high	in	interdependent	self-construal	express	lower	donation	support	for	out-group	charities	than	individuals	high	in	independent	self-	118	construal.	I	have	previously	argued	that	the	motivation	for	higher	out-group	giving	by	individuals	high	in	independent	self-construal	is	because	a)	they	have	more	flexible	notions	of	in-group	and	out-group	boundaries,	and	so	they	are	more	likely	to	view	out-group	members	as	potential	targets	to	recruit	as	in-group	members	and	b)	since	they	are	not	as	chronically	reassured	of	their	social	inclusion	status,	compared	to	individuals	with	high	interdependent	self-construal,	they	monitor	and	are	more	vigilant	for	cues	of	social	connection	potential	in	others.	As	a	result,	in	study	4,	when	the	out-group	charity	is	specifically	framed	as	similar,	I	argue	that	individuals	with	high	independent	self-construal	will	still	see	it	as	an	out-group,	but	will	take	the	“Animals	are	Similar”	frame	as	a	cue	of	social	connection	potential	and	will	therefore	also	donate	more	to	the	“Animals	are	Similar”	poster	frame.			Support	for	a	helping	based	on	similarity	hypothesis	is	longstanding.	Psychological	literature	has	demonstrated	that	even	arbitrarily	manipulated	cues	of	similarity	(shared	names,	birthdays,	etc.)	are	associated	with	positive	perceptions	towards	another	(Berger,	Messian,	Patel,	del	Prado,	&	Anderson,	2004;	Finche	&	Cialdini,	1989;	Gueguen,	Pichot,	&	Le	Dreff,	2005;	Heider,	2013).	Krebs	(1975)	found	that	when	strangers	were	perceived	as	more	similar	participants,	were	more	willing	to	assist	them	and	give	up	resources	to	do	so.	Even	animals	have	been	found	to	preference	helping	similar	others	(Quervel-Chaumette,	Dale,	Marshall-Pescini,	&	Range,	2015).	More	recently,	research	in	marketing	has	found	that	the	more	similar	to	humans	animals	appear,	the	more	favourably	they	are	treated	(Connell,	2013).	Costello	and	Hodson	(2009;	2010)	additionally	found	that	inducing	perceptions	of		119	human-animal	similarity	facilitated	more	inclusive	attitudes	to	both	animal	and	human	outsiders,	and	predicted	less	prejudicial	attitudes	towards	even	human	immigrants.	Researchers	working	in	the	area	of	conservation	and	HAI,	have	also	found	that	a	strong	relationship	between	perceived	similarity	to	humans	and	preference	for	animal	welfare	exists	(Butterfield,	Hill,	&	Lord,	2012;	Hills,	1995;	Kellert,	1980;	Kiesler	&	Kramer,	2007),	while	Batt	(2009)	suggests	that	humans	are	predisposed	to	liking	others	on	the	basis	of	perceived	shared	bio-behavioural	traits.			With	this	in	mind,	another	goal	of	study	4	was	to	investigate	whether	connection	to	cause,	and	empathy	to	cause,	would	moderate	(and	potentially	mediate),	specifically	based	on	perceived	similarity	to	cause.	My	prediction	was	that	an	“Animals	are	Similar”	poster	would	increase	both	empathy	and	connection	to	cause.			Limitations.	It	should	be	noted	that	one	limit	of	study	3	was	potential	power	for	detecting	the	interactions	in	the	main	regressions.	Although	G*Power	calculations	indicated	that	power	would	be	sufficient	for	detecting	the	main	(conditional)	effects,	interaction	effects	in	regressions	are	typically	smaller	(Aiken	&	West,	1991)	and	therefore	require	higher	power	to	detect.	Accurate	calculations	for	complex	regression	models	require	estimates	of	expected	effects	sizes	(unstandardized	regression	coefficients),	at	each	level	of	the	categorical	variable	(or	an	estimate	of	the	size	of	relationships	between	continuous	variable	interactions	and	other	predictor	variables)	for	all	the	conditional	and	interaction	effects	in	the	population,	which	often	may	not	be	known		120	a	priori	(Hayes,	2018).		While	G*Power	3.1	is	a	respected	and	commonly	used	tool	for	power	analysis	(Faul,	Erdfelder,	Buchner,	&	Lang,	2009),	it	has	been	argued	it	may	be	limited	in	correctly	detecting	power	for	significant	coefficients,	in	complex	regressions,	including	those	that	include	interactions	and	mediations	(Aberson,	2011;	Fritz,	&	MacKinnon,	2007).	Power	analysis	is	a	rapidly	evolving	field	and	tools,	such	as	pwr2ppl	for	R,	are	currently	in	development	for	power	detection	of	mediation	and	complex	moderation	results	(Aberson,	2011).		In	the	meantime	a	word	of	caution	must	be	made	for	the	interaction	results,	and	a	suggestion	for	future	replication	is	made,	in	order	to	verify	the	results	of	this	study.				3.5	Study	4		3.5.1	Overview.	Study	4	had	a	number	of	objectives.	First,	I	wished	to	investigate	whether	the	behaviour	observed	in	study	3	might	extend	into	something	with	real-life	financial	benefits–in	other	words,	actual	cash	donation	towards	an	out-group	(animal)	charity.	For	this	reason,	study	4	was	designed	to	collect	an	actual	cash	donation	measure,	as	well	as	a	three-item	measurement	of	donation	intentions.			Second,	I	wished	to	explore	more	of	the	whens	of	the	process	(Hayes,	2013)	by	examining	another	potential	moderator	of	the	mechanism	in	study	4.	Specifically,	I	aimed	to	investigate	whether	a	feeling	of	closeness	and	oneness	(vs.	distance	and	otherness)	with	the	target	(in	my	case	an	animal	charity),	may	influence	prosocial	behaviour,	and	if	so	whether	it	would	have	significant	effects	regardless	of	self-	121	construal	and	inclusion	status.	In	order	to	do	this	I	added	an	additional	factor	to	study	4:	that	of	framed	similarity,	or	difference,	to	the	cause	targets.	My	prediction	in	this	regard	was	that	under	normal	conditions	all	participants	would	donate	more	to	the	“Animals	are	Similar”	poster	frame.	Following	an	inclusion	manipulation	I	anticipated	that	participants	high	in	interdependent	self-construal	would	still	donate	more	to	the	“Animals	are	Similar”	poster	frame,	due	to	the	lack	of	impact	inclusion	manipulations	have	on	participants	high	in	interdependent	self-construal.	However,	for	participants	high	in	independent	self-construal,	I	predicted	a	different	response	following	an	inclusion	manipulation.	For	these	participants,	I	predicted	that	an	“Animals	are	Different”	poster	frame	would	prove	more	appealing,	since	such	a	frame	would	be	more	likely	to	act	as	a	reinforcement	of	in-group	and	out-group	boundaries.			To	clarify,	I	anticipated	that	an	inclusion	manipulation	would	do	two	things.	First,	it	would	act	to	reassure	participants	high	in	independent	self-construal	of	their	social	inclusion	status,	and	thereby	reduce	a	desire	to	seek	further	social	connection	with	similar	others.	While	I	predicted	that	the	“Animals	are	Similar”	poster	frame	would	appeal	to	these	participants	under	normal	conditions	(control)	precisely	because	it	indicates	social	connection	potential	and	encourages	a	relaxing	of	in-group	boundaries,	as	supported	by	prior	research	demonstrating	that	perceptions	of	similarity	increase	prosocial	behaviour	(Connell,	2013;	Costello	&	Hodson,	2009;	Krebs,	1975),	under	conditions	in	which	the	participants	had	recently	been	assured	of	their	social	inclusion	within	a	human	in-group	by	a	social	inclusion	manipulation,		122	I	predicted	that	it	would	hold	little	appeal,	and	might	even	present	itself	as	threatening.	This	latter	prediction	is	based	on	what	I	argue	may	be	the	second	downstream	result	of	an	inclusion	manipulation	on	individuals	with	a	high	sense	of	independent	self-construal:	an	activated	desire	to	bolster	the	status	of	the	in-group	and	clearly	differentiate	it	from	an	out-group.			As	previously	discussed,	I	argue	that	individuals	with	high	independent	self-construal	hold	an	essentially	more	precarious	sense	of	social	inclusion	or	belonging	than	do	individuals	with	high	interdependent	self-construal.		As	a	result,	while	they	are	willing	to	see	the	social	potential	and	similarity	of	animal	out-group	members	when	looking	to	increase	social	connection	and	inclusion,	when	they	are	satiated	with	belonging	they	will	be	likely	to	view	animals	as	more	out-group,	as	evidenced	by	the	findings	of	studies	1	and	2.		Under	recently	affirmed	inclusion	circumstances,	I	argue	it	should	be	more	beneficial	for	individuals	with	high	independent	self-construal	to	fortify	and	defend	in-group	boundaries	against	out-group	targets,	clearly	differentiating	themselves	and	their	in-group	from	an	out-group.	This	type	of	behaviour	has	been	previously	addressed	from	both	social	identity	theory	(SIT),	and	optimal	distinctiveness	(ODT)	theory	(Brewer,	1993;	Hornsey,	&	Hogg,	1999;	Spears,	Doosje,	&	Ellemers,	1997).	Donating	to	a	clearly	labelled	out-group	charity	(“Animal	are	Different”),	which	is	defined	by	its	inferior	properties,	would	therefore	act	to	signal,	and	solidify,	higher	in-group	status,	and	offer	more	potential	as	an	action	of	self-benefit,	than	donating	to	a	potentially	in-group	status	threatening	out-group	charity	(“Animal	are	Similar”).	In	summary	I	had	two	hypotheses	for	study	4.		123	First	I	predicted	(H8)	that	under	normal	conditions	(control	group)	people	with	a	higher	independent	self-construal	orientation	would	express	increased	donation	intentions	for	an	out-group	animal	charity	that	was	framed	as	similar	(vs.	different)	to	them,	since	it	would	offer	increased	connection	potential.	Second	I	predicted	(H10)	that	an	affirmation	of	social	inclusion	would	result	in	increased	donation	intentions	for	an	out-group	animal	charity	that	was	framed	as	different	(vs.	similar)	to	the	donor	for	individuals	with	a	higher	independent	self-construal	orientation,	since	it	would	offer	less	of	a	threat	to	in-group	distinctiveness.		This	argument	is	in	line	with	prior	research,	suggesting	much	prosocial	behaviour	by	individuals	high	in	independent	self-construal	is	motivated	by	a	desire	to	fulfill	personal	happiness	and	self	benefit	needs	(Duclos	&	Barasch,	2014).	Moreover,	it	is	also	in	line	with	prior	research,	suggesting	that	giving	to	lower	status	out-groups	acts	as	a	mechanism	for	signalling	in-group	strengths	and	bolstering	in-group	status	(Nadler,	Harpaz-Gorodeisky,	&	Ben-David,	2009).			With	the	above	in	mind,	a	third	goal	of	study	4	was	to	also	drill	a	little	deeper	into	the	how	of	the	mechanism	(Hayes,	2013),	to	see	if	a	feeling	of	connection	to	the	cause	might	be	underpinning	at	least	some	of	the	effect	seen	in	study	3.	Specifically,	I	wished	to	test	whether	perceived	connection	to	cause	would	act	as	a	mediator	for	donation	support	(donation	intentions	and	actual	cash	donation),	and	whether	it	would	mediate	primarily	with	the	similar	(vs.	different)	prime.	My	prediction	(H11)	was	that	connection	to	cause	would	indeed	mediate	the	effects	of	independent	self-	124	construal	orientation,	inclusion	and	their	interaction,	on	the	two	dependent	variables.		Prior	studies	have	provided	evidence	that	feeling	more	similar	and	connected	to	others	enables	an	opening	of	admission	to	one’s	in-group	to	a	broader	set	of	others	(Cuddy,	Rock,	&	Norton,	2007).	Additionally,	past	research	on	intergroup	relations	argues	that	salient	in-group/out-group	categories	play	a	key	role	in	regulating	the	perception	of	self-other	similarities	(Sturmer	&	Snyder,	2010).	As	a	consequence,	people	come	to	perceive	in-group	members	as	more	similar	to	each	other	(and	the	self),	whereas	out-group	members	are	more	dissimilar	to	the	in-group	(and	the	self)	(Wilder,	1986).		As	Carnegie	(1936)	and	Cialdini	(2001)	have	both	notably	observed,	we	like	things	better	when	we	feel	we	share	commonalities.	Salience	of	the	differences	and	divisions	between	in-groups	and	out-groups,	on	the	other	hand,	may	provoke	negatives	feelings,	such	as	anxiety	or	threat	(Dijker,	1987;	Jackson	&	Sullivan,	1989;	Stephan	&	Stephan,	1985).	Sturmer	and	Snyder	(2010)	argue	that	when	there	is	perceived	dissimilarity	between	helper	and	target,	helping	is	more	likely	to	occur	as	a	function	of	perceived	cost	and	benefits	to	self.	The	only	exception,	it	seems,	is	when	there	is	a	question	as	to	whether	a	target	is	seen	to	be	a	typical,	or	atypical,	out-group	member,	and	under	these	circumstances	the	pattern	may	be	broken	(Manis,	Nelson,	&	Shedler,	1988).			In	sum,	an	integration	of	the	prosocial	helping	literature,	along	with	the	research	on	intergroup	processes,	suggests	that	different	processes	and	motivations	are	likely	to		125	drive	prosocial	responses	that	are	dependent	on	the	nature	of	the	in-group	vs.	out-group	relationship	between	helper	and	target,	and	perceptions	of	differences	or	similarities.			Finally,	in	study	4	I	wished	to	measure	empathy	to	the	cause	and,	further,	explore	its	relationship	with	the	other	variables.	While	study	3	found	that	empathy	was	not	predicted	by	inclusion	and	self-construal,	past	research	has	had	mixed	results,	and	findings	suggest	that	a	more	nuanced	relationship	between	empathy	and	prosocial	behaviour	may	exist	(Underwood	&	Moore,	1982).	Cialdini,	Brown,	Lewis,	Luce,	and	Neuberg	(1997),	in	their	exploration	of	the	mechanisms	by	which	empathy	works,	found	evidence	to	support	an	argument	that	empathetic	concern	only	increases	helping	behaviour	through	its	relation	to	perceived	oneness,	or	self-other	overlap.	They	provide	evidence	that	empathy	itself	does	not	increase	prosocial	behaviour,	when	perceived	oneness	is	eliminated,	and	have	gone	on	to	question	the	empathy-altruism	model,	as	a	result.	Cialdini	and	colleagues,	additionally,	offer	the	added	insight	that	if,	rather	than	empathy,	it	is	self-other-overlap	that	promotes	prosocial	behaviour,	then	helping	under	these	circumstances	would	not	be	selfless,	but	should	rather	be	seen	as	helping	directed	towards	the	self.				This	argument	(Cialdini	et	al.,	1997),	is	supported	by	other	research,	demonstrating	that	empathy	may	only	predict	helping	of	in-group	targets,	and	not	that	of	out-group	targets	(Sturmer	&	Snyder,	2010).	Specifically,	Sturmer	&	Snyder	(2010)	found	that	when	there	is	perceived	dissimilarity	between	helper	and	target,	empathy	was	less		126	likely	to	act	as	a	motivator	to	help.	In	study	3	I	had	offered	no	in-group	choice,	which	may	have	been	a	reason	for	empathy	not	being	significantly	predicted	in	any	of	my	models.	However,	in	study	4	I	was	offering	the	two	different	poster	frames	(“Animals	are	Similar”	vs.	“Animals	are	Different”),	and	based	on	past	findings	(Cialdini,	Brown,	Lewis,	Luce,	&	Neuberg,	1997;	Sturmer	&	Snyder,	2010),	my	prediction	was	that	empathy	ratings	would	be	higher	for	the	poster	that	emphasised	self-other	overlap.			I	also	wished	to	measure	gender	in	study	4,	expecting	it	again	to	predict	donation	intentions.	In	addition,	I	also	opted	to	measure	pet	ownership	status	in	my	demographics,	as	a	potential	participant	variable	that	might	impact	the	donation	and	empathy	dependent	variables,	since	pet	ownership	has	previously	been	found	to	correlate	with	higher	concern	for	other	animals	(Bowd,	1984;	Bjerk,	Østdahl,	&	Kleiven,	2003;	Pagani,	2011;	Paul	&	Serpell,	1993).			3.5.2	Procedure.	Study	four	took	the	form	of	a	2	(condition:	control	vs.	inclusion	manipulation)	x	2	(poster	frame:	animals	are	similar	vs.	animals	are	different)	mixed	experimental	design,	with	self-construal	measured	as	a	moderator	variable. A	convenience	sample	of	399	American	participants	(51%	female)	was	recruited,	using	Mechanical	Turk,	and	completed	the	study	in	exchange	for	monetary	compensation.		Condition	was	assigned	randomly	using	randomizing	embedded	coding	within	Qualtrics	survey	software.			127	The	sample	size	was	based	on	standard	calculations,	using	G*Power	3.1,	for	a-priori	multiple	regression	with	four	predictor	variables	(1	continuous,	2	categorical,	and	the	interaction),	using	random	sampling	at	95%	confidence	level.	Residual	variance	was	estimated	at	1-(R2)	=	0.85.	Variance	explained	by	special	effect	(self-construal	as	a	continuous	variable),	was	estimated	at	0.08.	Calculations	estimated	a	sample	size	of	120,	or	greater,	would	be	needed	for	each	regression	model,	in	order	to	achieve	a	power	of	>0.80.				As	with	the	previous	study,	participants	were	told	that	they	would	be	completing	a	series	of	assignments,	that	would	include	a	short	writing	skills	task,	as	well	as	a	visual	and	text	evaluation	task.	After	filling	in	a	number	of	trait	scales,	ostensibly	to	assess	personality,	participants	then	completed	one	of	two	versions	(control	vs.	inclusion)	of	the	same	writing	task	that	was	used	in	study	3,	and	has	been	previously	used	to	manipulate	social	inclusion	(e.g.	Arndt,	Greenberg,	&	Cook,	2002	,	Knowles	&	Garner,	2008;	Pickett,	Gardner,	&	Knowles,	2004;	Manner,	DeWall,	Baumeister,	&	Schaller,	2007;	Waytz	&	Epley,	2012).	Following	this	participants	were	shown	one	of	two	posters	(animals	are	similar	vs.	animals	are	different),	designed	to	frame	similarity	or	dissimilarity	to	the	charity.	After	viewing	each	poster	and	accompanying	text,	participants	were	asked	to	express	donation	intentions	towards	the	cause	in	the	same	3-item	measure	used	in	study	three,	as	well	as	to	state	how	much	empathy,	and	connection,	they	felt	for	the	cause.	Finally	they	were	asked	how	similar	they	felt	to	the	cause.	Participants	were	then	asked	how	donating	made	them	feel	on	a	personal	level.		128	Following	this,	participants	were	informed	that	they	had	completed	the	final	task	and	would	be	receiving	a	bonus	of	an	extra	50	cents	for	their	work.	They	were	advised	that	they	could	choose	to	keep	the	entire	bonus,	or	donate	part,	or	all,	of	it	to	the	cause	that	they	had	been	rating.		Demographics	were	then	collected,	and	lastly	participants	were	asked	which	writing	task	they	had	completed,	before	being	debriefed	and	thanked.	All	of	the	money	selected	by	participants	to	be	donated,	was	donated	to	the	BC	SPCA	at	the	close	of	the	study.		Materials.		Participants	were	asked	to	complete	the	same	24	item	self-construal	scale	as	used	in	Study	1,	2	and	3.	(scale	reliability:	interdependent	self-construal	α	=.850;	independent	self-construal	α	=	.796).			Following	this	a	16-item	Need	Fulfillment	scale	was	administered,	which	is	a	scale	adapted	by	Pfundmair,	Graupmann,	Frey	&	Aydin,	2015)	from	Zadro,	Williams,	and	Richardson,	(2004)	and	is	designed	to	measure	four	fundamental	needs	using	4-items	for	each	need.	Examples	and	scale	reliability	are	as	follows:	belonging	(e.g.	I	feel	poorly	accepted; α =. 855);	self-esteem	(e.g.	I	feel	others	fail	to	perceive	me	as	worthy	and	likeable;	α	=.797);	autonomy/control	(e.g.	I	feel	in	control	of	my	life; α		=.851);	and	meaningful	existence	(e.g.	I	feel	my	existence	is	meaningless;	α =.831).	Similar	to	the	Interpersonal	Needs	Questionnaire	(INQ),	of	Van	Orden,	Cukrowicz,	Witte	and	Joiner	Jr.,	(2012),	used	in	study	3,	the	Need	Fulfillment	scale	contains	questions	designed	to	measure	the	level	of	a	person’s	unmet	needs	and	each	statement	was		129	measured	using	a	7-point	scale	(1=not	at	all	true,	7=	very	true;	see	appendix	A.13	for	full	scale).	I	opted	to	utilize	this	scale,	in	preference	to	the	(INQ),	since	in	Study	4	I	wished	to	separate	and	identify	other	unmet	fundamental	needs,	in	addition	to	only	belonging.	My	prediction	was,	primarily,	that	higher	unmet	belonging	needs	would	correlate	with	higher	independent	self-construal.	This	expectation	is	based	on	past	research,	suggesting	that	independent	self-construal	may	be	associated	with	lower	sense	of	belonging	(Cushman,	1990;	Gardner,	Gabriel,	&	Lee,	1999;	Ren,	Wesselmann,	&	Williams,	2013).	Regarding	the	other	needs,	I	predicted	that	independent	self-construal	would	be	positively	correlated	with	self-esteem	and	autonomy,	in	line	with	previous	research	(Feather	&	McKee,	1993;	Gardner,	Gabriel,	&	Lee,	1999).			Participants	were	then	randomly	assigned	to	one	of	two	conditions:	control,	or	inclusion.	Dependent	on	condition,	participants	were	asked	to	complete	the	writing	task,	previously	used	in	study	three;	the	subject	of	which	participants	were	told	was	randomly	generated.	In	fact,	participants	were	either	told	to	“think	about	a	time	in	which	you	went	to	the	grocery	store“	(control	condition),	or	to	“think	about	a	time	in	which	you	felt	a	strong	sense	of	being	included	or	belonging”	(inclusion	condition).	As	before,	participants	were	asked	to	recall	the	event	as	vividly	as	possible,	then	to	write	a	paragraph	describing	the	past	event,	in	as	much	detail	as	they	were	able.			Following	the	manipulation,	participants	were	thanked	and	moved	on	to	the	next	task,	which	they	were	told	was	a	visual	and	text	evaluation	task	that	involved	assessing	a	charity	poster	appeal	for	a	British	Columbia	Society	for	Prevention	of		130	Cruelty	to	Animals	(BC	SPCA)	animal	shelter.		Participants	were	shown	the	same	poster	picture	as	in	study	3,	with	one	of	two	pieces	of	text	accompanying	it.	The	first	piece	of	text	was	designed	to	frame	feelings	of	similarity	(“Animals	are	Similar”)	and	began	“While	you	look	at	this	poster	please	consider	how	similar	animals	are	to	human	beings”.	The	second	piece	of	text	was	designed	to	frame	feelings	of	dissimilarity	(“Animals	are	Different”)	and	began	“While	you	look	at	this	poster	please	consider	how	different	animals	are	to	human	beings”.	See	appendix	A.14	for	picture	and	full	text	of	both.		After	viewing	each	poster	and	accompanying	text,	participants	were	asked	to	express	donation	intentions	towards	the	cause	in	the	same	3-item	measure	used	in	study	three	(scale	reliability:	α	=	.952),	as	well	as	to	state	how	much	empathy	and	connection	they	felt	for	the	cause	on	a	7-point	scale	(1=not	at	all	true,	7=	very	much).	Following	this,	they	were	asked	how	similar	they	felt	to	the	cause	on	a	7-point	scale	(1=not	at	all	true,	7=	very	much).	Lastly,	participants	were	asked	answer	how	donating	makes	them	feel	and	why	they	do	it	in	an	open	ended	response.			The	final	task	for	the	participants	was	to	select	how	much,	if	any,	of	a	50	cent	bonus	they	wished	to	allot	to	the	charity	cause,	using	a	slider	scale.	Demographics	were	collected,	which	included	income	and	pet	ownership,	participants	were	asked	to	recall	which	writing	task	they	had	completed	as	a	manipulation	check,	and	then	debriefed	and	thanked.					131	3.5.3	Results.	My	study	exclusion	policy	stipulated	that	participants	would	be	screened	out	if	they	failed	all	attention	checks,	or	if	they	accurately	guessed	the	key	study	hypotheses.	Nineteen	participants	failed	all	attention	checks,	and	were	removed	from	the	study,	leaving	n=380	remaining.			Power.	Post-hoc	power	analyses,	using	G*Power,	detected	sufficient	power	(>.80)	to	detect	correlations	of	r(378)=.150	and	higher,	effect	sizes	of	d=0.29	and	higher	for	independent	samples	t-tests	for	gender	(group	sizes:	186,191)	at	95%	Confidence.	See	appendix	A.17	for	power	analysis.			Manipulation	check.		To	assess	whether	the	manipulation	had	been	effective,	participants	were	asked	to	recall	the	writing	task	they	received,	in	a	multiple-choice	question	at	the	end	of	the	study	after	demographics	were	taken.	When	asked	to	recall	their	writing	task	97.3%	of	participants	were	able	to	do	so	with	complete	accuracy.						Demographics.	Demographics	collected	showed	that	participants	in	the	sample	identified	as	78.9%	Caucasian,	6.1%	Black,	4.2%	Latin	American,	4.2%	Asian,	0.5%	South	East	Asian,	and	6.1%	mixed	other	(see	figure	21).	Results	also	showed	that	94.2%	of	participants	reported	having	lived	in	the	USA	for	over	21	years.	Participants		132	identified	as	49.2%	female.	Age	demographics	were	as	follows:	1.8%	under	21	years	old;	32.4%	21-30	years	old;	34.2%	31-40	years	old;	17.9%	41-50	years	old;	8.9%	51-60	years	old;	and	4.7%	over	60	years	old	(see	figure	22).	In	terms	of	pre-tax	income	20.8%	reported	under	$20,000,	24.2%	reported	$20,000-$39,999,	23.2%	reported	$40,000-$59,999,	13.7%	reported	$60,000-$79,999,	8.2%	reported	$80,000-$99,999	and	10%	reported	over	$99,999.			Demographic,	trait	and	other	correlations.	Results	of	study	4	revealed	that	independent	self-construal	was	positively	correlated	with	the	following	(INQ)	individual	needs:	self-esteem	(r(378)=.472,	p	<.001);	belonging	(r(378)=.403,	p	<.001);	control	(r(378)=.453,	p	<.001);	and	meaningful	existence	(r(378)=.416,	p	<.001).	Independent	self-construal	was	also	negatively	correlated	with	loneliness	(r(378)=-.289,	p	<.001).	Interdependent	self-construal	was	not	correlated	with	self-esteem	(r(378)=.039,	p	=.45)	or	meaningful	existence	(r(378)=.082,	p	=.11)	and	only	moderately	with	belonging	(r(378)=.180,	p	<.001)	and	control	(r(378)=.142,	p	=.006).	Interdependent	self-construal	was	not	correlated	with	loneliness	(r(378)=-.06,	p	=.22).	Unmet	belonging	needs	measured	in	the	INQ	were	also	negatively	correlated	with	Donation	intentions	(r(378)=-.129,	p=.01),	Connection	to	Cause	(r(378)=-.119,	p=.02)	,	and	Empathy	to	Cause	(r(378)=-.157,	p=.002).	Put	another	way,	the	more	you	felt	your	belonging	needs	were	met,	the	higher	your	expressed	donation	intentions,	connection	to	cause,	and	empathy	to	cause,	as	well	as	independent	self-construal.	My	prediction	had	been	that	higher	unmet	belonging	needs,	specifically,	would	correlate	with	high	independent	self-	133	construal	based	on	past	research,	but	my	research	results	did	not	support	this	finding.	Furthermore,	I	predicted	that	higher	unmet	belonging	needs	would	attenuate	responses	in	the	inclusion	condition.	Controlling	for	unmet	belonging	did	improve	prediction	in	the	main	regressions,	when	unmet	belonging	was	entered	as	a	covariate.	The	conditional	effect	of	unmet	belonging	was	also	marginally	significant	at	predicting	donation	intentions	on	its	own.	See	regression	analyses	in	appendix	A.15	for	further	details.		In	terms	of	donation	measures	the	three-item	donation	intentions	score	was	found	to	be	significantly	correlated	with	the	actual	cash	donation	measure	(r(377)=.368,	p	<.001).		Means	comparisons.	An	independent	samples	t-test	was	also	carried	out	to	test	for	a	difference	between	genders	as	separate	analysis.	Results	showed	that	females	(n=192)	(M=14.75,	SD=5.43)	indicated	significantly	higher	donation	intentions	than	males	(n=187)		(M=12.11,	SD=5.65),	t(377)=4.636,	p<.001,	d=	0.476,		as	well	as	higher	connection	to	cause	ratings	(females	M=5.01,	SD=1.87;	males	M=4.08,	SD=1.90)	,	t(377)=4.785,	p<.001,	d=	0.490	higher	similarity	to	cause	ratings	(females	M=4.53,	SD=2.05;	males	M=3.76,	SD=1.81)	,	t(374)=3.886,	p<.001,	d=	0.398	and	higher	empathy	to	cause	ratings	(females	M=5.86,	SD=1.54;	males	M=5.24,	SD=1.66)	,	t(373)=3.789,	p<.001,	d=	0.389.	They	also	made	higher	(if	not	significantly)	actual	cash	(cents)	donations	(females	M=15.60,	SD=19.08;	males	M=12.45,	SD=17.56),	t(373)=1.672,	p=.095,	d=	0.172.	I	therefore	carried	out	the	primary	regression	models	controlling	for	gender.				134	An	independent	samples	t-test	was	also	carried	out	to	test	for	a	difference	between	pet	ownership	vs.	non-ownership	as	a	separate	analysis.	Pet	owners	significantly	differed	on	donation	intentions,	connection	to	cause,	similarity	to	cause	and	empathy	for	cause	ratings	(although	not	as	strongly	as	gender),	and	differed	with	near	significance	on	actual	cash	donation.	See	appendix	A.15	for	detailed	analysis.		Regression	tests	for	moderation.	Moderated	multiple	regressions	were	carried	out	in	SPSS,	using	process	v.3	(Hayes,	2018)	to	examine	a)	whether	poster	style	(Animals	are	Similar	vs.	Animals	are	Different)	influenced	participant	ratings	of	connection	to	cause,	similarity	to	cause,	and	empathy	to	cause,	as	well	as	donation	intentions,	and	actual	cash	donation,	and	b)	whether	these	effects	were	moderated	by	the	manipulation	(inclusion	vs.	control	condition)	and	independent	self-construal	orientation.	In	consideration	of	the	fact	that,	based	on	previous	research	and	the	results	of	study	3,	interdependent	self-construal	was	not	expected	to	predict	any	of	the	primary	dependent	variables	it	was	not	entered	into	any	of	the	main	regression	models	of	study	4.	One	exception	was	made	to	investigate	whether	interdependent	self-construal	interacted	with	poster	type	to	predict	donation	intentions	in	the	“Animals	are	Similar”	vs.	“Animals	are	Different”	frame.				Following	Cohen	and	Cohen	(1983)	and	Wendorf	(2004),	I	dummy	coded	the	condition	for	the	inclusion	manipulation	(vs.	control),	as	well	as	for	the	poster	style	different	(vs.	similar).	Independent	self-construal	was	mean	centered,	in	accordance		135	with	guidelines,	stipulating	that	this	practice	renders	subsequent	tests	of	hypotheses	and	regression	coefficients	for	X	and	M	more	meaningful	and	substantially	interpretable,	as	well	as	to	reduce	the	likelihood	of	errors	in	interpretation	(Hayes,	2013).	All	dependent	measures	were	subjected	to	separate	moderated	multiple	regressions	with	each	manipulation	condition	and	the	control,	independent	self-construal	scores,	and	their	interactions	simultaneously	entered	as	predictor	variables.	As	in	the	previous	studies,	unstandardized	coefficients	(B)	are	reported	rather	than	betas	(β).	Because	βs	are	not	properly	standardized	in	interaction	terms	they	are	not	interpretable,	whereas	B	represents	the	difference	between	the	un-weighted	means	of	the	groups	involved	(see	Cohen	et	al.	2003).			3.5.3.1	Role	of	inclusion	manipulation,	poster	frame	and	interdependent	self-construal	in	donation	intentions.	Summary	of	results.	In	a	regression	model	(model	3;Hayes,	2017)	with	interdependent	self-construal,	as	expected	the	inclusion	manipulation	did	not	significantly	predict	donation	intentions	in	terms	of	a	conditional	effect.	Neither	did	interdependent	self-construal	predict	donation	intentions,	nor	the	poster	frame.	There	was	no	significant	three-way	interaction	between	interdependent	self-construal,	poster	frame	and	the	inclusion	manipulation,	which	again	was	as	predicted.	Of	interest,	however,	was	the	significant	two-way	interaction	between	interdependent	self-construal,	and	poster	frame.				136	In	a	two-way	regression	model	(model	1;Hayes,	2017)	with	interdependent	self-construal,	neither	interdependent	self-construal,	nor	the	poster	frame	predicted	donation	intentions.	However,	there	was	a	significant	two-way	interaction	between	interdependent	self-construal,	and	poster	frame.	As	predicted,	whether	donation	intentions	were	greater	depended	on	the	poster	frame,	as	well	as	participant’s	interdependent	self-construal,	with	donation	intentions	related	to	interdependent	self-construal	only	in	the	“Animals	are	Similar”	poster	frame	and	not	in	the	“Animals	are	Different”	poster	frame.	See	detailed	analysis	following.			Regression	model	of	donation	intentions.		I	submitted	the	donation	intentions	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	interdependent	self-construal	and	Poster	type	entered	as	independent	variables.	The	regression	model	in	total	was	highly	significant	R2	=	.047,	p	<.001,	F(3,	376)=	6.205.	I	found	no	conditional	effect	for	the	poster	type	B=-0.532,	SE=0.573,	t(380)=-0.930,	p=.35,	95%	CI	[-1.66,	0.59]	nor	for	the	interdependent	self-construal	measure	on	its	own	B=0.035,	SE=0.035,	t(380)=1.005,	p=.32,	95%	CI	[-0.03,	0.10.	There	was,	however,	a	significant	interaction	between	poster	type	and	interdependent	self-construal	B=0.105,	SE=0.049,	t(380)=2.161,	p=.03,	95%	CI	[0.01,	0.20].	The	two-way	interaction	showed	that	individuals	with	higher	interdependent	self-construal	indicated	higher	donation	intentions	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	A	test	of	highest	order	unconditional	interaction	produced	an	ΔR2	of	.012	F(1,376)	=	4.670,	p=	.03	as	a	result	of	the	two-way	interaction.			137	A	test	of	the	conditional	interaction	(Poster	Style	X	interdependent	self-construal)	revealed	that	the	effect	on	donation	intentions	was	significant	at	the	lower	values	of	interdependent	self-construal	(θXàY	l	W=	Interdependent	SC-11.67	=-1.7602,	SE=	0.818,	p=.03)	but	not	significant	at	the	higher	levels	(θXàY	l	W=	Interdependent	SC12.37		=0.770,	SE=0.819,	p=.35).		See	figure	23	for	a	visualization	with	the	Johnson-Neyman	Point	showing	the	region	of	interdependent	self-construal	values	(filled	area	below	-8.1141)	for	which	a	floodlight	test	would	reveal	significant	differences	between	the	two	groups,	and	figure	24	for	a	visualization	of	the	corresponding	floodlight	analysis	with	confidence	intervals.		3.5.3.2	Role	of	inclusion	manipulation	(vs.	control),	poster	frame	and	independent	self-construal	in	donation	intentions.	Summary	of	results.	Following	the	procedures	previously	outlined,	I	regressed	donation	intentions	on	the	dummy	(control	vs.	inclusion)	and	moderator	variables.		In	a	regression	model	with	independent	self-construal,	the	inclusion	manipulation	did	not	significantly	predict	donation	intentions	in	terms	of	a	conditional	effect,	neither	did	independent	self-construal,	nor	the	poster	frame.	Of	interest,	however,	was	the	significant	three-way	interaction	between	independent	self-construal,	poster	frame	and	the	inclusion	manipulation.	As	predicted,	and	in	line	with	study	3,	whether	donation	intentions	were	greater	for	the	inclusion	condition,	over	the	control	condition,	depended	on	the	poster	frame,	as	well	as	participant’s	independent	self-construal.		In	view	of	the	result	of	the	t-test,	and	my	prediction	that	gender	would	significantly	predict		138	donation	intentions,	I	produced	a	second	regression	model	that	included	gender	as	a	covariate.	This	model	was	significant	and	increased	significance	in	the	majority	of	the	conditional	and	interaction	effects.			Regression	model	of	donation	intentions.		I	first	submitted	the	donation	intentions	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	not	significant	R2	=	.031,	p	=	.11,	F(7,	372)=	1.708.	I	found	no	conditional	effect	of	the	inclusion	condition	B=0.104,	SE=0.837,	t(380)=0.124,	p=.90,	95%	CI	[-1.54,	1.75]	nor	for	independent	self-construal	measure	on	its	own	B=-0.086,	SE=0.054,	t(380)=-1.604,	p=.11,	95%	CI	[-0.19,	0.02],	or	for	poster	type	on	its	own	B=-0.344,	SE=0.804,	t(380)=-0.428,	p=.67,	95%	CI	[-1.93,	1.24].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.214,	SE=0.077,	t(380)=2.770,	p=.01,	95%	CI	[0.06,	0.37].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.0001,	SE=1.167,	t(380)=-0.0001,	p=1.00,	95%	CI	[-2.29,	2.29].	However,	there	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.187,	SE=0.077,	t(380)=2.428,	p=.02,	95%	CI	[0.04,	0.34],	and	most	interestingly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.317,	SE=0.113,	t(380)=-2.815,	p=.01,	95%	CI	[-0.54,	-0.10].				The	three-way	interaction	showed	that	in	the	control	condition	individuals	with	high	independent	self-construal	indicated	significantly	higher	donation	intentions		139	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	individuals	with	high	independent	self-construal	preferred	the	“Animals	are	Different”	to	us	text.		A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.021	F(1,372)	=	7.924,	p=	.01	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal	revealed	that	the	effect	on	donation	intentions	was	significant	with	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-3.3836,	p=.04)	and	low	self-construal	(θXWàY	l	Z=	-10.33	=3.2750,	p=.05).		 Covariate	analyses.	A	regression	model	was	then	produced	entering	in	gender	as	covariate.	Gender	was,	as	predicted,	a	strong	predictor	of	donation	intentions.	See	results	below.			Regression	model	of	donation	intentions	controlling	for	gender.	This	time	controlling	for	gender	I	once	more	submitted	the	donation	intentions	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	significant	R2	=	.094,	p	<.001,	F(8,	371)=	4.820.	I	found	no	conditional	effect	of	the	inclusion	condition	B=0.444,	SE=0.813,	t(380)=0.546,	p=.59,	95%	CI	[-1.16,	2.04]	nor	for	poster	type	on	its	own	B=-0.076,	SE=0.780,	t(380)=-0.097,	p=.92,	95%	CI	[-1.61,	1.46].	There	was,	however,	a	significant	conditional	effect	of	the	independent	self-construal	measure	on	its	own	B=-0.108,	SE=0.052,	t(380)=-2.077,	p=.04,	95%	CI		140	[-0.21,	-0.01],.	There	was	also	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.249,	SE=0.075,	t(380)=3.314,	p=.001,	95%	CI	[0.10,	0.40].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.432,	SE=1.133,	t(380)=-0.382,	p=.70,	95%	CI	[-2.66,	1.80].	There	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.200,	SE=0.075,	t(380)=2.680,	p=.008,	95%	CI	[0.05,	0.35],	and	most	importantly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.355,	SE=0.109,	t(380)=-3.252,	p=.001,	95%	CI	[-0.57,	-0.14].		There	was	also	a	significant	conditional	effect	found	for	gender	B=-2.852,	SE=0.561,	t(380)=-5.080,	p<001,	95%	CI	[-3.96,	-1.75].				The	three-way	interaction	showed	that	(as	before	without	gender)	in	the	control	condition	individuals	with	high	independent	self-construal	indicated	significantly	higher	donation	intentions	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	individuals	with	high	independent	self-construal	preferred	the	“Animals	are	Different”	to	us	text.		A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.0258,	F(1,371)	=	10.575,	p=	.001,	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal	revealed	that	the	effect	on	donation	intentions	was	significant	with	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-4.2259,	p=.01)	and	low	self-construal	(θXWàY	l	Z=	-10.33	=3.2398,	p=.04).	See	appendix	figure	25	for	a	visual		141	of	the	moderated	moderation	model,	controlling	for	gender	and	figure	26	for	a	visual	of	the	corresponding	floodlight	analysis	with	JN	points	and	confidence	bands.		Probing	the	conditional	effects	of	the	focal	predictor,	at	values	of	the	moderators,	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=	-2.1394,	SE=	1.0672,	t(380)=	-2.0046,	p=.05,	95%	CI	[-4.24,	-0.04],	approaching	significant	in	the	Control	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=	2.0558,	SE=	1.427,	t(380)=	1.7991,	p=.07,	95%	CI	[-0.19,	4.30],	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=	1.1004,	SE=1.1859,	t(380)=	0.9279,	p=.35,	95%	CI	[-1.23,	3.43],	and	approaching	significant	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67)	Effect=	-2.1701,	SE=	1.1569,	t(380)=	-1.857	p=.06,	95%	CI	[-4.45,	0.11].			Covariate	analyses.	A	regression	model	was	also	produced,	entering	in	pet	ownership	as	a	covariate,	and	another	model	produced,	with	unmet	belonging	as	a	covariate.	Pet	ownership	predicted	donation	intentions,	but	less	strongly	than	gender.	Likewise,	unmet	belonging	predicted	donation	intentions,	but	again	not	as	strongly	as	gender.	See	appendix	A.15	for	pet	ownership	and	unmet	belonging	results.			3.5.3.3	Role	of	inclusion	manipulation	(vs.	control),	poster	frame	and	independent	self-construal	in	actual	cash	donation.		142	Summary	of	results.	In	a	regression	model	with	independent	self-construal,	the	inclusion	manipulation	did	not	significantly	predict	actual	cash	donation,	in	terms	of	a	conditional	effect,	neither	did	independent	self-construal,	nor	the	poster	frame.	Likewise,	there	was	no	significant	three-way	interaction	between	independent	self-construal,	poster	frame	and	the	inclusion	manipulation.	In	view	of	the	result	of	the	t-test,	and	my	prediction	that	gender	would	significantly	predict	cash	donation,	I	produced	a	second	regression	model	that	included	gender	as	a	covariate.	This	delivered	no	significant	effects	or	interactions,	although	significance	was	approaching	at	times.	A	visualisation	(figure	27.)	shows	that	the	patterns	were	in	the	predicted	direction	with	individuals	who	were	high	in	independent	self-construal	showing	higher	actual	cash	donation	for	a	Similar	Poster	than	for	a	Different	Poster	under	normal	(control)	conditions,	with	a	reversal	of	the	pattern	following	an	inclusion	manipulation.			Regression	model	for	actual	cash	donation.	I	submitted	the	actual	cash	donation	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	not	significant	R2	=.015,	p	=	.57,	F(7,	371)=	0.817.	I	found	no	conditional	effect	of	the	inclusion	condition	B=2.883,	SE=2.732,	t(379)=1.055,	p=.29,	95%	CI	[-2.49,	8.26]	nor	for	independent	self-construal	measure	on	its	own	B=-0.173,	SE=0.175,	t(379)=-0.990,	p=.32	95%	CI	[-0.52,	0.17].	There	was	a	marginally	significant	effect	for	poster	type	on	its	own	B=4.947,	SE=2.617,	t(380)=1.890,	p=.06,	95%	CI	[-0.20,	10.09].	There	was	no	significant	interaction	between	inclusion	and	independent	self-construal	B=0.191,	SE=0.251,		143	t(379)=0.759,	p=.45,	95%	CI	[-0.30,	0.69].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-6.280,	SE=3.803,	t(379)=-1.651,	p=.10,	95%	CI	[-13.76,	1.20].	Moreover,	there	was	no	significant	interaction	between	poster	type	and	independent	self-construal	B=0.358,	SE=0.251,	t(379)=1.431,	p=.15,	95%	CI	[-0.13,	0.85],	and	no	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.423,	SE=0.367,	t(379)=-1.154,	p=.25,	95%	CI	[-1.14,	0	.30].	A	test	of	higher	order	unconditional	interaction	produced	a	non-significant	ΔR2	of	.004,	F(1,371)=1.331,	p=.25	as	a	result	of	the	three-way	interaction.			Regression	model	for	actual	cash	donation	controlling	for	gender.	This	time	controlling	for	gender,	I	submitted	the	actual	cash	donation	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	not	significant	R2	=.026,	p	=.29,	F(8,	370)=1.209.	I	found	no	conditional	effect	of	the	inclusion	condition	B=3.351,	SE=2.732,	t(379)=1.227,	p=.22,	95%	CI	[-2.02,	8.72],	nor	for	independent	self-construal	measure	on	its	own	B=-0.202,	SE=0.175,	t(379)=-1.156,	p=.25,	95%	CI	[-0.55,	0.14],	was	just	significant	for	poster	type	on	its	own	B=5.297,	SE=2.612,	t(379)=2.027,	p=.04,	95%	CI	[0.16,	10.43].	There	was	a	marginally	significant	conditional	effect	found	for	gender	B=-3.724,	SE=1.883,	t(379)=-1.978,	p=.05,	95%	CI	[-7.43,	-0.02].		There	was	no	significant	interaction	between	inclusion	and	independent	self-construal	B=0.236,	SE=0.251,	t(379)=0.940,	p=.35,	95%	CI	[-0.26,	0.73].	There	was	an	approaching	significant	interaction	between	inclusion	and	poster	type	B=-6.868,	SE=3.800,	t(379)=-1.807,	p=.07,	95%		144	CI	[-14.34,	0.60].	There	was	no	significant	interaction	between	poster	type	and	independent	self-construal	B=0.375,	SE=0.250,	t(379)=1.503,	p=.13,	95%	CI	[-0.12,	0.87],	and	no	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.473,	SE=0.366,	t(379)=-1.292,	p=.20,	95%	CI	[-1.19,	0.25].	A	test	of	higher	order	unconditional	interaction	produced	a	non-significant	ΔR2	of	.004,	F(1,370)	=	1.670,	p=	.20	as	a	result	of	the	three-way	interaction.	See	Figure	27.	for	a	visual	representation	controlling	for	gender.			3.5.3.4	Role	of	inclusion	manipulation	(vs.	control),	poster	frame	and	independent	self-construal	in	connection	to	cause.	Following	the	procedures	previously	outlined,	I	regressed	connection	to	cause	on	the	dummy	(control	vs.	inclusion)	and	moderator	variables.			Summary	of	results.	In	a	regression	model	with	independent	self-construal,	the	inclusion	manipulation	did	not	significantly	predict	connection	to	cause	in	terms	of	a	conditional	effect,	neither	did	independent	self-construal,	nor	the	poster	frame.	Of	interest,	however,	was	the	significant	three-way	interaction	between	independent	self-construal,	poster	frame	and	the	inclusion	manipulation.	As	predicted,	and	in	line	with	the	donation	intentions	models,	whether	connection	to	cause	were	greater	for	the	inclusion	condition,	over	the	control	condition,	depended	on	the	poster	frame,	as	well	as	participant’s	independent	self-construal.		In	view	of	the	results	of	the	t-test	and	previous	regression	models,	I	produced	a	second	regression	model	that		145	included	gender	as	a	covariate.	This	model	was	a	better	predictor	and	there	was	increased	significance	in	the	majority	of	the	conditional	and	interaction	effects,	as	a	result	of	adding	gender.			Regression	model	for	connection	to	cause.	I	next	submitted	the	connection	to	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	approaching	significant	R2	=.034,	p	=	.08,	F(7,	372)=	1.857.	Once	again	I	found	no	conditional	effect	of	the	inclusion	condition	B=0.207,	SE=0.284,	t(380)=0.729,	p=.47,	95%	CI	[-0.35,	0.77]	nor	for	independent	self-construal	measure	on	its	own	B=-0.030,	SE=0.018,	t(380)=-1.620,	p=.11,	95%	CI	[-0.07,	0.01],	or	for	poster	type	on	its	own	B=0.238,	SE=0.273,	t(380)=0.872,	p=0.38,	95%	CI	[-0.30,	0.77].	There	was	however	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.073,	SE=0.026,	t(380)=2.768,	p=.01,	95%	CI	[0.02,	0.12].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.191,	SE=0.396,	t(380)=-0.484,	p=.63,	95%	CI	[-0.97,	0.59].	However,	there	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.068,	SE=0.026,	t(380)=2.591,	p=.01,	95%	CI	[0.02,	0.12],	and	most	interestingly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.113,	SE=0.038,	t(380)=-2.955,	p=.003,	95%	CI	[-0.19,	-0.04].		The	three-way	interaction	showed	that	in	the	control	condition	individuals	with	high	independent	self-construal	indicated	significantly	higher	connection	to	cause	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion		146	condition	individuals	with	high	independent	self-construal	preferred	the	“Animals	are	Different”	to	us	text.	A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.023,	F(1,372)=	8.7318,	p=.003	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal,	revealed	that	the	effect	on	donation	intentions	was	significant	with	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-1.3959,	p=.02)	and	near	significant	with	low	self-construal	(θXWàY	l	Z=	-10.33	=0.9746,	p=.08).			Probing	the	conditional	effects	of	the	focal	predictor	at	values	of	the	moderators,	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=0.959,	SE=	0.3993,	t(380)=2.4021,	p=.02,	95%	CI	[0.17,	1.74],	but	not	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	-0.4607,	SE=	0.3737,	t(380)=-1.2328,	p=.22,	95%	CI	[-1.20,	0.27],	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	0.5139,	SE=.4153,	t(380)=1.2373,	p=.22,	95%	CI	[-0.30,	1.33],	nor	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	-0.4369,	SE=0.4041,	t(380)=-1.0811,	p=.28,	95%	CI	[-1.23,	0.36].			Regression	model	for	connection	to	cause	controlling	for	gender.		I	submitted	the	connection	to	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018)	controlling	for	gender.	The	regression	model	in	total	was	significant	R2	=	.098,	p	<	.001,	F(8,	371)=	5.053.	Once	again	I	found	no	conditional	effect	of	the	inclusion		147	condition	B=0.324,	SE=0.276,	t(380)=1.176,	p=.24,	95%	CI	[-0.22,	0.87].	However,	there	was	a	significant	conditional	effect	for	independent	self-construal	measure	on	its	own	B=-0.037,	SE=0.018,	t(380)=-2.100,	p=.04,	95%	CI	[-0.07,	-0.002].	I	found	no	conditional	effect	for	poster	type	on	its	own	B=0.330,	SE=0.264,	t(380)=1.247,	p=.21,	95%	CI	[-0.19,	0.85].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.085,	SE=0.025,	t(380)=3.322,	p=.001,	95%	CI	[0.04,	0.14],	and	between	independent	self-construal	and	poster	type	B=0.072,	SE=0.025,	t(380)=2.853,	p=.005,	95%	CI	[0.02,	0.12].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.340,	SE=0.384,	t(380)=-0.886,	p=0.38,	95%	CI	[-1.10,	0.42].	Lastly	there	was	a	significant	three-way	interaction	between	poster	type	and	independent	self-construal	B=-0.126,	SE=0.037,	t(380)=-3.404,	p<.001,	95%	CI	[-0.20,-0.05].		There	was	a	significant	conditional	effect	for	gender	on	its	own	B=-0.980,	SE=0.190,	,t(380)=-5.151,	p<.001,	95%	CI	[-1.35,	-0.61].	The	three-way	interaction	showed	that	in	the	control	condition,	individuals	with	high	independent	self-construal	indicated	significantly	higher	donation	intentions	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	individuals	with	high	independent	self-construal	preferred	the	“Animals	are	Different”	to	us	text.		A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.028,	F(1,371)	=	11.5859,	p<	.001	as	a	result	of	the	three-way	interaction.				A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal	revealed	that	the	effect	on	donation	intentions	was		148	significant	with	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-1.6853,	p=.002)	and	near	significant	with	low	self-construal	(θXWàY	l	Z=	-10.33	=0.9625,	p=.08).		See	figure	28	for	a	visual	showing	the	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	connection	to	cause	(Y)	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender,	and	figure	29	for	a	visual	of	the	corresponding	floodlight	analysis,	with	JN	points	and	confidence	bands.				Probing	the	conditional	effects	of	the	focal	predictor	at	values	of	the	moderators	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67)	Effect=1.0987,	SE=0.3872,	t(380)=	2.8376	p=.005,	95%	CI	[0.34,	1.86],	but	not	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=0.4147,	SE=0.3616	,	t(380)=	-1.1467	p=.25,	95%	CI	[-1.13,	0.30],	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=0.5478,	SE=	0.4018,	t(380)=	1.3634	p=.17,	95%	CI	[-0.24,	1.34],	and	not	significant	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=-0.5866,	SE=0.392,	t(380)=	-1.4965	p=.14,	95%	CI	[-1.36,	0.18].			3.5.3.5	Role	of	inclusion	manipulation	(vs.	control),	poster	frame	and	independent	self-construal	in	empathy	for	cause.		149	Following	the	procedures	previously	outlined,	I	regressed	empathy	for	cause	on	the	dummy	(control	vs.	inclusion)	and	moderator	variables.			Summary	of	results.	In	a	regression	model	with	independent	self-construal,	the	inclusion	manipulation	did	not	significantly	predict	empathy	for	cause	in	terms	of	a	conditional	effect,	neither	did	independent	self-construal,	nor	the	poster	frame.	Of	interest	however,	was	the	small	but	significant	three-way	interaction	between	independent	self-construal,	poster	frame	and	the	inclusion	manipulation.	As	predicted,	and	in	line	with	the	donation	intentions	and	connection	to	cause	models,	whether	empathy	for	cause	were	greater	for	the	inclusion	condition,	over	the	control	condition,	depended	on	the	poster	frame,	as	well	as	participant’s	independent	self-construal.		In	view	of	the	results	of	the	t-test	and	previous	regression	models,	I	produced	a	second	regression	model	that	included	gender	as	a	covariate.	This	model	was	a	better	predictor	and	there	was	increased	significance	in	the	majority	of	the	conditional	and	interaction	effects,	as	a	result	of	adding	gender.			Regression	model	for	empathy	for	cause.	I	next	submitted	the	empathy	for	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	significant	R2	=.039,	p	=.04,	F	(7,	372)=	1.857.	Once	again,	I	found	no	conditional	effect	of	the	inclusion	condition	B=-0.054,	SE=0.239,	t(380)=-0.225,	p=.82,	95%	CI	[-0.52,	0.42]	nor	for	independent	self-construal	measure	on	its	own	B=-0.013,	SE=0.015,	t(380)=-0.845,	p=.40,	95%	CI		150	[-0.04,	0.02],	or	for	poster	type	on	its	own	B=0.019,	SE=0.229,	t(380)=0.081,	p=.94,	95%	CI	[-0.43,	0.47].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.054,	SE=0.022,	t(380)=2.437,	p=.02,	95%	CI	[0.01,	0.10].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=0.187,	SE=0.333,	t(380)=0.564,	p=.57,	95%	CI	[-0.47,	0.84].	However,	there	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.053,	SE=0.022,	t(380)=2.418,	p=.02,	95%	CI	[0.01,	0.10],	and	most	interestingly,	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.081,	SE=0.032,	t(380)=-2.516,	p=.01,	95%	CI	[-0.14,	-0.02].		The	three-way	interaction	showed	that	in	the	control	condition	individuals	with	high	independent	self-construal	indicated	significantly	higher	empathy	for	cause	ratings	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	the	Poster	Style	made	little	difference	to	empathy	for	cause	ratings.		A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.016	F(1,372)	=	6.3312,	p=	.01	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal	revealed	that	the	effect	on	empathy	ratings	was	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=1.0222,	p=.03)	but	not	significant	with	high	self-construal	(θXWàY	l	Z=	10.67	=-0.675,	p=.16).			151	Probing	the	conditional	effects	of	the	focal	predictor	at	values	of	the	moderators	revealed	the	effect	to	be	near	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=0.5848,	SE=0.3357,	t	(380)=1.7421	p=.08,	95%	CI	[-0.08,	1.25],	and	near	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=-0.5294,	SE=0.3142,	t	(380)=-1.685	p=.09,	95%	CI	[-1.15,	0.09],	but	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=0.4928,	SE=0.3492,	t	(380)=1.4111	p=.16,	95%	CI	[-0.19,	1.18],	nor	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=-0.0902,	SE=0.3398,	t	(380)=-0.2654	p=.79,	95%	CI	[-0.76,	0.58].			Regression	model	for	empathy	to	cause	controlling	for	gender.		I	submitted	the	empathy	for	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018)	controlling	for	gender.	The	regression	model	in	total	was	significant	R2	=.079,	p	<.001,	F	(8,	371)=	3.996.	Once	again	I	found	no	conditional	effect	of	the	inclusion	condition	B=0.025,	SE=0.235,	t(380)=0.105,	p=.92,	95%	CI	[-0.44,	0.49]	nor	for	independent	self-construal	measure	on	its	own	B=-0.018,	SE=0.015,	t(380)=-1.199,	p=.23,	95%	CI	[-0.05,	0.01],	or	for	poster	type	on	its	own	B=0.080,	SE=0.225,	t(380)=0.357,	p=.72,	95%	CI	[-0.36,	0.52].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.062,	SE=0.022,	t(380)=2.848,	p=.01,	95%	CI	[0.02,	0.10].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=0.088,	SE=.327,	t(380)=0.269,	p=.79,	95%	CI	[-0.56,	0.73].	However,	there	was	a	significant	interaction	between	poster	type	and		152	independent	self-construal	B=0.056,	SE=0.022,	t(380)=2.605,	p=.01,	95%	CI	[0.01,	0.10],	and	most	interestingly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.090,	SE=0.032,	t(380)=-2.842,	p=.01,	95%	CI	[-0.14,	-0.03].		I	found	a	conditional	effect	of	gender	on	its	own	B=-0.657,	SE=0.162,	t(380)=4.053,	p<.001,	95%	CI	[-0.98,	-0.34].	The	three-way	interaction	showed	that	in	the	control	condition	individuals	with	high	independent	self-construal	indicated	significantly	higher	empathy	for	cause	ratings	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	the	Poster	Style	made	little	difference	to	empathy	for	cause	ratings.	A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.020,	F(1,371)	=	8.0777,	p=	.005	as	a	result	of	the	three-way	interaction.			See	figure	30	for	visual	of	the	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	empathy	for	cause	(Y)	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender,	and	figure	31	for	a	corresponding	visual	of	the	JN	points	along	with	confidence	bands.		A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	empathy	for	cause	was	near	significant	at	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-0.8689,	p=.07)	and	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=1.0141,	p=.03).				153	Probing	the	conditional	effects	of	the	focal	predictor	at	values	of	the	moderators	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67)	Effect=0.6784,	SE=0.3298,	t(380)=	2.0572,	p=.04,	95%	CI	[0.03,	1.33],	but	not	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=-0.4986,	SE=0.3080,	t(380)=	-1.6188,	p=.11,	95%	CI	[-1.10,	0.11],	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33)	Effect	=0.5155,	SE=	0.3422,	t(380)=	1.5063,	p=.13,	95%	CI	[-0.16,	1.19],	and	not	significant	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67)	Effect	=-0.1905,	SE=0.3339,	t(380)=-0.5707,	p=.57,	95%	CI	[-0.85,	0.47].			3.5.3.6	Role	of	inclusion	manipulation	(vs.	control),	poster	frame	and	independent	self-construal	in	similarity	to	cause.	Following	the	procedures	previously	outlined,	I	regressed	similarity	to	cause	on	the	dummy	(control	vs.	inclusion)	and	moderator	variables.			Summary	of	results.	In	a	regression	model	with	independent	self-construal,	the	inclusion	manipulation	did	not	significantly	predict	similarity	to	cause	in	terms	of	a	conditional	effect,	neither	did	independent	self-construal,	nor	the	poster	frame.	Of	interest,	however,	was	the	small	but	significant	three-way	interaction	between	independent	self-construal,	poster	frame	and	the	inclusion	manipulation.	As	predicted,	and	in	line	with	the	donation	intentions	and	connection	to	cause	models,	whether	similarity	to		154	cause	ratings	were	greater	for	the	inclusion	condition,	over	the	control	condition,	depended	on	the	poster	frame,	as	well	as	participant’s	independent	self-construal.		In	view	of	the	results	of	the	t-test	and	previous	regression	models,	I	produced	a	second	regression	model	that	included	gender	as	a	covariate.	This	model	was	a	better	predictor	and	there	was	increased	significance	in	the	majority	of	the	conditional	and	interaction	effects,	as	a	result	of	adding	gender.			Regression	model	for	similarity	to	cause.	I	next	submitted	the	similar	to	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	significant	R2	=	.041,	p	=	.03,	F	(7,	372)=	2.292.	Once	again	I	found	no	conditional	effect	of	the	inclusion	condition	B=0.460,	SE=0.288,	t(380)=1.597,	p=.11,	95%	CI	[-0.11,	1.03]	nor	for	independent	self-construal	measure	on	its	own	B=-0.012,	SE=0.019,	t(380)=-0.638,	p=.52,	95%	CI	[-0.05,	0.03],	or	for	poster	type	on	its	own	B=0.304,	SE=0.277,	t(380)=1.099,	p=.27,	95%	CI	[-0.24,	0.85].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.061,	SE=0.027,	t(380)=2.301,	p=.02,	95%	CI	[0.01,	0.11].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.282,	SE=0.402,	t(380)=-0.703,	p=.48,	95%	CI	[-1.07,	0.51].	However,	there	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.053,	SE=0.027,	t(380)=2.014,	p=.05	95%	CI	[0.001,	0.11],	and	most	interestingly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.105,	SE=0.039,	t(380)=-2.696,	p=.007	95%	CI	[-0.18,	-0.03].		The	three-way	interaction	showed	that	in	the	control	condition		155	individuals	with	high	independent	self-construal	indicated	significantly	higher	ratings	of	similarity	to	the	cause	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	individuals	with	high	independent	self-construal	provided	higher	ratings	of	similarity	in	the	“Animals	are	Different”	to	us	text.	A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.019,	F(1,372)	=	7.2661,	p=	.007	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Type)	at	values	of	independent	self-construal	revealed	that	the	effect	on	similarity	ratings	was	not	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=0.7974,	p=.16)	but	was	significant	with	high	self-construal	(θXWàY	l	Z=	10.67	=-1.3974,	p=.02).		Probing	the	conditional	effects	of	the	focal	predictor	at	values	of	the	moderators	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67),	Effect	=0.8733,	SE=0.4052,	t(380)=	2.155,	p=.03,	95%	CI	[0.08,	1.67],	but	not	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33),	Effect	=-0.2467,	SE=0.3793,	t(380)=	-0.6504,	p=.52,	95%	CI	[-0.99,	0.50],	and	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33),	Effect	=0.5507,	SE=	0.4215,	t(380)=	1.3064,	p=.19,	95%	CI	[-0.28,	1.38],	nor	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67),	Effect	=-0.5241,	SE=0.4102,	t(380)=	-1.2779,	p=.20,	95%	CI	[-1.33,	0.28].				156	Regression	model	for	similarity	to	cause	controlling	for	gender.		I	next	submitted	the	similar	to	cause	measure	to	the	MMR	analysis	(Model	3,	Hayes	2018).	The	regression	model	in	total	was	significant	R2	=	.085,	p	<	.001,	F	(8,	371)=	4.304.	There	was	a	marginal	conditional	effect	of	the	inclusion	condition	B=0.558,	SE=0.283,	t(380)=1.973,	p=.05,	95%	CI	[0.001,	1.11]	but	no	significant	effect	for	independent	self-construal	measure	on	its	own	B=-0.018,	SE=0.018,	t(380)=-1.002,	p=.32,	95%	CI	[-0.05,	0.03],	or	for	poster	type	on	its	own	B=0.381,	SE=0.271,	t(380)=1.405,	p=.30,	95%	CI	[-0.15,	0.92].	There	was,	however,	a	significant	interaction	between	inclusion	and	independent	self-construal	B=0.071,	SE=0.026,	t(380)=2.727,	p=.007,	95%	CI	[0.02,	0.12].	There	was	no	significant	interaction	between	inclusion	and	poster	type	B=-0.407,	SE=0.394,	t(380)=-1.032,	p=.30,	95%	CI	[-1.18,	0.37].	However	there	was	a	significant	interaction	between	poster	type	and	independent	self-construal	B=0.057,	SE=0.026,	t(380)=2.201,	p=.03,	95%	CI	[0.01,	0.11],	and	most	interestingly	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	B=-0.821,	SE=0.038,	t(380)=-3.040,	p=.003,	95%	CI	[-0.19,	-0.04].		There	was	a	significant	conditional	effect	of	gender	B=-0.821,	SE=0.195,	t(380)=-4.203,	p<.001,	95%	CI	[-1.21,	-0.44].	The	three-way	interaction	showed	that	in	the	control	condition,	individuals	with	high	independent	self-construal	indicated	significantly	higher	ratings	of	similarity	to	the	cause	with	the	“Animals	are	Similar”	to	us	text,	than	they	did	with	the	“Animals	are	Different”	to	us	text.	However,	in	the	inclusion	condition	individuals	with	high	independent	self-construal	provided	higher	ratings	of	similarity	in	the	“Animals	are		157	Different”	to	us	text.	A	test	of	higher	order	unconditional	interaction	produced	an	ΔR2	of	.023,	F(1,371)=	9.2389,	p=.003	as	a	result	of	the	three-way	interaction.			A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	similarity	to	cause	was	significant	at	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-1.6398,	p=.004)	but	not	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=0.7873,	p=.16).			See	figure	32	for	a	visual	of	the	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	similarity	to	cause	(Y)	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender,	and	figure	33	for	the	corresponding	JN	points	and	confidence	intervals.		Probing	the	conditional	effects	of	the	focal	predictor,	at	values	of	the	moderators	revealed	the	effect	to	be	significant	with	the	Control	Condition	at	high	independent	self-construal	(Z=10.67),	Effect=0.9902,	SE=0.3974,	t(380)=	2.4916,	p=.01,	95%	CI	[0.21,	1.77],	but	not	significant	in	the	Control	Condition	at	low	independent	self-construal	(Z=-10.33),	Effect	=-0.2082,	SE=0.3712,	t(380)=	-0.5607,	p=.58,	95%	CI	[-0.94,	0.52],	not	significant	in	the	Inclusion	Condition	at	low	independent	self-construal	(Z=-10.33),	Effect	=0.5791,	SE=0.4125,	t(380)=	1.4041,	p=.16,	95%	CI	[-0.23,	1.39],	and	not	significant	in	the	Inclusion	Condition	at	high	independent	self-construal	(Z=10.67),	Effect	=-0.6496,	SE=0.4024,	t(380)=	-1.6143,	p=.11,	95%	CI	[-1.44,	0.14].				158	3.5.3.7	Tests	of	mediated	moderation.	Conditional	process	analysis	for	connection	to	cause.		My	next	goal	was	to	test	a	moderated	mediation	model	(model	11;Hayes,	2017)	using	conditional	process	analysis.	Conditional	process	analysis	was	carried	out	to	investigate	the	mediation	of	connection	to	cause	on	donation	intentions	and	actual	cash	donation,	whilst	being	moderated	by	inclusion	(X),	self-construal	(W)	and	poster	type	(Z).	Conditional	process	analysis	is	a	process	by	which	mediation	analysis	and	moderation	analysis	are	combined	in	a	single	model	of,	in	my	case	moderated	mediation,	and	is	used	to	understand	and	describe	the	conditional	nature	of	the	mechanisms	by	which	a	variable	transmits	its	effect	on	another	and	to	test	hypothesis	about	such	contingent	effects	(Hayes	2018).				Covariate	analysis.	In	view	of	the	results	of	the	t-test	and	previous	regression	models,	I	carried	out	the	conditional	process	analyses	with	connection	to	cause	as	a	mediator	with,	and	without,	gender	as	a	covariate.			Regression	model	for	connection	to	cause	as	a	mediator.	My	previous	MMR	analyses	established	that	both	the	Manipulation	(inclusion	vs.	control)	X	Independent	Self-construal	interaction,	as	well	as	the	three-way	interaction,	Manipulation	(inclusion	vs.	control)	X	Independent	Self-construal	X	Poster	Type,	was	associated	with	Connection	to	the	Cause;	as	well	as	with	Donation	intentions;	and	with	Actual	Cash	Donation.	A	regression	test	that	placed	Connection		159	to	Cause	in	as	a	covariate	confirmed	that	the	mediator	(Connection	to	the	Cause)	was	related	to	both	Donation	intentions	B=2.372,	t(380)=26.08,	p<.001	and	Actual	Cash	Donation	B=3.352,	t(380)=7.173,	p<.001.	To	test	the	potential	mediation	effect	I	followed	the	bootstrapping	method	(with	5000	iterations)	advocated	by	Preacher,	Rucker	&	Hayes	(2007).	This	procedure	tests	the	null	hypothesis	that	the	indirect	path	from	the	interaction	term	to	the	dependent	variable	via	the	mediator	does	not	significantly	differ	from	zero.	If	zero	is	not	contained	within	the	confidence	intervals	(CI)	computed	by	the	bootstrapping	procedure,	then	one	may	conclude	that	the	indirect	effect	is	indeed	significantly	different	from	zero	at	p<.05.			a) Connection	to	cause	on	donation	intentions.	A	first	stage	moderated	moderated	mediation	model	(Model	11,	Hayes	2018)	delivered	an	index	test	of	moderated	moderated	mediation	with	a	slope	that	was	statistically	different	from	zero,	Index	=	-0.2687,	Boot	SE=0.0991,	Bootstrap	95%	CI	=	-0.4605	to	-0.0645.	Probing	the	interaction	we	may	see	that	the	indirect	effect	of	inclusion	on	donation	intentions	through	connection	to	cause	(mediator)	is	positive	for	those	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	2.333,	BootSE=	1.0389,	95%	CI	from	0.2110	to	4.3025,	but	not	different	from	zero	for	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Similar”	Poster	type:	(point	estimate:	-0.9897,	BootSE=	0.9984,	95%	CI	from	-2.8774	to	1.0316).	Nor	was	it	different	from	zero	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	-1.2888,	BootSE=	0.9041,	95%	CI	from	-	160	3.0322	to	0.4843),	nor	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	1.031,	BootSE=	0.9805,	95%	CI	from	-0.9909	to	2.8833).		a) Connection	to	cause	on	donation	intentions	controlling	for	gender.	A	first	stage	moderated	moderated	mediation	model	(Model	11,	Hayes	2018)	this	time	controlling	for	gender	delivered	an	index	test	of	moderated	moderated	mediation	with	a	slope	that	was	statistically	different	from	zero,	Index	=	-0.2958,	Boot	SE=0.0957,	Bootstrap	95%	CI	=	-0.4806	to	-0.0995.	Probing	the	interaction	we	may	see	that	the	indirect	effect	of	inclusion	on	donation	intentions	through	connection	to	cause	(mediator)	is	positive	for	those	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	2.8742,	BootSE=	0.9865,	95%	CI	from	0.8684	to	4.7654,	but	not	different	from	zero	for	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	-1.0793,	BootSE=	0.9412,	95%	CI	from	-2.8757	to	0.8576).	Nor	was	it	different	from	zero	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	-1.2872,	BootSE=	0.8747,	95%	CI	from	-3.0054	to	0.4236),	nor	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	0.9707,	BootSE=	0.9703,	95%	CI	from	-1.0410	to	2.7953).				161	b) Connection	to	cause	on	actual	cash	donation.	A	first	stage	moderated	moderated	mediation	model	delivered	an	index	test	of	moderated	moderated	mediation	with	a	slope	that	was	statistically	different	from	zero,	Index	=	-0.3813,	Boot	SE=0.1445,	Bootstrap	95%	CI	=	-0.6632	to	-0.0912.	Probing	the	interaction	we	may	see	that	the	indirect	effect	of	inclusion	on	cash	donation	through	connection	to	cause	(mediator)	is	positive	for	those	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	3.2558,	BootSE=	1.5567,	95%	CI	from	0.2761	to	6.3952,	but	not	different	from	zero	for	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	-1.4034,	BootSE=	1.4489,	95%	CI	from	-4.2967	to	1.3948).	Nor	was	it	different	from	zero	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	-1.8871,	BootSE=	1.2873,	95%	CI	from	-4.3426	to	0.7992),	nor	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	1.4619,	BootSE=	1.3686,	95%	CI	from	-1.2827	to	4.1469).		a) Connection	to	cause	on	actual	cash	donation	controlling	for	gender.	A	first	stage	moderated	moderated	mediation	model	delivered	an	index	test	of	moderated	moderated	mediation	with	a	slope	that	was	statistically	different	from	zero,	Index	=	-0.4238,	Boot	SE=0.1390,	Bootstrap	95%	CI	=	-0.6964	to	-0.1480.	Probing	the	interaction	we	may	see	that	the	indirect	effect	of	inclusion	on	cash	donation	through	connection	to	cause	(mediator)	is	positive	for	those	individuals		162	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	4.0821,	BootSE=	1.4947,	95%	CI	from	1.2862	to	7.0748,	but	not	different	from	zero	for	individuals	with	high	independent	self-construal	(10.67)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	-1.5458,	BootSE=	1.3806,	95%	CI	from	-1.2862	to	7.0748).	Nor	was	it	different	from	zero	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Different”	Poster	type	(point	estimate:	-1.8813,	BootSE=	1.2475,	95%	CI	from	-4.2902	to	0.6948),	nor	for	individuals	with	low	independent	self-construal	(-10.33)	in	the	“Animals	are	Similar”	Poster	type	(point	estimate:	1.3914,	BootSE=	1.3686,	95%	CI	from	-1.4272	to	3.9766).		Conditional	process	analysis	for	empathy	and	similarity	to	cause.	As	a	final	test	I	also	produced	moderated	mediation	models	(model	11;Hayes,	2017)	using	conditional	process	analysis	to	investigate	the	mediation	of	similarity	to	cause	and	well	as	empathy	for	cause,	on	both	donation	intentions	and	actual	cash	donation,	whilst	being	moderated	by	inclusion	(X),	self-construal	(W)	and	poster	type	(Z).	The	empathy	to	cause	and	similarity	to	cause	models	were	only	produced	as	simple	models,	without	gender	added	as	a	covariate	since	findings	did	not	change	substantially	although,	in	line	with	previous	findings,	gender	did	improve	the	prediction	of	both	models.		Both	of	these	mediation	analyses	can	be	found	in	the	appendix	(A.15).				163	3.5.4	Discussion.	Summary.	The	results	of	study	4	repeated	the	findings	of	study	3,	by	demonstrating	that	following	an	inclusion	manipulation,	donation	intentions	significantly	reduced	for	individuals	that	are	high	in	independent	self-construal,	as	compared	to	donation	intentions	under	normal	(control)	conditions.		In	an	attempt	to	advance	the	findings	of	study	3,	study	4	further	probed	the	mechanisms	involved	in	prosocial	donation,	through	the	use	of	a	similar	vs.	different	poster	frame	for	the	out-group	cause.		Under	normal	(control)	circumstances	individuals	that	are	high	in	independent	self-construal	expressed	higher	donation	intentions	for	an	out-group	cause	that	was	framed	to	highlight	similarities	with	them	(“Animals	are	Similar	to	Humans”).	However,	following	an	inclusion	manipulation,	donation	intentions	reduced	for	this	cause.	Following	an	inclusion	manipulation,	individuals	that	are	high	in	independent	self-construal	expressed	higher	donation	intentions	for	an	out-group	cause	that	was	framed	to	highlight	differences	with	them	(“Animals	are	Different	to	Humans”).			In	doing	so,	study	4	offered	support	for	H8:	that	under	normal	conditions	(control	group),	people	with	an	independent	self-construal	orientation	will	express	higher	donation	intentions	for	an	out-group	animal	charity	that	is	framed	as	similar	(vs.	different)	to	them.	The	results	for	ratings	of	connectedness	to	the	cause	mirrored	the	pattern	of	donation	intentions,	as	did	the	result	for	actual	cash	donations	(albeit	not	significantly),	in	a	simple	moderated	moderation	model.					164	Furthermore,	study	4	extended	the	findings	beyond	donation	intentions,	directed	towards	an	out-group	animal	charity	cause,	into	actual	cash	donations,	in	a	model	that	was	mediated	by	connection	to	cause.		A	series	of	mediation	analyses	showed	that	feelings	of	connection	to	cause	acted	as	a	significant	mediator,	and	produced	a	significant	3-way	moderated	moderated	mediation	model	for	both	donation	intentions	and	actual	cash	donations.	See	detailed	discussion	following.		One	of	the	specific	goals	of	study	4	was	to	investigate	whether	a	feeling	of	closeness	and	oneness	(vs.	distance	and	otherness)	with	the	target	could	be	manipulated,	in	order	to	influence	prosocial	behaviour,	and	if	so	whether	it	would	have	significant	effects	regardless	of	self-construal	and	inclusion	status.	In	order	to	do	this	I	had	constructed	two	posters	that	framed	similarity	“Animals	are	Similar”,	or	difference	“Animals	are	Different”,	to	the	cause.	My	prediction	had	been	that	under	normal	circumstances	(control	conditions),	all	participants	would	prefer	the	“Animals	are	Similar”	frame,	and	this	would	be	expressed	by	an	increase	in	donation	intentions.			The	results	of	study	4	supported	my	prediction	that	participants	high	in	interdependent	self-construal	would	normally	donate	more	to	the	“Animals	are	Similar”	poster	frame.	Regression	analysis	showed	a	significant	two-way	interaction	between	poster	type	and	interdependent	self-construal	regarding	donation	intentions.	Put	another	way,	participants	with	high	interdependent	self-construal	expressed	higher	donation	intentions	for	the	“Animals	are	Similar”	cause	in	which	the	poster	framed	the	targets	as	similar	to	humans,	and	lower	donation	intentions		165	for	the	“Animals	are	Different”	cause	in	which	the	poster	framed	the	targets	as	different	to	humans.		In	line	with	my	predictions,	I	found	no	significant	evidence	to	suggest	that	interdependent	self-construal	appeared	to	interact	with	the	inclusion	condition,	to	predict	either	donation	intentions	or	actual	cash	donations	in	any	regression	models.	This	finding	also	supports	my	hypothesis	(H12)	and	extends	past	research	by	offering	evidence	that	individuals	high	in	interdependent	self-construal	are	not	only	little	impacted	by	exclusion	threats	(Gardner,	Pickett,	&	Knowles,	2005;	Powers,	Worsham,	Freeman,	Wheatley,	&	Heatherton,	2014;	Uskul	&	Over,	2014),	but	that	they	are	also	little	impacted	by	inclusion	promises.		However,	as	expected,	interdependent	self-construal	did	interact	with	the	similar/different	prime,	in	that	higher	interdependent	self-construal	predicted	donation	intentions	for	the	“Animals	are	Similar”	poster	and	cause.	This	finding	is	also	in	line	with	past	research,	which	has	found	interdependent	self-construal	to	be	associated	with	in-group	giving,	as	well	as	with	increased	interpersonal	closeness	(Holland,	Roeder,	van	Baaren,	Brandt,	&	Hannover,	2004).			As	outlined,	study	4	was	able	to	offer	evidence	in	support	of	my	prediction	(H8)	that	individuals	high	in	independent	self-construal	would	preference	the	“Animals	are	Similar”	poster	frame,	under	normal	(control)	conditions.	My	argument	for	this	prediction	is	that	individuals	high	in	independent	self-construal	will	feel	more	affinity	and	connection	to	out-group	causes	because,	unlike	individuals	high	in	interdependent	self-construal,	they	do	not	see	out-groups	members	as	rigidly	out-group.	This	is	because,	as	previously	argued,	individuals	high	in	independent	self-	166	construal	a)	have	more	flexible	notions	of	in-group	and	out-group	boundaries,	and	so	are	more	likely	to	view	out-group	members	as	potential	targets	to	recruit	as	in-group	members	and	b)	are	not	as	chronically	reassured	of	their	social	inclusion	status,	compared	to	individuals	with	high	interdependent	self-construal,	and	so	monitor	for	cues	of	social	connection	potential	in	others.			As	a	result,	in	study	4	when	the	out-group	charity	is	specifically	framed	as	similar	I	predicted	that	individuals	with	high	independent	self-construal	would	still	see	it	as	an	out-group,	but	would	take	the	“Animals	are	Similar”	frame	as	a	cue	of	social	connection	potential	and	would	therefore	donate	more	accordingly	to	the	“Animals	are	Similar”	poster	frame.	Regression	analysis	in	study	4	supported	this	prediction	by	showing	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	for	donation	intentions.	Under	normal	circumstances	(control),	participants	with	high	independent	self-construal	did	express	higher	donation	intentions	for	the	“Animals	are	Similar”	cause,	in	which	the	poster	framed	the	targets	as	similar	to	humans,	and	lower	donation	intentions	for	the	“Animals	are	Different”	cause,	in	which	the	poster	framed	the	targets	as	different	to	humans.				This	support	for	the	“Animals	are	Similar”	poster	frame	in	study	4	additionally	offers	support	for	the	argument	(H11)	that	in	normal	circumstances	(control),	individuals	with	high	independent	self-construal	express	higher	donation	intentions	to	out-group	animal	charities	because	they	feel	a	similarity,	connection,	or	oneness	with	them.	Such	a	feeling	of	oneness	may	have	purpose,	in	that	it	is	likely	to	boost	a		167	sense	of	belonging	or	social	inclusion,	which	may	be	especially	motivational	for	individuals	high	in	independent	self-construal.	Additionally,	as	we	have	previous	noted,	it	make	sense	for	individuals	seeking	social	connection	to	reduce	barriers	to	in-group	entry,	and	to	be	more	relaxed	regarding	how	they	implement	in-group	boundaries.	Past	research	supports	the	behaviour	of	relaxing	boundaries	when	we	feel	others	are	similar	(Cuddy,	Rock,	&	Norton,	2007).			Following	an	inclusion	manipulation,	however,	I	predicted	that	participants	high	in	independent	self-construal	would	respond	differently	to	participants	high	in	interdependent	self-construal.	I	predicted	(H10)	that	an	“Animals	are	Different”	poster	frame	would	prove	more	appealing	for	participants	high	in	independent	self-construal	following	an	inclusion	manipulation.	The	reason	for	this	prediction	is	that	I	anticipated	that	an	inclusion	manipulation	would	reassure	participants	high	in	independent	self-construal	of	their	social	inclusion	status,	satiate	their	belonging	needs	(DeWall,	Baumeister,	&	Vohs,	2008),	and	so-by	reduce	a	desire	to	find	similarities	with	out-group	members	or	seek	further	social	connection	with	others.	While	the	“Animals	are	Similar”	poster	frame	would	appeal	to	these	participants	under	normal	conditions	(control),	precisely	because	it	indicates	social	connection	potential	and	encourages	a	relaxing	of	in-group	boundaries,	under	conditions	in	which	the	participants	had	been	previously	assured	of	their	social	inclusion	within	a	human	in-group,	I	predicted	that	the	“Animals	are	Similar”	poster	frame	would	hold	little	appeal.				168	Once	assured	of	belonging	and	inclusion	there	is	little	to	be	gained	from	finding	similarities	with	out-group	members.	In	fact,	it	may	even	be	threatening	to	do	so,	since	finding	similarities	with	out-group	members	might	potentially	challenge	the	recently	affirmed	social	inclusion	and	in-group	status,	maybe	even	presenting	itself	as	a	threat	to	group	distinctiveness.	Under	these	circumstances,	it	may	be	seen	as	more	beneficial	to	reinforce	in-group	and	out-group	categories,	and	donate	more	to	a	poster	that	clearly	affirms	that	animals	are	different	from	humans.	This	argument	is	supported	by	a	wide	body	of	intergroup	research	on	group	differentiation	in	the	face	of	comparison	groups	(Brown,	1984;	Jetten,	Spears,	&	Manstead,	1999;	Jetten,	Spears,	&	Postmes,	2004;	Van	Knippenberg	&	Ellemers,	1990).		This	finding	is	also	in	line	with	past	research	on	intergroup	processes,	suggesting	that,	dependent	on	the	nature	of	the	in-group	vs.	out-group	relationship	between	helper	and	target	and	perceptions	of	differences	or	similarities,	different	processes	and	motivations	are	likely	to	drive	responses.	Sturmer	and	Snyder	(2010)	argue	that	when	there	is	perceived	dissimilarity	between	helper	and	target,	helping	is	more	likely	to	occur	as	a	function	of	perceived	cost	and	benefits	to	self.	The	only	exception,	it	seems,	is	when	there	is	a	question	as	to	whether	a	target	is	seen	to	be	a	typical	or	atypical	out-group	member,	and	under	these	circumstances	the	pattern	may	be	broken	(Manis,	Nelson,	&	Shedler,	1988).			Study	4	was	also	able	to	build	on	study	3	by	providing	further	evidence	in	support	of	H11,	that	connection	to	the	cause	might	act	as	a	successful	mediator	for	donation	intentions.	My	prediction	(H11)	had	been	that	perceptions	of	connection	to	cause		169	would	mediate	the	effects	of	independent	self-construal	orientation,	inclusion	and	their	interaction,	on	the	two	dependent	variables.	The	findings	of	study	4	supported	this	in	a	model	with	poster	type,	inclusion	and	independent	self-construal	predicting	donation	intentions,	and	additionally	provided	some	boundary	conditions	to	demonstrate	how	self-construal	and	inclusion	status	would	impact	this	mechanism.			For	the	cash	donation	dependent	variable,	a	moderated	regression	analysis	showed	no	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	for	actual	cash	donations,	although	the	patterns	were	in	the	same	direction	as	with	donation	intentions,	and	actual	cash	donations	and	donation	intentions	correlated	significantly.		Controlling	for	gender	and	pet	ownership	improved	significance	and	effect	size	marginally.	A	conditional	process	analysis,	using	a	first	stage	moderated	moderated	mediation	model,	however,	confirmed	that	connection	to	cause	significantly	mediated	in	a	model	with	independent	self-construal,	poster	type	and	inclusion,	and	significantly	predicted	both	donation	intentions	and	actual	cash	donations.		Controlling	for	gender	increased	the	effect	size	and	significance	level.			As	with	study	3,	gender	in	study	4	was	found	to	positively	predict	donation	intentions	and	actual	donations,	with	females	expressing	significantly	greater	donation	intentions	and	making	a	near	significantly	greater	actual	donations.	This	is	in	line	with	most	past	research,	which	shows	a	consistent	trend	of	females	donating	more	to	charities	(Einolf,	2011;	Hodgkinson	&	Weitzman,	1992;	1994;	1996;	Kirsh,		170	Hume,	&	Jalnadoni,	1999;	Mesche,	Rooney,	Steinberg,	&	Denton,	2006;	Winterich,	Mittal,	&	Ross	Jr.,	2009).	Study	4	additionally	measured	pet	ownership	and,	as	might	be	expected,	this	also	positively	predicted	donation	intentions	and	actual	cash	donations	to	a	domestic	animal	charity.	Study	4	also	measured	individual	belonging	needs,	and	again	as	expected	this	also	predicted	donation	intentions,	although	not	strongly.	None	of	the	covariates	measured	(gender,	pet	ownership,	unmet	belonging)	interacted	with	the	other	independent	variables	however.			Study	4	also	provided	evidence	to	show	that	empathy	to	the	cause	may	be	predicted	by	perceptions	of	similarity	vs.	difference	to	the	cause,	as	a	function	of	inclusion	status	and	self-construal.		This	is	in	line	with	prior	research	(Cialdini,	Brown,	Lewis,	Luce,	&	Neuberg,	1997),	which	has	offered	evidence	that	empathetic	concern	only	increases	helping	behaviour	through	a	relationship	to	perceived	oneness,	or	self-other	overlap.	A	regression	analysis	showed	a	significant	three-way	interaction	between	poster	type,	inclusion	and	independent	self-construal	regarding	empathy	to	cause.	Under	normal	circumstances	(control),	participants	with	high	independent	self-construal	expressed	higher	empathy	to	cause	for	the	“Animals	are	Similar”	cause	when	the	poster	framed	the	targets	as	similar	to	humans,	and	lower	empathy	to	cause	for	the	“Animals	are	Different”	cause	when	the	poster	framed	the	targets	as	different	to	humans.	However,	following	an	inclusion	manipulation	this	pattern	reversed	for	the	“Animals	are	Different”	cause.	Participants	with	high	independent	self-construal	expressed	higher	empathy	to	cause	for	the	“Animals	are	Different”	cause,	following	a	belonging	manipulation.	There	was	no	significant	impact	of		171	independent	self-construal	seen	for	empathy	to	cause	in	the	“Animals	are	Similar”	cause	poster	following	a	belonging	manipulation.				In	study	4,	a	conditional	process	analysis	did	not	find	evidence	to	support	a	claim	that	empathy	to	cause	was	significantly	mediating	donation	intentions	or	actual	cash	donations,	as	a	function	of	poster	type,	inclusion	status,	and	self-construal.			Taken	together,	the	findings	of	study	4	support	prior	research	on	the	helping	based	on	similarity	hypothesis.	Krebs	famously	(1975)	found	that	when	strangers	were	perceived	as	more	similar,	participants	were	more	willing	to	assist	them	and	give	up	resources	to	do	so.	More	recently,	research	in	marketing	has	found	that	the	more	similar	to	humans	animals	appear	the	more	favourably	they	are	treated	(Connell,	2013)	and	that	anthropomorphism	may	promote	a	sense	of	kinship	(Veer,	2013).		Costello	and	Hodson	(2010)	found	that	inducing	perceptions	of	human	animal	similarity	facilitated	more	inclusive	attitudes	to	both	animal	and	human	outsiders,	and	predicted	less	prejudicial	attitudes	towards	immigrants.	Likewise,	Batt	(2009),	working	in	the	area	of	conservation,	found	a	strong	relationship	between	perceived	similarity	to	humans	and	preference	existed,	which	suggested	that	humans	are	predisposed	to	liking	species	on	the	basis	of	perceived	shared	bio-behavioural	traits.			Limitations.	One	limit	of	study	4	again	was	potential	power	for	detecting	the	interactions	in	the	main	regressions.	Although	G*Power	calculations	indicated	that	power	would	be		172	sufficient	for	detecting	the	main	(conditional)	effects,	interaction	effects	in	regressions	are	typically	smaller	(Aiken	&	West,	1991)	and	therefore	require	higher	power	to	detect.	As	previously	noted,	accurate	calculations	for	complex	regression	models	require	estimates	of	expected	effects	sizes	(unstandardized	regression	coefficients)	at	each	level	of	the	categorical	variable	(or	an	estimate	of	the	size	of	relationships	between	continuous	variable	interactions	and	other	predictor	variables)	for	all	the	conditional	and	interaction	effects	in	the	population,	which	often	may	not	be	known	a	priori	(Hayes,	2018).		While	G*Power	3.1	is	a	respected	and	commonly	used	tool	for	power	analysis	(Faul,	Erdfelder,	Buchner,	&	Lang,	2009),	it	has	been	argued	it	may	be	limited	in	correctly	detecting	power	for	significant	coefficients	in	complex	regressions,	especially	for	those	that	include	mediations	as	well	as	interactions	(Aberson,	2010;	Fritz,	&	MacKinnon,	2007;	Hayes,	2017).	Power	analysis	is	a	rapidly	evolving	field	and	tools	such	as	pwr2ppl	for	R	are	currently	in	development	for	power	detection	of	mediation	and	complex	moderation	results	(Aberson,	2010),	in	addition	to	other	methodology	specifically	aimed	at	addressing	the	problem	of	indirect	effects	within	mediated	regression	models	(Fritz,	&	MacKinnon,	2007;	Ma	&	Zeng,	2014;	Zhang,	2014).	In	the	meantime,	a	word	of	caution	must	be	offered	for	the	interaction	and	mediation	results,	and	a	suggestion	for	future	replication	is	made	in	order	to	verify	the	results.							173	3.6	Christmas	Giving	Study	(Study	5)	3.6.1	Overview.	The	Christmas	Giving	study	(Study	5)	had	two	goals.	The	main	aim	was	to	examine	the	roles	that	self-construal	and	inclusion	would	play	in	donation	behaviour,	in	the	real-world	setting	of	a	field	study.	The	hope	was	to	replicate	the	previous	findings,	in	a	setting	that	would	enable	my	research	to	provide	practical	feedback	that	might	assist	others	when	devising	strategies	for	cause	marketing	campaigns.	The	second	aim	of	study	5	was	to	test	a	novel	new	manipulation	for	inclusion	in	the	form	of	a	charity	poster,	with	wording	designed	to	manipulate	feelings	of	inclusion	(vs.	control).	I	therefore	had	two	predictions;	first	(H7)	that	under	normal	conditions	(control	group)	people	with	a	higher	independent	self-construal	orientation	would	express	increased	donation	intentions	for	an	out-group	animal	charity	than	people	with	a	lower	independent	self-construal	orientation,	and	second	(H9)	that	an	affirmation	of	social	inclusion	(inclusion	manipulation)	would	result	in	reduced	donation	intentions	for	an	out-group	animal	charity	in	people	with	a	higher	independent	self-construal	orientation,	as	compared	to	under	normal	conditions.			3.6.2	Procedure.	Three	research	assistants,	using	intercept	methods	and	paper	surveys,	administered	the	field	study	in	early	December,	in	the	indoor	mall	of	a	Canadian	city	suburb.		In	total	118	participants	were	recruited	to	participate	in	the	study	in	exchange	for	$3.00.	Participants	were	told	that	the	study	was	for	non-commercial	academic	research	looking	at	the	topic	of	donation	interest	in	animal	causes,	and	would	take		174	between	5	and	10	minutes	to	complete.	A	consent	form	was	administered	to	all	participants	prior	to	beginning	the	study.	Of	the	118	participants	recruited,	17	participants	either	did	not	complete	the	survey,	or	withdrew	consent,	resulting	in	a	final	sample	of	101	participants	(55%	female,	Mode	age	=	61	years	and	older).			Participants	filled	in	the	24-item	Self-Construal	scale,	used	in	previous	studies.	They	were	then	shown	a	small	poster	for	an	animal	charity	and	asked	to	rate	the	cause	on	a	variety	of	measures.	Following	this	a	3-item	donation	measure	was	taken,	as	well	as	single	item	empathy	measure.	A	short	belonging	scale	was	administered	and	demographic	questions	were	collected.	Participants	were	then	thanked	and	given	a	debriefing	form.	Finally	participants	were	given	$3.00	as	compensation	for	the	study	in	single	dollar	coins.	After	the	money	had	been	passed	over,	but	before	they	had	left,	participants	were	informed	that	the	administrator	was	collecting	donations	for	the	charity	on	the	poster	and	they	were	offered	the	opportunity	to	donate	any	or	all	of	the	compensation	that	they	had	been	given	should	they	wish.	This	served	as	a	second	dependent	variable.	All	proceeds	were	donated	to	the	BC	SPCA.		Materials.	Participants	first	completed	the	24-item	Self-construal	scale	(Singelis,	1994;	independent	self-construal	α	=	.638,	interdependent	self-construal	α	=	.093)	used	in	the	previous	studies.	One	item	in	the	interdependent	self-construal	scale,	“It	is	important	to	me	to	respect	decisions	made	by	the	group”,	showed	extremely	poor		175	correlation	with	all	the	other	items,	and	was	flagged	for	removal	increasing	the	shortened	scale	reliability	(α	=	.700).			Participants	were	then	shown	one	of	two	small	advertising	posters	and	asked	to	look	at	it.	Both	posters	were	identical	except	for	the	text	on	them,	which	was	either	neutral	(control),	or	belonging	(inclusion	manipulation)	in	content.	In	the	inclusion	condition	participants	were	encouraged	to	think	during	the	holiday	season,	about	how	much	they	were	cherished	by	others,	and	to	celebrate	their	sense	of	belongingness.	In	the	neutral	condition	participants	were	encouraged	to	think	during	the	holiday	season,	about	taking	time	off	to	relax,	and	enjoying	the	opportunity	to	do	things	they	wanted.	Both	also	urged	participants	to	think	about	donating	to	the	BC	SPCA	at	this	time	of	year	(see	appendix	A.16	for	poster	materials).			Participants	were	then	asked	to	evaluate	the	cause,	on	five	criteria	using	a	7-point	scale	(1=	very	much,	7	=	not	at	all;	scale	reliability	α	=	.962).	Specifically,	they	were	asked	how	much	they	liked	the	cause,	were	favourable	to	the	cause,	were	positive	about	the	cause,	considered	the	cause	desirable,	and	considered	the	cause	good.	A	3-item	donation	intentions	measure	(α	=	.953)	was	then	taken	as	the	dependent	variable.	This	measure	required	participants	to	answer	on	a	7-point	scale	regarding	how	likely,	how	inclined	and	how	willing	they	were	to	donate	to	a	campaign	for	the	cause	on	the	poster	they	had	been	shown.	Following	this	participants	were	asked	how	much	empathy	they	had	for	the	cause,	and	a	3-item	belonging	scale	(α	=	.810)		176	was	administered:	“I	feel	connected	with	others”,	“I	feel	isolated	from	the	rest	of	the	world”,	“	I	have	a	sense	of	belonging”	using	a	7-point	scale.	Brief	demographics	were	collected	(age,	gender,	and	pre-tax	income)	in	category	form.			3.6.3	Results.	The	study	was	by	necessity	short	in	length	and	administered.	As	a	consequence	no	attention	checks	were	included,	nor	was	a	question	regarding	the	research	hypothesis	asked.	No	participants	who	completed	the	survey	and	gave	consent	were	excluded.			Power.	A	post	hoc	power	analysis	was	conducted,	using	G*Power	3.1.	With	a	sample	size	of	101	in	a	multiple	regression	model	with	3	predictors	(2	predictors	and	1	interaction	term)	and	an	observed	R2	of	0.12,	ρ2	was	calculated	at	0.127,	which	provided	a	power	estimation	of	0.83	at	95%	confidence	level.		A	post	hoc	power	test	conducted	using	G*Power	3.1	for	an	independent	sample	t-test	revealed	that	with	a	sample	size	of	101	(50	per	condition)	an	observed	d=0.57,	with	critical	t	of 1.984,	was	needed	for	a	power	estimation	of	>0.80	at	95%	confidence	level.	Furthermore,	a	post	hoc	power	test	for	correlations	revealed	that	with	a	sample	size	of	101,	an	observed	correlation	of	.28,	was	needed	for	a	power	estimation	of	>0.80	at	95%	confidence	level.	See	appendix	A.17	for	power	analysis	output.				177	Manipulation	check.	The	3-item	belonging	scale	was	intended	to	serve	as	a	manipulation	check.	It	was	expected	that	higher	ratings	of	belonging	would	be	seen	in	the	inclusion	condition.	However,	the	short	belong	scale	measure	was	not	found	to	vary	by	condition,	with	the	control	condition	(n=50;	M=10.20,	SD=4.34)	not	being	significantly	different	from	the	inclusion	condition	(n=51;	M=10.18,	SD=5.75)	according	to	an	independent	sample	t-test	t(99)=0.023,	p=.98.			Demographics.	Demographic	data	was	not	collected	on	ethnicity	within	the	survey,	however,	administrators	confirmed	that	the	sample	was	highly	Caucasian	in	content.	Survey	results	showed	that	15.8%	were	18-30	years,	16.8%	were	31-40	years,	18.8%	were	41-50	years,	20.8%	were	51-60	years,	and	27.7%	were	over	60	years	old	(see	figure	34).	In	terms	of	pre-tax	income	14.9%	reported	under	$20,000,	13.9%	reported	$20,000-$40,000,	22.8%	reported	$41,000-$70,000,	14.9%	reported	$71,000-$100,000,	and	27.7%	reported	over	$100,000,	with	5.9%	missing.	Participants	identified	as	55%	female.			Demographic,	trait	and	other	correlations.	Results	of	the	study	5	revealed	a	significant	positive	correlation	(r(99)=.371,	p<.001)		between	donation	intentions	(measured	in	the	3-item	measure)	and	independent	self-construal,	compared	to	between	interdependent	self-construal	and	donation	intentions	(r(99)=.185,	p=.06).	Independent	self-construal	also	correlated		178	positively	with	cause	rating	(r(99)=.306,	p=.002),	but	not	with	actual	cash	donations	(r(99)=.013,	p=.90).	Empathy	for	the	cause	did	not	correlate	with	either	independent	self-construal	(r(99)=-.022,	p=.83).	nor	interdependent	self-construal(r(99)=-.113,	p=.26).		Actual	cash	donations	did	not	correlate	with	donation	intentions	(r(99)=.072,	p=.47).	There	was	a	positive	correlation	between	actual	cash	donations	and	age	r(99)=.228,	p=.02,	but	not	between	donation	intentions	and	age	r(99)=.093,	p=.35.			Means	comparisons.		An	independent	sample	t-test	showed	that	females	(n=55;	M=30.07,	SD=5.71)	rated	the	cause	significantly	higher	than	did	males	(n=46;	M=27.65,	SD=6.48)	t(99)=	-1.996,	p=.049,	d=	0.40,	however	neither	gender	were	significantly	more	likely	to	express	higher	intentions	to	donate	t(99)=	0.106,	p=.92,	or	give	higher	cash	donations	t(99)=	1.094,	p=.28.			An	independent	sample	t-test	showed	that	participants	in	the	control	condition	expressed	significantly	higher	empathy	for	the	cause	(n=50;	M=4.46,	SD=1.72)	than	did	participants	in	the	inclusion	condition	(n=51;	M=3.41,	SD=2.07)	t(99)=	2.767,	p=.007	d=0.55.	Participants	in	the	control	condition,	however,	expressed	significantly	lower	donation	intentions	for	the	cause	(n=55;	M=11.98,	SD=4.80)	than	did	participants	in	the	inclusion	condition	(n=51;	M=14.96,	SD=4.83)	t(99)=	-3.109,	p=.002	d=0.60.	There	was	no	difference	on	cash	donations	between	conditions	t(99)=	-0.596,	p=.55.		179	Regression	tests	for	moderation.	Moderated	multiple	regressions	were	carried	out	in	SPSS	using	process	v.3	(Hayes,	2018)	to	examine	whether	poster	style	(Control	vs.	Inclusion)	influenced	participant	donation	intentions,	actual	cash	donations,	and	empathy	to	cause,	and	whether	these	effects	were	moderated	by	self-construal	orientation.			Following	Cohen	&	Cohen	(1983)	and	Wendorf	(2004),	I	dummy	coded	the	condition	for	the	inclusion	manipulation	vs.	control.	Independent	and	interdependent	self-construal	were	mean	centered,	in	accordance	with	guidelines	(Hayes,	2013)	stipulating	that	this	practice	renders	subsequent	tests	of	hypotheses	and	regression	coefficients	for	X	and	M	more	meaningful	and	substantially	interpretable,	as	well	as	to	reduce	the	likelihood	of	errors	in	interpretation.	All	dependent	measures	were	subjected	to	separate	moderated	multiple	regressions	with	each	manipulation	condition	and	the	control,	each	self-construal	scores,	and	their	interactions	simultaneously	entered	as	predictor	variables.	As	in	the	previous	studies	unstandardized	coefficients	(B)	are	reported	rather	than	betas	(β).	Because	βs	are	not	properly	standardized	in	interaction	terms	they	are	not	interpretable,	whereas	B	represents	the	difference	between	the	un-weighted	means	of	the	groups	involved	(see	Cohen	et	al.	2003).			3.6.3.1	Role	of	poster	and	self-construal	in	donation	intentions,	cash	donation,	and	empathy.	Summary	of	results.		180	Against	expectations,	neither	independent	self-construal,	nor	interdependent	self-construal	had	any	significant	impact	on	donation	intentions	or	cash	donations	in	study	5.	Furthermore,	neither	self-construal	significantly	interacted	with	the	poster	condition	(control	vs.	inclusion).	The	poster	condition	appeared	to	have	a	strong	conditional	effect,	with	inclusion	positively	predicting	donation	intentions	and	empathy	to	the	cause,	in	separate	regression	models,	with	both	independent	and	interdependent	self-construal.	See	detailed	analyses	below.				Moderating	influence	of	poster	on	independent	self-construal	for	donation	intentions.	I	submitted	the	donation	intentions	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	independent	self-construal	and	poster	(control	vs.	inclusion)	entered	as	independent	variables.	The	regression	model	in	total	was	significant	R2	=	.218,	p	<.001,	F(3,	97)=	9.028.	I	found	a	significant	conditional	effect	for	the	poster	B=2.783,	SE=1.006,	t(101)=2.767,	p=.007,	95%	CI	[0.79,	4.78]	and	for	the	independent	self-construal	measure	on	its	own	B=0.224,	SE=0.097,	t(101)=2.296,	p=.02,	95%	CI	[0.03,	0.42.	There	was	no	significant	interaction	between	poster	and	independent	self-construal	B=0.016,	SE=0.122,	t(101)=0.131,	p=.90,	95%	CI	[-0.23,	0.26].		Moderating	influence	of	poster	on	independent	self-construal	for	actual	cash	donations.		I	submitted	the	cash	donations	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	independent	self-construal	and	poster	(control	vs.	inclusion)	entered	as	independent	variables.	The	regression	model	in	total	was	not	significant	R2	=.014,	p		181	=.72,	F(3,	97)=	0.442.	I	found	no	conditional	effect	for	the	poster	B=-0.362,	SE=0.374,	t(101)=-0.967,	p=.34,	95%	CI	[-1.10,	0.38]	nor	for	the	independent	self-construal	measure	on	its	own	B=0.014,	SE=0.027,	t(101)=-0.502,	p=.62,	95%	CI	[-0.07,	0.04].	There	was	no	significant	interaction	between	poster	and	independent	self-construal	B=0.044,	SE=0.045,	t(101)=0.981,	p=.33,	95%	CI	[-0.05,	0.13].			Moderating	influence	of	poster	on	independent	self-construal	for	empathy	to	cause.	I	submitted	the	empathy	to	cause	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	independent	self-construal,	poster	(control	vs.	inclusion),	and	the	interaction	between	them,	entered	as	independent	variables.	The	regression	model	in	total	was	significant	R2	=.092,	p	=.02,	F(3,	97)=	0.442.	I	found	an	approaching	significant	conditional	effect	for	the	poster	B=-0.767,	SE=0.424,	t(101)=-1.807,	p=.07,	95%	CI	[-1.61,	0.08]	but	not	for	the	independent	self-construal	measure	on	its	own	B=0.045,	SE=0.041,	t(101)=1.096,	p=.28,	95%	CI	[-0.04,	0.13].	There	was	no	significant	interaction	between	poster	and	independent	self-construal	B=-0.075,	SE=0.051,	t(101)=-1.462,	p=.15,	95%	CI	[-0.18,	0.03].			Moderating	influence	of	poster	on	interdependent	self-construal	for	donation	intentions.	I	submitted	the	donation	intentions	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	interdependent	self-construal	and	poster	(control	vs.	inclusion)	entered	as	independent	variables.	The	regression	model	in	total	was	significant	R2	=	.111,	p	=.009,	F(3,	97)=	4.05.	I	found	a	significant	conditional	effect	for	the	poster	B=2.795,		182	SE=0.966,	t(101)=2.894,	p=.01,	95%	CI	[0.88,	4.71]	but	not	for	the	interdependent	self-construal	measure	on	its	own	B=0.065,	SE=0.095,	t(101)=0.691,	p=.49,	95%	CI	[-0.12,	0.25].	There	was	no	significant	interaction	between	poster	and	interdependent	self-construal	B=0.037,	SE=0.120,	t(101)=0.309,	p=.76,	95%	CI	[-0.20,	0.27].			Moderating	influence	of	poster	on	interdependent	self-construal	for	actual	cash	donations.	I	submitted	the	cash	donations	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	interdependent	self-construal,	poster	(control	vs.	inclusion),	and	the	interaction	between	them,	entered	as	independent	variables.	The	regression	model	in	total	was	not	significant	R2	=.014,	p	=.72,	F(3,	97)=	0.442.	I	found	no	conditional	effect	for	the	poster	B=-0.362,	SE=0.374,	t(101)=-0.967,	p=.34,	95%	CI	[-1.10,	0.38]	nor	for	the	independent	self-construal	measure	on	its	own	B=0.014,	SE=0.027,	t(101)=-0.502,	p=.62,	95%	CI	[-0.07,	0.04.	There	was	no	significant	interaction	between	poster	and	independent	self-construal	B=0.044,	SE=0.045,	t(101)=0.981,	p=.33,	95%	CI	[-0.05,	0.13].			Moderating	influence	of	poster	on	interdependent	self-construal	for	empathy	to	cause.	I	submitted	the	empathy	to	cause	measure	to	the	MMR	analysis	(Model	1,	Hayes	2018)	with	interdependent	self-construal,	poster	(control	vs.	inclusion),	and	the	interaction	between	them,	entered	as	independent	variables.	The	regression	model	in	total	was	significant	R2	=.134,	p	=.003,	F(3,	97)=	4.996.	I	found	a	significant		183	conditional	effect	for	the	poster	B=-1.041,	SE=0.373,	t(101)=-2.789,	p=.006,	95%	CI	[-1.78,	-0.30]	but	not	for	the	interdependent	self-construal	measure	on	its	own	B=0.054,	SE=0.036,	t(101)=1.481,	p=.14,	95%	CI	[-0.02,	0.13].	There	was	a	significant	interaction	between	poster	and	interdependent	self-construal	B=-0.116,	SE=0.046,	t(101)=-2.506,	p=.01,	95%	CI	[-0.21,	0.02].			3.6.4	Discussion.	Results	for	study	5	were	mixed.	There	appeared	to	be	a	main	effect	of	condition,	with	participants	in	the	control	condition	expressing	significantly	lower	donation	intentions	for	the	cause	compared	to	participants	in	the	inclusion	condition,	according	to	independent	sample	t-tests.	This	finding	was	corroborated	by	the	regressions,	which	showed	a	conditional	effect	for	the	condition	also.	H9	was	therefore	not	supported	by	the	results.	There	also	appeared	to	be	a	main	effect	found	for	empathy,	but	in	a	reverse	direction	with	reduced	empathy	reported	in	the	inclusion	condition.	No	clear	effect	was	found	for	either	self-construal,	in	terms	of	donation	intentions	or	cash	donations.	H7	was	therefore	also	not	supported	by	the	results.	Interdependent	self-construal	interacted	with	the	inclusion	condition	to	produce	a	negative	prediction	for	empathy;	in	that	empathy	was	reduced	following	the	inclusion	manipulation.		Neither	rating	of	cause	nor	empathy	for	cause	were	found	to	mediate	donation	intentions.	While	previous	research	has	found	that	gender	moderates	prosocial	behaviour,	and	may	be	mediated	by	empathy	(Willer,	Wimer,	&	Owens,	2015),	this	finding	was	not	supported	by	study	5.				184	The	somewhat	mixed	findings	of	study	5	were	disappointing,	but	possibly	speak	to	the	complexities	of	running	field	studies	where	it	is	hard	to	control	for	confounding	variables.	Intercept	surveys	by	their	nature	must	be	kept	short	and	it	is	hard	to	collect	sufficient	data	to	manage	for	other	variables.	The	sample	size	was	relatively	small,	due	in	part	to	the	time	window	I	had	been	able	to	negotiate,	in	which	to	collect	data.			Moreover,	the	demographics	in	this	survey	varied	dramatically	from	all	previous	studies,	with	the	age	range	of	participants	being	mainly	over	50	years.		Research	shows	that	belonging	needs	may	be	experienced	differently	over	a	lifespan,	and	that	responses	to	inclusion	and	exclusion	threats	may	differ	with	age	(Andersson,	1998;	Minichiello,	Browne,	&	Kendig,	2000;	Newsom	&	Schulz,	1996;	Nicolaisen,	&	Thorsen,	2014;	Schultz	Jr.,	&	Moore,	1984).	Specifically	research	has	suggested	that	loneliness	has	a	curvilinear	relationship	with	age	(Lasgaard,	Friis,	&	Shevlin,	2016),	with	those	under	25	years	and	over	65	years	experiencing	the	highest	levels	of	loneliness	(Victor,	&	Yang,	2012).	In	view	of	these	findings	the	skewed	age	demographic	in	my	sample	may	have	impacted	experiences	of	the	manipulation,	as	compared	to	my	previous	studies.		Another	question	hovers	around	how	successful	my	novel	manipulation	had	been	at	increasing	feelings	of	inclusion.	In	consideration	of	the	fact	that	the	manipulation	check	noted	problems,	in	that	the	3-item	belonging	scale	failed	to	offer	significantly	different	results	between	conditions,	it	may	be	questioned	as	to	whether	the		185	inclusion	manipulation	that	I	had	constructed,	functioned	as	planned.	For	some	people,	perhaps	those	with	a	good	network	of	social	support,	it	may	indeed	have	functioned	to	increase	a	sense	of	belonging;	but	for	others,	perhaps	lonely	and	isolated	it	may	have	had	a	reverse	effect	and	made	current	loneliness	more	salient.	It	is	possible,	therefore,	that	it	may	have	functioned	as	both	an	inclusion	and	exclusion	manipulation.	This	is	clearly	of	concern,	and	offers	a	warning	to	future	researchers,	who	may	be	looking	to	try	out	novel	inclusion	and	exclusion	manipulations.			A	final	aspect	of	this	study,	that	caused	challenges,	was	the	experiences	of	all	the	research	assistants	when	collecting	the	data.	There	was	often	confusion	over	questions	on	the	part	of	the	respondents	and	some	respondents	particularly	struggled	with	the	self-construal	scales.	This	is	may	be	seen	evidenced	in	the	poorer	scale	reliability	for	these	scales,	than	may	be	normally	expected,	and	the	need	to	remove	one	item	entirely	from	the	interdependent	self-construal	scale.		In	consideration	of	the	fact	that	this	scale	was	developed	with	students	in	mind	it	is	possible	that	it	lacks	reliability	and	validity	across	older	age	groups	in	the	general	population.		In	addition	it	was	challenging	to	administer	the	cash	donation	dependent	variable	at	times,	and	the	results	may	reflect	this	struggle.	Many	people	appeared	to	feel	awkward	taking	part	of	the	compensation	payment,	and	donating	another	part,	and	so	there	was	often	an	all-or-nothing	response.	Furthermore,	being	forced	to	make	the	choice	regarding	donation	in	a	public	setting	may	have	influenced	the	response.	Past	research	has	found	that	people	will	respond		186	differently	to	charity	appeals,	as	a	factor	of	whether	they	are	being	made	in	public,	or	private,	setting	(Froming,	Walker,	&	Lopyan,	1982;	Kristofferson,	White	&	Peloza,	2014;	Simpson,	White	&	Laran,	2018;	White	&	Peloza,	2009;	Wu,	Gao,	&	Mattila,	2017).	It	is	also	possible	that	the	fact	that	participants	were	filling	in	the	survey	in	public	with	an	administrator,	may	also	have	influenced	responses	to	the	donation	intentions	dependent	variable.	Research	has	noted	that	reputational	incentives	may	influence	donations	and	may	depend	on	the	situational	(public	vs.	private)	context	(Reinstein,	&	Riener,	2012;	Simpson,	&	Willer,	2008).	Studies	1-4	had	all	taken	place	in	a	private	setting,	in	which	participants	could	be	relatively	sure	that	their	responses	to	questions	regarding	donation	support	and	inclusive	behaviour	were	to	be	kept	confidential.	However,	study	5	was	conducted	in	public	and	as	such	presented	a	major	difference	in	context.			In	sum,	study	5	was	not	able	to	offer	much	support	for	my	predictions,	and	did	not	support	my	expectation	that	independent	self-construal	would	interact	with	an	inclusion	manipulation	to	reduce	donation	intentions.				 		187	Chapter	4:	General	Discussion		4.1	General	Discussion	Results	across	four	studies	provide	converging	evidence	in	support	of	H5	and	H9;	that	an	affirmation	of	social	inclusion	or	belonging	will	reduce	out-group	directed	prosocial	behaviour,	in	individuals	with	high	independent	self-construal.	Study	1	offers	evidence	to	support	H1:	that	under	normal	conditions	individuals	with	high	independent	self-construal	may	behave	more	inclusively	towards	out-group	targets	in	a	grouping	task,	whilst	individuals	with	high	interdependent	self-construal	behave	less	inclusively.	Study	1	also	provides	evidence	for	a	relationship	between	trait	anthropomorphism	(finding	human	attributes	in	objects	and	non-human	others)	and	high	independent	self-construal,	and	supports	H2:	that	high	independent	self-construal	is	positively	related	to	the	inclusion	of	more	items	in	the	in-group	that	rate	high	in	potential	for	social	interaction.	This	is	supportive	of	the	argument	that	individuals	with	high	independent	self-construal	may	use	‘social	potential’	as	a	cue	for	in-group	selection.			Study	2	replicates	the	findings	of	study	1	and	offers	evidence	in	support	of	H5:	that	the	pattern	reverses	following	an	affirmation	of	social	inclusion	status,	after	which	individuals	with	high	independent	self-construal	behave	less	inclusively	towards	out-group	members,	in	addition	to	offering	support	for	H6;	and	that	they	will	also	offer	reduced	ratings	of	social	anthropomorphism	in	a	photo	task.	Taken	together	study	1	and	2	offer	evidence	that	individuals	with	high	independent	self-construal		188	may	be	evaluating	and	including	out-group	members,	as	a	result	of	the	social	resource	potential	they	offer,	but	that	this	behaviour	ceases	following	confirmation	of	their	own	social	inclusion	status.	This	finding	is	supported	by	prior	work	of	Pickett,	Gardner,	and	Knowles	(2004),	demonstrating	that	a	higher	need	for	belonging	correlates	with	enhanced	monitoring	of	social	cues.	Moreover,	I	extend	the	work	of	Pickett,	Gardner,	and	Knowles	(2004)	into	the	domain	of	self-construal,	by	offering	evidence	to	support	the	proposition	that	under	normal	circumstances	(control),	individuals	high	in	independent	self-construal	will	be	more	inclined	to	monitor	for	social	cues,	which	may	indicate	a	potential	for	increased	social	connection.	Study	2	also	shows	that	the	pattern	reverses	following	an	affirmation	of	inclusion,	and	I	suggest	this	may	be	as	a	result	of	a	reduced	need	for	social	connection	and	an	activated	desire	to	signal	in-group	distinctiveness	following	an	affirmation	of	belonging.	Study	2	also	offers	evidence,	supporting	H12,	to	show	that	individuals	high	in	interdependent	self-construal	are	buffered	against	affirmations	of	inclusion.		Study	3	extends	the	pattern	into	the	domain	of	prosocial	behaviour,	that	of	donation	support	directed	towards	an	out-group	cause,	and	offers	some	evidence	to	support	H8:	that	individuals	with	high	independent	self-construal	will	normally	express	higher	donation	support	for	out-group	causes,	but	that	following	an	affirmation	of	inclusion	they	will	express	less	donation	support.	Study	4	repeats	the	findings	of	study	3	and	offers	evidence	to	demonstrate	that	individuals	with	high	independent	self-construal	prefer	to	donate	to	out-groups	that	seem	similar	to	them,	under		189	normal	conditions,	but	that	following	an	affirmation	of	inclusion	they	prefer	to	donate	to	out-groups	that	appear	different	from	their	in-group.	This	finding	supports	H10.	Study	4	also	provides	evidence	in	support	of	H11:	that	a	perception	of	connection	to	the	cause	mediates	donations	towards	the	cause	for	individuals	with	high	independent	self-construal.			Taken	together	the	research	provides	support	for	an	argument	that	individuals	with	high	independent	self-construal	may	be	more	open	to	admitting	out-group	members	to	an	in-group,	and	more	supportive	of	out-group	causes,	because	they	are	looking	for	opportunities	to	connect	with	others,	and	therefore	are	vigilant	for	cues	of	social	connection	potential	in	others.	In	doing	so,	my	research	provides	support	for	an	argument	that	individuals	with	high	independent	self-construal	may	be	equally,	if	not	more,	invested	in	managing	their	social	inclusion	status,	compared	to	individuals	with	high	interdependent	self-construal.	In	addition,	I	argue	that	following	a	confirmation	of	social	inclusion,	individuals	with	high	independent	self-construal	experience	a	reduced	desire	to	seek	social	resources,	and	an	increased	desire	to	signal	and	defend	in-group	status,	which	may	result	in	a	reduction	of	inclusive	behaviour	towards	out-groups	and	a	consolidation	and	differentiation	of	in-group	and	out-group	boundaries	(Tajfel,	1979;	Tajfel	&	Turner,	1979;	Yuki	&	Takemura,	2013)	in	order	to	maintain	self	esteem	(Branscombe,	Ellemers,	Spears,	&	Doosje,	1999).					190	4.2	Theoretical	Implications	This	paper	contributes	to	the	growing	literature	examining	the	impacts	of	social	exclusion	and	inclusion	on	prosocial	behaviour	(Beest	&	Williams,	2011;	Catanese	&	Tice,	2005;	Cuadrado,	Tabernero	&	Steinel,	2015;	Lee	&	Shrum,	2012;	Maner,	DeWall,	Baumeister,	&	Schaller,	2007;	Twenge,	Baumeister,	DeWall,	Ciarocco,	&	Bartels,	2007).	While	most	previous	work	has	focused	on	the	negative	downstream	consequences	of	social	exclusion,	I	have	instead	focused	on	social	inclusion,	and	have	built	on	past	research	detailing	the	negative	downstream	results	of	social	exclusion	(Twenge,	Baumeister,	DeWall,	Ciarocco,	&	Bartels,	2007),	by	demonstrating	that	there	may	also	be	negative	downstream	consequences	to	inclusion,	especially	with	regard	to	prosocial	support	for	out-group	targets.			Furthermore	this	dissertation	seeks	to	extend	the	existing	research	into	the	impacts	of	self-construal	orientation	on	prosocial	behaviour	(Duclos	&	Barasch,	2014;	Kemmelmeier,	Jambor,	&	Letner,	2006;	Pfundmair,	Graupmann,	Frey,	&	Aydin,	2015;	Ren,	Wesselmann,	&	Williams,	2013;	Simpson,	White,	&	Laran,	2018),	to	propose	some	boundary	conditions	regarding	the	whens	and	the	hows.	Specifically,	regarding	the	former,	I	offer	evidence	that,	when	individuals	with	high	independent	self-construal	experience	an	affirmation	of	inclusion,	they	will	express	lower	donation	intentions,	and	behave	less	inclusively,	than	under	normal	conditions.	Regarding	the	latter,	I	demonstrate	in	my	research	that	the	impacts	of	social	inclusion	appear	to	primarily	affect	those	individuals	high	in	independent	self-construal,	whereas	individuals	high	in	interdependent	self-construal	remain		191	relatively	impervious.	This	finding	is	in	line	with	past	research	on	the	responses	of	individuals	high	in	interdependent	self-construal	to	social	exclusion.	Furthermore,	I	make	a	novel	proposition,	that	this	happens	as	a	result	of	individuals	high	in	independent	self-construal	seeing	prosocial	inclusive	behaviour	as	a	means	by	which	to	improve	social	connection	and	to	feel	more	socially	included.		On	the	other	hand,	individuals	who	are	high	in	interdependent	self-construal	will	consider	themselves	chronically	assured	of	inclusion,	as	a	result	of	their	longstanding	rigid	in-group	standing,	and	will	therefore	be	relatively	unmoved	by	exclusion	or	inclusion	manipulations.	To	support	this	proposition,	I	offer	preliminary	evidence	that	individuals	high	in	independent	self-construal	make	inclusion	choices	based	on	the	social	potential	of	others	(study	1),	and	are	more	likely	to	anthropomorphize	others	on	attributes	relating	to	social	potential	(study	2).	Furthermore,	I	demonstrate	(study	4)	that	for	individuals	with	high	independent	self-construal,	donation	intentions	towards	out-group	targets	is	mediated	by	feelings	of	connection	to	the	target,	and	moderated	by	perceptions	of	similarity	vs.	difference	with	the	target.			Additionally,	I	contribute	to	the	small	body	of	research	looking	into	anthropomorphism,	as	a	function	of	social	exclusion	(Epley,	Waytz,	Akalis	&	Cacioppo,	2008;	Eyssel,	&	Reich,	2013;	Powers,	Worsham,	Freeman,	Wheatley,	&	Heatherton,	2014;	Waytz,	Epley,	&	Cacioppo,	2010;	Waytz	&	Epley,	2012)	and	provide	tentative	evidence	that	not	only	is	independent	self-construal	specifically,	and	positively,	related	to	a	tendency	to	socially	anthropomorphise,	but	also	that		192	feelings	of	social	inclusion	may	interact	with	independent	self-construal	to	reduce	ratings	of	social	anthropomorphism—a	finding	that	builds	on	the	recent	work	of	Bartz,	Tchalova,	and	Fenerci	(2016).				Taken	together,	this	research	adds	a	degree	of	depth	and	nuance	to	the	existing	body	of	literature	by	demonstrating	the	moderating	influence	of	self-construal,	and	extends	previous	work	on	the	negative	downstream	effects	of	social	exclusion	to	include	that	of	social	inclusion.	Finally,	this	research	also	demonstrates	that	the	findings	extend	beyond	human	out-groups,	struggling	to	enjoy	adequate	prosocial	support,	by	evidencing	that	the	findings	are	also	relevant	in	the	context	of	animal	out-groups.	In	doing	so	it	builds	on	recent	research	that	is	looking	at	human-animal	relations	through	an	intergroup	lens	(Amiot	&	Bastian,	2017).		4.3	Practical	Implications		From	a	managerial	perspective	the	current	research	provides	important	insights	for	marketing	practice,	especially	in	the	realms	of	cause	marketing	and	prosocial	behaviour.	In	studies	3	and	4	I	explicitly	investigated	whether	the	more	inclusive	behaviour	of	the	earlier	studies	might	extend	into	other	forms	of	prosocial	behaviour	with	direct	financial	benefits,	and	provided	evidence	in	the	affirmative.	The	potential	of	increasing	donations	for	out-group	charities	has	clear	direct	benefits	for	real	world	setting,	since	many	conservation	charities,	as	well	as	charities	aimed	at	benefiting	third	world	causes,	struggle	to	attract	sufficient	donors	and	funding.	However,	so	far,	little	academic	research	has	examined	the		193	psychological	mechanisms	behind	this	behaviour.	My	hope	is	that	this	research	goes	some	way	to	improving	understanding	of	the	motivational	mechanisms	that	increase	and	decrease	prosocial	behaviour,	towards	out-groups	and	distant	others,	and	will	be	able	to	offer	valuable	insights	for	conservation	organizations,	animal	welfare	and	humane	charities,	and	even	third	world	appeals,	wishing	to	leverage	these	concepts	to	improve	prosocial	behaviour.				One	important	result	of	my	research	was	the	finding	that	individuals	that	are	high	in	independent	self-construal	may	behave	less	prosocially	when	they	are	assured	of	their	belonging	status.	While	there	are	clear	ethical	problems	associated	with	encouraging	people	to	feel	lonely	expressly	in	order	to	boost	donation	behaviour,	charities	should	ensure	that	they	do	not	prime	the	opposite	(social	inclusion),	as	this	may	well	reduce	donation	behaviour.	Furthermore	it	should	be	noted	that	manipulations	of	belonging	can	be	tricky,	as	may	be	seen	in	my	final	study	(study	5),	when	a	novel	attempt	at	an	inclusion	manipulation	appeared	to	backfire,	potentially	causing	some	of	the	older	study	participants	to	become	more	aware	of	their	lack	of	social	connections.			One	other	final	finding	from	my	research	is	the	corroboration	of	past	findings,	showing	that	individuals	high	in	interdependent	self-construal	do	appear	to	give	less	to	out-group	charities	than	do	individuals	high	in	independent	self-construal	(Duclos	&	Barasch,	2014;	Kemmelmeier,	Jambor,	&	Letner,	2006;	Pfundmair,	Graupmann,	Frey,	&	Aydin,	2015;	Ren,	Wesselmann,	&	Williams,	2013).	Charities		194	with	tight	budgets	may	wish	to	plan	accordingly	with	this	in	mind	in	order	to	maximise	return	on	their	marketing	investments.			Mitigating	this	issue	somewhat,	my	research	points	to	the	need	to	highlight	self-other	similarities	in	charity	appeals,	which	should	meet	with	increased	support	regardless	of	self-construal	orientation.	This	builds	on	previous	marketing	research,	suggesting	that	the	more	similar	to	humans	animals	appear,	the	more	favourably	they	will	be	treated	(Connell,	2013).	As	has	been	previously	noted,	a	strong	relationship	between	perceived	similarity	to	humans	and	preference	exists,	which	suggests	that	humans	are	predisposed	to	liking	species	on	the	basis	of	perceived	shared	bio-behavioural	traits	(Batt,	2009).		However,	as	also	noted,	the	framing	of	appeals	that	highlight	similarities	needs	to	be	carefully	worded	in	order	to	avoid	the	dangers	of	priming	mortality	salience	(Costello	&	Hodson,	2010).			Moreover,	as	my	research	notes,	cues	of	similarity	may	come	in	many	forms.	Although	I	used	direct	text	to	highlight	similarity	in	study	4,	study	2	highlighted	another	form	of	‘seeing	similar’	–	that	of	anthropomorphism.	Anthropomorphism	has	been	found	to	promote	a	sense	of	kinship	(Veer,	2013),	and	its	use	is	prevalent	in	marketing	fields,	as	evidenced	by	the	familiar	use	of	brand	spokes-characters	such	as	Kellogg’s	Tony	the	Tiger,	and	the	United	States	Forestry	Service’s	Smokey	the	Bear.	Not	surprisingly,	conservation	workers	have	found	that	animals	that	appear	to	embody	more	human	characteristics	are	indeed	perceived	as	more	charismatic	and	appealing	by	the	public	(Ducarme,	Luque,	&	Courchamp,	2012),	and		195	conservation	charities	have	already	attempted	to	leverage	anthropomorphism	as	a	conservation	tool	(Chan	2012;	Tam,	Lee,	&	Chao,	2013),	with	some	degrees	of	success.	As	has	been	previously	noted,	however,	individuals	have	a	tendency	to	anthropomorphize	more	when	they	are	socially	excluded	(Epley,	Waytz,	&		Cacioppo	2007),	and	to	dehumanize	more	when	they	are	socially	included	(Lucas,	&	Livingston,	2014;	Waytz,	&	Epley,	2012).	This	research	serves	to	sound	a	note	of	caution	regarding	the	potential	damage	that	might	conceivably	arise	from	incautious	use	of	such	tools.	Further	research	looking	into	the	mechanisms	by	which	anthropomorphism	works	in	this	setting	would	therefore	be	recommended.			In	summary,	this	research	suggests	that	marketers	involved	in	cause	marketing	and	prosocial	appeals	would	do	well	to	make	an	effort	to	understand	the	specific	motivations	of	their	donors,	and	carefully	tailor	their	charitable	marketing	materials	accordingly,	to	ensure	that	they	are	relevant	and	appealing	to	their	key	supporters	and	customers.			4.4	Research	Limitations	and	Directions	for	Future	Research	There	are	a	number	of	general	areas	highlighted	by	this	research	that	would	benefit	from	additional	future	research.	One	area	I	did	not	have	time	to	fully	explore	was	whether	anthropomorphism	could	be	harnessed	in	any	positive	ways,	to	increase	feelings	of	similarity	and	social	connection	towards	out-group	targets.	In	view	of	past	research	on	intergroup	comparison	and	anthropomorphism	(Leyens,	Rodriguez-Perez,	Rodriguez-Torres,	Gaunt,	Paladino,	Vaes,	&	Demoulin,	2001),	this		196	appears	to	be	an	area	worthy	of	future	study.		Likewise,	I	was	not	able	to	explicitly	explore	whether	anthropomorphism	impacted	donation	behaviour,	rather	than	just	inclusive	behaviour,	and	therefore	this	would	be	another	area	that	would	benefit	from	further	research.				Another	connected	area	of	research	that	I	was	not	able	to	pursue	is	the	area	of	mimicry.	Mimicry,	in	this	case,	is	defined	as	the	apparently	unconscious,	or	automatic,	imitation	of	gestures,	behaviours,	facial	expressions,	speech	and	movements.	Not	only	has	past	research	indicated	that	mimicking	increases	a	sense	of	rapport,	interpersonal	closeness,	and	liking	and	potentially	boosts	prosocial	behaviour	(Chartrand	&	Bargh,	1999;	Lakin,	Jefferis,	Cheng,	&	Chartrand,	2003;	van	Baaren,	Holland,	Kawakami,	&	van	Knipperberg,	2004),	but	people	have	also	been	shown	to	mimic	more	when	they	feel	they	are	similar,	even	if	similarity	is	incidental	(Guéguen,	&	Martin,	2009).	Likewise,	evidence	has	been	offered	in	support	of	the	argument	that	connectedness	to	the	mimicry	target	may	moderate	the	behaviour	(van	Baaren,	Janssen,	Chartrand,	&	Dijksterhuis,	2009).	Further	research	also	shows	that	people	are	more	likely	to	mimic	each	other	when	they	are	excluded	and	seeking	to	socially	connect	(Lakin	&	Chartrand,	2003;	Lakin,	Chartrand,	&	Arkin,	2008;	Uldall,	Hall,	&	Chartrand,	2003).	Evidence	suggests	that	people	with	a	more	interdependent	self-construal	may	mimic	more	than	people	with	independent	self-construal	(Ashton-James,	van	Baaren,	Chartrand,	Decety,	Karremans,	2007;	van	Barren,	Maddux,	Chartrand,	de	Bouter,	&	van	Knippenberg,	2003),	making	mimicry		197	a	potentially	potent	tool	for	use	with	individuals	high	in	interdependent	self-construal.			On	a	more	specific	level,	future	research	in	this	area	could	benefit	from	exploring	whether	these	findings	are	applicable	in	the	context	of	cultural	self-construal.	Past	research	has	demonstrated	that	individualistic	regions	within	a	wider	country	may	be	more	predisposed	to	donating	to	charity	(Kemmelmeier,	Jambor,	&	Letner,	2006),	and	it	would	be	useful	to	examine	how	experiences	of	inclusion	might	moderate	cultural	self-construal.	Future	research	might	also	look	to	replicate	these	results,	by	priming	individual	self-construal	within	a	study	framework,	as	opposed	to	measuring	it.	Study	5	also	highlighted	another	potential	line	of	future	research:	that	of	the	influence	of	public/private	settings	on	the	mechanisms	involved	in	this	current	research.	Whether	these	findings	only	hold	true	in	a	private	setting	would	be	a	very	valuable	area	to	explore.			In	terms	of	generalizability,	it	would	be	useful	to	replicate	some	of	these	studies	in	the	field	if	at	all	possible,	as	well	as	with	more	diverse	and	greater	samples,	in	order	to	be	sure	of	the	robustness	of	the	results.	Furthermore,	I	measured	only	inclusive	behaviour	and	charitable	donation.	Future	research	might	look	to	ascertain	whether	these	findings	might	extend	to	other	prosocial	dependent	variables,	such	as	volunteer	work.	Moreover	my	research	looked	at	one	specific	out-group,	non-human	animals.	Past	research	has	found	that	human-animal	similarity	primes	may	result	in	reduced	prejudice	towards	out-groups	containing	other	humans	(Hodson	&		198	Costello,	2010).	It	would,	therefore,	be	interesting	to	see	if	the	effects	with	self-construal,	witnessed	in	my	research,	extend	to	these	groups	as	well.	On	the	topic	of	non-human	species	it	may	also	be	noted	that	the	future	belonging	manipulation	used	in	study	2	is	worded	to	specifically	manipulate	feelings	of	human	belonging.	While	this	manipulation	has	been	used	to	good	effect	in	past	research	it	is	possible	that	it	may	have	primed	‘human	association’	as	opposed	to	general	belonging.	In	consideration	of	the	interspecies	context	of	this	research,	future	research	might	look	to	test	whether	the	effects	hold	true	in	a	non-human	specific	belonging	context	moving	forward.				Another	area	that	would	also	be	potentially	valuable	to	explore	in	terms	of	future	research	would	be	to	further	tease	apart	what	other	self-benefits,	beyond	social	connection,	may	drive	prosocial	behaviour	as	a	factor	of	self-construal	orientation.	As	previously	noted,	self-benefits	may	include	personal	benefits,	such	as	those	that	involve	increasing	pleasure/decreasing	suffering	(for	example:	increasing	social	inclusion,	personal	group	status;	highlighting	individual	distinctiveness),	or	social	identity	based	benefits,	such	as	those	that	confirm	a	positive	group	stereotype,	or	disconfirm	a	negative	group	stereotype.	More	broadly,	future	research	might	chose	to	explore	the	connection	between	empathy	and	the	mechanisms	detailed,	as	well	as	to	look	at	extending	this	research	into	the	existing	research	on	guilt,	self-construal	and	similarity	to	target	(Chen	&	Moosmayer,	2018)	in	a	prosocial	setting.				199	Research	Limitations.	In	terms	of	research	limitations,	as	previously	noted	many	of	the	studies	may	be	noted	to	lack	power,	in	places	at	least,	and	at	times	p-values	approached	significance	only.		Whilst	study	3	and	4	were	well	powered	to	detect	the	conditional	effects	in	each	regression	model,	they	may	not	have	been	sufficiently	well	powered	to	detect	all	of	the	smaller	interaction	effects.	In	view	of	this,	the	fact	that	the	trends	and	patterns	seen	in	the	interactions	were	repeated	over	a	number	of	samples	in	the	different	studies,	adds	at	least	some	robustness	to	the	findings.	These	are	admittedly	first	steps	in	a	novel	area	of	research,	and	to	be	more	sure	of	the	robustness	of	the	findings	it	would	be	advisable	to	attempt	to	further	replicate	these	in	future	research.			Another	limitation	of	my	findings	may	result	from	the	demographic	differences	seen	in	my	samples.	Whilst	student	samples,	typically	drawn	from	Western,	Educated,	Industrialized,	Rich,	and	Democratic	(WEIRD)	societies,	are	commonly	used	in	behaviour	science	research,	it	is	acknowledged	that	they	may	have	unusual	features	that	limit	universal	replicability	(Heinrich,	Heine,	&	Norenzayan,	2010).	In	particular	attention	should	be	drawn	to	the	high	Asian	ethnic	bias	in	many	of	my	lab	sample	demographics,	compared	to	my	MTurk	samples.	Demographic	differences,	in	terms	of	both	age	and	ethnicity,	were	quite	large	between	all	of	my	MTurk	studies	and	my	lab	studies.	These	demographic	differences	may	be	considered	a	weakness,	or	a	strength,	considering	the	patterns	across	all	studies.	The	fact	that	interaction	trends	were	repeated	with	reasonable	consistency,	despite	these	changes	in		200	demographics,	may	speak	to	the	potential	for	my	findings	to	be	generalizable	beyond	specific	populations.	A	word	of	caution	must	be	noted,	however,	regarding	the	failure	of	the	short	field	study	to	replicate	the	results	of	the	previous	4	studies.			4.5	Concluding	Remarks		In	summary,	this	dissertation	provides	novel	insights	into	the	prosocial	actions	and	intentions	directed	towards	animal	out-groups,	as	a	function	of	self-construal.	More	specifically,	I	focus	on	the	whens	and	hows	that	describe	the	circumstances	under	which	self-construal	orientation	and	inclusion	status	will	interact	to	produce	more,	or	less,	prosocial	results.	Over	five	empirical	studies,	the	current	research	examines	a	key	factor	(social	inclusion)	that	interacts	with	independent	self-construal	to	shift	prosocial	behaviour,	in	the	form	of	inclusive	behaviour	and	donation	intentions	towards	out-group	members	that	are	non-human	animals.	I	also	investigate	how	perceptions	of	similarity	(vs.	difference)	with	an	out-group	may	moderate	prosocial	intentions,	and	how	assessments	of	connection	to	the	cause	may	mediate	prosocial	intentions,	both	as	a	function	of	self-construal.			In	study	1	and	2,	I	demonstrate	that	independent	self-construal	is	positively	related	to	the	inclusion	of	out-group	members	into	a	self-formed	in-group	under	normal	circumstances,	but	that	following	an	affirmation	of	social	inclusion	the	relationship	reverses.	Furthermore,	I	suggest	that	this	pattern	is	as	a	result	of	a	desire	to	fulfill	social	connection	needs	for	individuals	with	a	high	independent	self-construal,	and	I	show	that	this	effect	is	not	as	a	result	of	positive	affect.	In	study	3	and	4,	I		201	demonstrate	the	same	pattern	in	individuals	high	in	independent	self-construal,	only	this	time	with	donation	intentions	as	expressed	towards	an	out-group	cause.		Additionally,	I	show	that	framing	the	cause	as	similar	(vs.	different)	has	the	effect	of	increasing	donation	support	for	individuals	high	in	independent	self-construal.	I	propose	that	this	is	as	a	result	of	an	increased	interest	for	social	connectedness,	and	demonstrate	that	connection	to	the	cause	acts	as	a	mediator	to	predict	donation	intentions.		Taken	together,	the	studies	make	theoretical	and	practical	contributions	to	established	literature	streams	in	the	areas	of	prosocial	behaviour,	self-construal,	and	the	downstream	effects	of	exclusion	and	belonging	on	the	self.	I	hope	that	this	dissertation	provides	useful	insights	in	the	area	of	prosocial	behaviour	as	specifically	directed	towards	out-groups,	including	animal	out-groups.					 		202	Chapter	5:	Figures		Figure	1.	Social	Identity	vs.	Personal	Identity							©	Brewer,	1991.			 		203	Figure	2.	Optimal	Distinctiveness	Model	of	the	Self				©	Brewer,	1991		 		204	Figure	3.	Construals	of	Self			a) Independent	view	of	self,	b)			Interdependent	view	of	self			©	Markus	&	Kitayama,	1991.		 		205	Figure	4.	Self-Construal	and	Prosocial	Behaviour	Path	(Studies	1,	2	and	3)						 																													 	Independent SC Interdependent SC Social	Connection	Monitoring	Assured Belonging  Out-group recipient  In-group recipient 	206	Figure	5.	Self-Construal	and	Prosocial	Behaviour	Path	(Study	4)																					Independent SC Social	Connection	Monitoring	Assured Belonging  Similar out-group recipient  Different out-group recipient Activates	in-group	bias	and	differentiation		207	Figure	6.	Sample	Ethnicity	by	Percentage	(Pre-test	Study)			 				 	0	10	20	30	40	50	60	70	Chinese	 South	Asian	Korean	 South	East	Asian	Caucasian	Japanese	 Other	Ethnicity	by	Percentage		208	Figure	7.	Sample	Age	by	Percentage	(Pre-test	Study)						 	0	10	20	30	40	50	60	Under	21	years	old	21-25	years	old	 26-30	years	old	 Over	30	years	old	Age	by	Percentage		209	Figure	8.	Sample	Ethnicity	by	Percentage	(Study	1)										 	0	10	20	30	40	50	60	70	80	90	Caucasian	 Black	 Other	Ethnicity	by	Percentage		210	Figure	9.	Sample	Age	by	Percentage	(Study	1)							 	0	5	10	15	20	25	30	35	40	Under	21	years	old	21-30	years	old	31-40	years	old	41-50	years	old	51-60	years	old	Over	60	years	old	Age	by	Percentage		211	Figure	10.	Sample	Ethnicity	by	Percentage	(Study	2)									 	0	10	20	30	40	50	60	70	80	90	Asian	 Caucasian	 East	Indian	 Black	 Other	Ethnicity	by	Percentage		212	Figure	11.	Proportion	of	In-group	made	up	of	Animals	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	2)					Figure	11.	a)	 	 	 	 	 Figure	11.	b)		Regression	Lines	with	Johnson-Neyman	Points	showing	the	regions	of	a)	independent	self-construal	and	b)	interdependent	self-construal	values	(filled	areas	above	-0.92	and	below	0.94	respectively)	for	which	a	floodlight	test	would	reveal	significant	differences	between	the	two	groups.					 		213	Figure	12.	Floodlight	Analysis	of	Proportion	of	In-group	made	up	of	Animals	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	2)							Figure	12.	a)	 	 	 	 	 Figure	12.	b)		Floodlight	analysis	of	the	conditional	effect	of	the	condition	(X;	control	vs.	inclusion)	on	inclusion	of	animals	in	the	in-group	θXàY	as	a	function	of	M	for	a)	independent	and	b)	interdependent	self-construal	orientation,	along	with	confidence	bands	(Bauer	&	Curran,	2005;	Preacher	et	al,	2006;	Rogosa,	1980;	Spiller	et	al.	2013.)			As	can	be	seen	in	figure	a)	when	M≥-.92	the	confidence	bands	are	entirely	above	the	zero	and	in	figure	b)	when	M≤.94	and	M≥16.77	the	confidence	bands	are	entirely	above	the	zero.				 		214	Figure	13.	Entitativity	Scores	of	the	In-group	as	a	Function	of	Condition	(Inclusion	vs.	Control)	and	Independent	Self-Construal	(Study	2)			13.	a)	Independent	Self-Construal	 							 13.	c)	Interdependent	Self-Construal			A	Visual	representation	of	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	ratings	of	entitativity	of	the	formed	in-group	(Y)	by	self-construal	(W)		Figure	13.a)	shows	that	while	normally	(control)	there	appears	to	be	a	strong	positive	correlation	between	independent	self-construal	and	subsequent	rating	of	the	entitativity	of	a	self-formed	in-group,	this	relationship	reduces	with	an	inclusion	condition.			Figure	13.c)	shows	no	change	in	entitativity	rating	as	a	result	of	the	inclusion	condition	for	interdependent	self-construal.				 		215	Figure	14.	Anthropomorphism	of	Dog	Picture	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	2)				Figure	14.	a)	 	 	 	 	 	 Figure	14.	b)		Figure	14.	a)	A	visual	representation	of	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	ratings	of	(social)	anthropomorphism	(Y)	by	independent	self-construal	(W).			Figure	14.	b)	A	visual	representation	of	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	ratings	of	(social)	anthropomorphism	(Y)	by	independent	self-construal	(W)	while	controlling	for	individual	differences	in	anthropomorphism.					 		216	Figure	15.	Floodlight	Analysis	of	Ratings	of	Anthropomorphism	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	2)										 																											95%	CI	Upper	Limit					 							Point	Estimate					 							95%	CI	Lower	Limit						Figure	15.	c)	The	conditional	effect	of	the	condition	(X;	control	vs.	inclusion)	on	ratings	of	(social)	anthropomorphism	(θXàY),	as	a	function	of	independent	self-construal	orientation	(W),	controlling	for	individual	differences	in	anthropomorphism.	As	can	be	seen	when	W	≥1.8451	the	confidence	bands	are	entirely	below	the	zero.						 		217	Figure	16.	Sample	Ethnicity	by	Percentage	(Study	3)							 	0	10	20	30	40	50	60	70	80	90	Caucasian	 Black	 LaMn	American	South	East	Asian	Asian	 Other	Ethnicity	by	Percentage		218	Figure	17.	Sample	Age	by	Percentage	(Study	3)						 	0	10	20	30	40	50	60	Under	21	years	old	21-30	years	old	31-40	years	old	41-50	years	old	51-60	years	old	Over	60	years	old	Age	by	Percentage		219	Figure	18.	Donation	Intentions	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	3)																				Figure	18.	b)	Regression	lines	showing	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	donation	intention	(Y)	by	independent	self-construal	(W)	while	controlling	for	gender.			Note:	Johnson-Neyman	Point	showing	the	region	of	independent	self-construal	values	(filled	area	above	9.1342)	for	which	a	floodlight	test	would	reveal	significant	differences	between	the	two	groups.				 		220	Figure	19.	Floodlight	Analysis	of	Donation	Intentions	as	a	Function	of	Condition	(Control	+	Affect	vs.	Inclusion)	and	Independent	Self-Construal	(Study	3)		 					 								 				 										 			95%	CI	Upper	Limit		 			Point	Estimate			 			95%	CI	Lower	Limit						Figure	19.	The	conditional	effect	of	the	condition	(X;	control	vs.	inclusion)	on	donation	intentions	(θXàY),	as	a	function	of	independent	self-construal	orientation	(W),	controlling	for	gender.	As	can	be	seen	when	W	≥	9.13	the	confidence	bands	are	entirely	below	the	zero.				 		221	Figure	20.	Connection	to	Cause	as	a	Function	of	Condition	(Control	vs.	Inclusion)	and	Self-Construal	(Study	3)															Figure	20.	a)	 	 	 	 	 Figure	20.b)			Figure	20.	a)	A	visual	representation	of	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	ratings	of	connection	to	cause	(Y)	by	independent	self-construal	(W).			Figure	20.	b)	A	visual	representation	of	the	moderation	of	the	effect	of	condition	(X;	control	vs.	inclusion)	on	ratings	of	connection	to	cause	(Y)	by	interdependent	self-construal	(W).			 		222	Figure	21.	Sample	Ethnicity	by	Percentage	(Study	4)											 	0	10	20	30	40	50	60	70	80	90	Caucasian	 Black	 LaMn	American	Asian	 South	East	Asian	Other	Ethnicity	by	Percentage		223	Figure	22.	Sample	Age	by	Percentage	(Study	4)							 	0	5	10	15	20	25	30	35	40	Under	21	years	old	21-30	years	old	31-40	years	old	41-50	years	old	51-60	years	old	Over	60	years	old	Age	by	Percentage		224	Figure	23.	Donation	to	Cause	as	a	Function	of	Poster	Frame	(Similar	vs.	Different)	and	Interdependent	Self-Construal	(Study	4)																					Figure	23.	Regression	lines	showing	the	moderation	of	the	effect	of	poster	(X;	similar	vs.	different)	on	donation	intentions	(Y)	by	interdependent	self-construal	(W).	Note:	Johnson-Neyman	Point	showing	the	region	of	interdependent	self-construal	values	(filled	area	below	-8.1141)	for	which	a	floodlight	test	would	reveal	significant	differences	between	the	two	groups.				 		225	Figure	24.	Floodlight	Analysis	of	Donation	Intentions	as	a	Function	of	Poster	Style	(Similar	vs.	Different)	and	Interdependent	Self-Construal	(Study	4)			 							 											95%	CI	Upper	Limit			 							 											Point	Estimate				 							 											95%	CI	Lower	Limit													Figure	24.	The	conditional	effect	of	the	poster	(X;	similar	vs.	different)	on	donation	intention	(θXàY),	as	a	function	of	interdependent	self-construal	orientation	(W).	As	can	be	seen	when	W	≤	-8.1141	the	confidence	bands	are	entirely	below	the	zero.				 		226	Figure	25.	Donation	to	Cause	as	a	Function	of	Condition	(Control	vs.	Inclusion)	Poster	Type	(Similar	vs.	Different)	and	Independent	Self-Construal	(Study	4)						High	Independent	Self-Construal	(Z)	=	10.67		θXWàY	l	Z=	10.67	=-4.2259,	p=.01												Low	Independent	Self-Construal	(Z)	=	-10.33		θXWàY	l	Z=	-10.33	=3.2398,	p=.04					Figure	25.	The	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	donation	intention	(Y),	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender.				A	test	of	the	conditional	interaction	(Condition	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	donation	intention	is	significant	at	both	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-4.2259,	p=.01)	as	well	as	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=3.2398,	p=.04).				 		227	Figure	26.	Floodlight	Analysis	of	Donation	Intentions	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)													 																			95%	CI	Upper	Limit		 																Point	Estimate																		95%	CI	Lower	Limit							Figure	26.	The	conditional	two-way	interaction	between	condition	and	poster	style	(θXWàY	)	as	a	function	of	independent	self-construal	(Z)	when	controlling	for	gender.	As	can	be	seen	when	Z	≤-9.828	and	≥6.0581	the	confidence	bands	are	entirely	above	and	below	the	zero	respectively.					 		228	Figure	27.	Actual	Cash	Donation	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)						Figure	27.	A	visual	representation	of	the	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	actual	cash	donation	(Y),	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender.						 		229	Figure	28.	Connection	to	Cause	as	a	Function	of	Independent	Self-Construal,	Inclusion	Manipulation	and	Similar	vs.	Difference	Poster	Frame	(Study	4)		 							High	Independent	Self-Construal	(Z)	=	10.67		θXWàY	l	Z=	10.67	=-1.6853,	p=.002												Low	Independent	Self-Construal	(Z)=-10.33		θXWàY	l	Z=	-10.33	=0.9625,	p=.08					Figure	28.	The	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	connection	to	cause	(Y),	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender.				A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	connection	to	cause	was	significant	at	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-1.6853,	p=.002)	and	near	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=0.9625,	p=.08).			 		230	Figure	29.	Floodlight	Analysis	of	Connection	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)															 			 																																	95%	CI	Upper	Limit			 																Point	Estimate			 																95%	CI	Lower	Limit						Figure	29.	The	conditional	two-way	interaction	between	condition	and	poster	style	(θXWàY	)	as	a	function	of	independent	self-construal	(Z)	when	controlling	for	gender.	As	can	be	seen	when	Z	≤-11.6991	and	≥3.6638	the	confidence	bands	are	entirely	above	and	below	the	zero	respectively.					 		231	Figure	30.	Empathy	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)			 		 				High	Independent	Self-Construal	(Z)	=	10.67		θXWàY	l	Z=	10.67	=-0.8689,	p=.07												Low	Independent	Self-Construal	(Z)	=	-10.33		θXWàY	l	Z=	-10.33	=1.0141,	p=.03						Figure	30.	The	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	empathy	for	cause	(Y),	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender.				A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	empathy	to	cause	was	near	significant	at	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-0.8689,	p=.07)	and	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=1.0141,	p=.03).	 	 		232	Figure	31.	Floodlight	Analysis	of	Empathy	for	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)														 95%	CI	Upper	Limit		 	 Point	Estimate			 	95%	CI	Lower	Limit							Figure	31.	The	conditional	two-way	interaction	between	condition	and	poster	style	(θXWàY	)	as	a	function	of	independent	self-construal	(Z)	when	controlling	for	gender.	As	can	be	seen	when	Z	≤-8.0906	and	≥11.9537	the	confidence	bands	are	entirely	above	and	below	the	zero	respectively.									233	Figure	32.	Similarity	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)		 				High	Independent	Self-Construal	(Z)	=	10.67		θXWàY	l	Z=	10.67	=-1.6398,	p=.004												Low	Independent	Self-Construal	(Z)	=	-10.33		θXWàY	l	Z=	-10.33	=0.7873,	p=.16						Figure	32.	The	conditional	effect	of	poster	style	(X;	similar	vs.	different)	on	similarity	to	cause	(Y),	as	a	function	of	condition	(W;	control	vs.	inclusion)	and	independent	self-construal	(Z)	from	a	moderated	moderation	model,	controlling	for	gender.				A	test	of	the	conditional	interaction	(Inclusion	X	Poster	Style)	at	values	of	Independent	Self-Construal	revealed	that	the	effect	on	similarity	to	cause	was	significant	at	high	independent	self-construal	(θXWàY	l	Z=	10.67	=-1.6398,	p=.004)	but	not	significant	with	low	independent	self-construal	(θXWàY	l	Z=	-10.33	=0.7873,	p=.16).					234	Figure	33.	Floodlight	Analysis	of	Similarity	to	Cause	as	a	Function	of	Poster	Style	(Similar	vs.	Different),	Condition	(Control	vs.	Inclusion)	and	Independent	Self-Construal	(Study	4)		 																		 95%	CI	Upper	Limit			 Point	Estimate			95%	CI	Lower	Limit					Figure	33.	The	conditional	two-way	interaction	between	condition	and	poster	style	(θXWàY)	as	a	function	of	independent	self-construal	(Z)	when	controlling	for	gender.	As	can	be	seen	when	Z	≤-15.6051	and	≥3.5873	the	confidence	bands	are	entirely	above	and	below	the	zero	respectively.						 		235	Figure	34	Sample	Age	by	Percentage	(Study	5)							 	0	5	10	15	20	25	30	18-30	years	old	31-40	years	old	41-50	years	old	51-60	years	old	Over	60	years	old	Age	by	Percentage		236	Figure	35.	Conceptual	Models	(Hayes,	2013;	2018)	used	in	Studies	2-5	Conceptual	Moderation	Model	1	(Hayes,	2013)	used	in	Study	2												Conceptual	Moderation	Model	1	(Hayes,	2013)	used	in	Study	3	&	4															Percentage	of	in-group	made	up	of	animals					Independent	self-construal					Inclusion	Manipulation					Donation	Intention					Independent	self-construal					Inclusion	Manipulation		237	Conceptual	Moderated	Mediation	Model	7	(Hayes,	2018)	used	in	Study	3													Conceptual	Moderated	Moderation	Model	3	(Hayes,	2018)	used	in	Study	4																	Donation	Intention				Independent	self-construal					Inclusion	Manipulation								Poster	Type				Donation	Intention			Inclusion	Manipulation					Independent	Self-construal					Connection	to	Cause		238	Conceptual	Moderated	Moderated	Mediation	Model	11	(Hayes,	2018)	used	in	Study	4															 			Independent	Self-construal				Donation	Intention				Poster	Type					Inclusion	Manipulation		 			 Connection	to	Cause		239	References			Aaker,	J.	L.	&	Lee,	A.	Y.	(2001).	“I”	seek	pleasures	and	“we”	avoid	pains:	The	role	of	self-regulatory	goals	in	information	processing	and	persuasion.	Journal	of	Consumer	Research,	28(1),	33-49.	Aaker,	J.	L.	&	Williams,	P.	(1998).	Empathy	versus	pride:	The	influence	of	emotional	appeals	across	cultures.	Journal	of	Consumer	Research,	25(3),	241-261.	Aberson,	C.	L.	(2011).	Applied	power	analysis	for	the	behavioural	sciences.	Routledge.	Abrams,	D.	&	Hogg,	M.	A.	(1988).	Comments	on	the	motivational	status	of	self-esteem	in	social	identity	and	intergroup	discrimination.	European	Journal	of	Social	Psychology,	18(4),	317-334.	Aderman,	D.	(1972).	Elation,	depression,	and	helping	behaviour.	Journal	of	Personality	and	Social	Psychology,	24(1),	91.	Aiken,	L.	S.,	West,	S.	G.,	&	Reno,	R.	R.	(1991).	Multiple	regression:	Testing	and	interpreting	interactions.	Sage.	Amiot,	C.	E.	&	Bastian,	B.	(2017).	Solidarity	with	Animals:	Assessing	a	relevant	dimension	of	social	identification	with	animals.	PLoS	One,	12(1).	Andersson,	L.	(1998).	Loneliness	research	and	interventions:	A	review	of	the	literature.	Aging	&	Mental	Health,	2(4),	264-274.	Aquino,	K.	&	Reed,	A.,	II.	(2002).	The	self-importance	of	moral	identity.	Journal	of	Personality	and	Social	Psychology,	83(6),	1423–1440.	Aristotle,	Jowett,	B.,	&	Davis,	H.	W.	C.	(1920).	Aristotle's	Politics.	Oxford:	At	the	Clarendon	Press.	Arndt,	J.,	Greenberg,	J.,	&	Cook,	A.	(2002).	Mortality	salience	and	the	spreading	activation	of	worldview-relevant	constructs:	Exploring	the	cognitive	architecture	of	terror	management.	Journal	of	Experimental	Psychology:	General,	131(3),	307.	Arnocky,	S.,	Stroink,	M.,	&	DeCicco,	T.	(2007).	Self-construal	predicts	environmental	concern,	cooperation,	and	conservation.	Journal	of	Environmental	Psychology,	27(4),	255-264.		240	Ashton–James,	C.,	Van	Baaren,	R.	B.,	Chartrand,	T.	L.,	Decety,	J.,	&	Karremans,	J.	(2007).	Mimicry	and	me:	The	impact	of	mimicry	on	self–construal.	Social	Cognition,	25(4),	518-535.	Bagozzi,	R.	P.,	Verbeke,	W.,	&	Gavino	Jr,	J.	C.	(2003).	Culture	moderates	the	self-regulation	of	shame	and	its	effects	on	performance:	The	case	of	salespersons	in	The	Netherlands	and	the	Philippines.	Journal	of	Applied	Psychology,	88(2),	219.	Bartz,	J.	A.,	Tchalova,	K.,	&	Fenerci,	C.	(2016).	Reminders	of	social	connection	can	attenuate	anthropomorphism:	A	replication	and	extension	of	Epley,	Akalis,	Waytz,	and	Cacioppo	(2008).	Psychological	Science,	27(12),	1644-1650.	Bastian,	B.,	Costello,	K.,	Loughnan,	S.,	&	Hodson,	G.	(2012).	When	closing	the	human–animal	divide	expands	moral	concern:	The	importance	of	framing.	Social	Psychological	and	Personality	Science,	3(4),	421-429.	Bastian,	B.,	Loughnan,	S.,	Haslam,	N.,	&	Radke,	H.	R.	(2012).	Don’t	mind	meat?	The	denial	of	mind	to	animals	used	for	human	consumption.	Personality	and	Social	Psychology	Bulletin,	38(2),	247-256.	Batson,	C.	D.	(1987).	Prosocial	motivation:	Is	it	ever	truly	altruistic?	In	Advances	in	Experimental	Social	Psychology	20,	65-122.	Academic	Press.	Batson,	C.	D.,	Ahmad,	N.,	&	Stocks,	E.	L.	(2011).	Four	forms	of	prosocial	motivation:	Egoism,	altruism,	collectivism,	and	principlism.	na.	Batson,	C.	D.	&	Shaw,	L.	L.	(1991).	Evidence	for	altruism:	Toward	a	pluralism	of	prosocial	motives.	Psychological	Inquiry,	2(2),	107-122.	Batson,	C.	D.	&	Shaw,	L.	L.	(1991).	Encouraging	words	concerning	the	evidence	for	altruism.	Psychological	Inquiry,	2(2),	159-168.	Batt,	S.	(2009).	Human	attitudes	towards	animals	in	relation	to	species	similarity	to	humans:	A	multivariate	approach.	Bioscience	Horizons,	2(2),	180-190.	Bauer,	D.	J.,	&	Curran,	P.	J.	(2005).	Probing	interactions	in	fixed	and	multilevel	regression:	Inferential	and	graphical	techniques.	Multivariate	Behavioural	Research,	40(3),	373-400.		241	Baumeister,	R.	F.,	Brewer,	L.	E.,	Tice,	D.	M.,	&	Twenge,	J.	M.	(2007).	Thwarting	the	need	to	belong:	Understanding	the	interpersonal	and	inner	effects	of	social	exclusion.	Social	and	Personality	Psychology	Compass,	1(1),	506-520.	Baumeister,	R.	F.,	DeWall,	C.	N.,	Ciarocco,	N.	J.,	&	Twenge,	J.	M.	(2005).	Social	exclusion	impairs	self-regulation.	Journal	of	Personality	and	Social	Psychology,	88(4),	589.	Baumeister,	R.	F.,	&	Leary,	M.	R.	(1995).	The	need	to	belong:	Desire	for	interpersonal	attachments	as	a	fundamental	human	motivation.	Psychological	Bulletin,	117(3),	497.	Bekoff,	M.	(2007).	Animals	matter:	A	biologist	explains	why	we	should	treat	animals	with	compassion	and	respect.	Shambhala	Publications.	Bernstein,	M.	J.,	Sacco,	D.	F.,	Brown,	C.	M.,	Young,	S.	G.,	&	Claypool,	H.	M.	(2010).	A	preference	for	genuine	smiles	following	social	exclusion.	Journal	of	Experimental	Social	Psychology,	46(1),	196-199.	Bjerke,	T.,	Østdahl,	T.,	&	Kleiven,	J.	(2003).	Attitudes	and	activities	related	to	urban	wildlife:	Pet	owners	and	non-owners.	Anthrozoös,	16(3),	252-262.	Block,	L.	G.	(2005).	Self-referenced	fear	and	guilt	appeals:	The	moderating	role	of	self-construal.	Journal	of	Applied	Social	Psychology,	35(11),	2290-2309.	Bond,	M.	H.	&	Hewstone,	M.	(1988).	Social	identity	theory	and	the	perception	of	intergroup	relations	in	Hong	Kong.	International	Journal	of	Intercultural	Relations,	12(2),	153-170.	Bowd,	A.	D.	(1984).	Fears	and	understanding	of	animals	in	middle	childhood.	Journal	of	Genetical	Psychology,	145,	143-144.		Branscombe,	N.	R.,	Ellemers,	N.,	Spears,	R.,	&	Doosje,	B.	(1999).	The	context	and	content	of	social	identity	threat.	Social	identity:	Context,	commitment,	content,	35-58.	Brewer,	M.	B.	(1979).	In-group	bias	in	the	minimal	intergroup	situation:	A	cognitive-motivational	analysis.	Psychological	Bulletin,	86(2),	307.	Brewer,	M.	B.	(1991).	The	social	self:	On	being	the	same	and	different	at	the	same	time.	Personality	and	Social	Psychology	Bulletin,	17(5),	475-482.		242	Brewer,	M.	B.	(1993).	Social	identity,	distinctiveness,	and	in-group	homogeneity.	Social	Cognition,	11(1),	150-164.	Brewer,	M.	B.	(1999).	The	psychology	of	prejudice:	Ingroup	love	and	outgroup	hate?	Journal	of	Social	Issues,	55(3),	429-444.	Brewer,	M.	B.	&	Kramer,	R.	M.	(1986).	Choice	behaviour	in	social	dilemmas:	Effects	of	social	identity,	group	size,	and	decision	framing.	Journal	of	Personality	and	Social	Psychology,	50(3),	543.	Brewer,	M.	B.	&	Gardner,	W.	(1996).	Who