UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reconciling resistance : women's postnatal physical activity decision-making Liva, Sarah 2017

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_february_liva_sarah.pdf [ 3.48MB ]
Metadata
JSON: 24-1.0363008.json
JSON-LD: 24-1.0363008-ld.json
RDF/XML (Pretty): 24-1.0363008-rdf.xml
RDF/JSON: 24-1.0363008-rdf.json
Turtle: 24-1.0363008-turtle.txt
N-Triples: 24-1.0363008-rdf-ntriples.txt
Original Record: 24-1.0363008-source.json
Full Text
24-1.0363008-fulltext.txt
Citation
24-1.0363008.ris

Full Text

RECONCILING RESISTANCE: WOMEN’S POSTNATAL PHYSICAL ACTIVITY DECISION-MAKING by   Sarah Liva   MSN, The University of British Columbia, 2011   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Nursing)   THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)   December 2017   © Sarah Liva, 2017 ii  Abstract  Background: Nearly half of women have reported reducing their physical activity below recommended levels during the postpartum period and evidence suggests these reductions are sustained beyond the postnatal year. The modestly effective interventions to increase postnatal physical activity may indicate a lack of understanding about how women make postnatal physical activity choices. Research Question: What are women’s decision-making processes about physical activity in the postnatal period? Methods: I used a qualitative grounded theory methodology to develop a core category and interrelated categories explaining women’s physical activity decision-making processes during the postpartum period. Thirty women within a year of childbirth, were recruited using flyers and social media, and interviewed after they completed 3-day diaries. I open-coded the data to identify a core category, and then selectively coded, and theoretically coded to articulate relationships between the core category and other categories using constant comparative analysis. Findings: The core category, reconciling resistance, comprised three processes: gauging, engaging, and adjusting, and explained how women resolved their main concern in making physical activity decisions to minimize discord between their physical activity desires and actual physical activity patterns. Participants experienced personal embodiment, relational, environmental and physical activity centrality resistance that they reconciled through gauging the risks and accessibility of choosing their desired activities and deciding whether they regarded the activities as essential for them postnatally. The women engaged to either push through, hold back, or hold still on pursuing their desired activities and adjusted their choices of strategies in response to their experiences of engaging. Only the women who constructed their physical activities as supporting their own and others’ needs and experienced limited personal embodiment resistance (e.g., fatigue, injury) pushed through. Most women held back or held still by redefining the types of activity that they believed could support their own and others’ needs. Implications: Providing mothers with anticipatory guidance and education about safely returning to physical activity is an important strategy for health care providers to support postnatal physical iii  activity. Increasing knowledge about physical activity choices in the context of women’s need satisfaction could support strategies that are acceptable to women within the postpartum period. iv  Lay Summary  This study explored new mothers’ physical activity decision-making to understand why nearly half of new mothers have reported reducing their physical activity to below recommended levels and healthcare interventions to promote their activity have demonstrated mixed effectiveness. Thirty mothers with babies under one year old completed activity diaries and were interviewed. Most women held back from pursuing their desired physical activities because of resistance (beliefs) that made those activities seem non-essential, too risky, or non-accessible in the postnatal context. Only women who believed physical activity was necessary to meet their own needs, could contribute to others’ needs, and who experienced limited personal embodiment resistance (e.g., fatigue, birth recovery) pursued their desired activities. New mothers need support and education to engage in physical activity in the context of personal embodiment resistance. Research exploring relationships between activity choices and new mother’s need satisfaction could support the development of more effective healthcare strategies. v   Preface  While I received advice and support from my committee, I was responsible for conducting all parts of this research study, including: recruitment, obtaining consent, interviews, storing data, analysis, and writing-up. A transcriptionist transcribed 14 of the interviews. Ethical approval was obtained by the University of British Columbia Behavioural Research Ethics Board: ID number H14-01688. vi  Table of Contents Abstract	...........................................................................................................................................................................	ii	Lay	Summary	................................................................................................................................................................	iv	Preface	.............................................................................................................................................................................	v	Table	of	Contents	........................................................................................................................................................	vi	List	of	Tables	.............................................................................................................................................................	xiii	List	of	Figures	............................................................................................................................................................	xiv	List	of	Abbreviations	and	Conventions	..............................................................................................................	xv	Acknowledgements	.................................................................................................................................................	xvi	Dedication	................................................................................................................................................................	xviii	Chapter	1:	Introduction	.............................................................................................................................................	1	Background	....................................................................................................................................................................................................	1	Research	Problem	........................................................................................................................................................................................	2	Lack	of	clarity	around	effective	strategies	to	promote	physical	activity	..............................................................................................	2	Importance	of	seeking	women’s	perspectives	on	decision-making	within	the	postnatal	period	.....................................	5 Postnatal	physical	activity	literature	..............................................................................................................................................................	7	Purpose,	Definitions,	and	Significance	................................................................................................................................................	8	Summary	and	Thesis	Organization	......................................................................................................................................................	9	Chapter	2:	Literature	Review	................................................................................................................................	11	Introduction	................................................................................................................................................................................................	11	vii  Use	of	The	Literature	...............................................................................................................................................................................	11	General	Physical	Activity	Literature	and	Critique	......................................................................................................................	12	Ecological	physical	activity	research.	...........................................................................................................................................................	17	Intrapersonal	factors	..........................................................................................................................................................................................	18	Interpersonal	influences:	social	and	media	influences	.......................................................................................................................	19	Policy	and	environmental	influences	..........................................................................................................................................................	21	Description	of	Postnatal	Physical	Activity	Literature	...............................................................................................................	23	Postpartum	Physical	Activity	Research	Areas	and	Critique	..................................................................................................	23	Barrier/facilitator	research.	..................................................................................................................................................................................	23	Feminist	and	gendered	perspectives	................................................................................................................................................................	26	Physical	Activity	Literature	Summary	.............................................................................................................................................	29	Physical	Activity	and	Decision-Making	...........................................................................................................................................	29	The	influence	of	habit	..............................................................................................................................................................................................	30	Emotions	and	decision-making	...........................................................................................................................................................................	32	Eccles’	Expectancy-Value	Decision-Making	Framework	.........................................................................................................	33	Chapter	Summary	.....................................................................................................................................................................................	35	Chapter	3:	Theoretical	Perspectives	and	Methodology	...............................................................................	37	Symbolic	Interactionism	........................................................................................................................................................................	37	The	object	......................................................................................................................................................................................................................	38	Symbols,	the	act,	and	symbolic	interactionism	.............................................................................................................................................	39	Emergence	of	the	act	................................................................................................................................................................................................	40	Mind	.................................................................................................................................................................................................................................	41	Self.	...................................................................................................................................................................................................................................	42	Society	.............................................................................................................................................................................................................................	43	Application	of	symbolic	interactionism	to	the	study	...........................................................................................................................	44	Symbolic	Interactionism	Critique	......................................................................................................................................................	46	viii  Response	to	the	astructural	bias	critique	........................................................................................................................................................	49	Response	to	the	power	and	agency	critique	..................................................................................................................................................	50	Pragmatism	and	the	goal	of	symbolic	interactionist	work.	.....................................................................................................................	51	Introduction	to	Grounded	Theory	.....................................................................................................................................................	52	Grounded	theory’s	alignment	with	symbolic	interactionism	.................................................................................................................	55	Sensitizing	concepts	...........................................................................................................................................................................................	56	Navigating	a	Grounded	Theory	Methodological	Course	and	Epistemological/Ontological	Stance	......................	57	Divergence	in	grounded	theory	approaches	..................................................................................................................................................	57	The	shifting	location	of	grounded	theory	........................................................................................................................................................	59	My	grounded	theory	location	.........................................................................................................................................................................	61	Theoretical	and	methodological	considerations	for	my	proposed	approach.	...........................................................................	64	Additional	rigour	and	ethical	considerations	..........................................................................................................................................	65	Chapter	Summary	.....................................................................................................................................................................................	67	Chapter	4:	Methods	..................................................................................................................................................	69	Introduction	................................................................................................................................................................................................	69	Study	Design	and	Research	Questions	.............................................................................................................................................	69	Definition	of	terms	....................................................................................................................................................................................................	69	Ethical	Considerations	............................................................................................................................................................................	70	Sampling	.......................................................................................................................................................................................................	72	Inclusion	and	exclusion	criteria	..........................................................................................................................................................................	73	Approaches	to	sampling	..........................................................................................................................................................................................	73	Data	Collection	Sources	and	Strategies	...........................................................................................................................................	75	Participant	diaries	.....................................................................................................................................................................................................	75	Interviews	.....................................................................................................................................................................................................................	77	Field	notes	.....................................................................................................................................................................................................................	81	Theoretical	sensitivity	.......................................................................................................................................................................................	82	ix  Data	Analysis	............................................................................................................................................................................	86	Data	management	......................................................................................................................................................................................................	87	Coding	.............................................................................................................................................................................................................................	87	Open	coding	............................................................................................................................................................................................................	88	Development	of	main	concern	and	core	category	.................................................................................................................................	90	Memos	and	theoretical	integration	....................................................................................................................................................................	99	Rigour	........................................................................................................................................................................................	102	Reflexivity	and	Power	........................................................................................................................................................	105	Reflexivity	..................................................................................................................................................................................................................	105	Power	...........................................................................................................................................................................................................................	110	Reciprocity	.................................................................................................................................................................................................................	111	Summary	..................................................................................................................................................................................	112	Chapter	5:	Reconciling	Resistance	....................................................................................................................	113	Introduction	...........................................................................................................................................................................	113	Demographics	........................................................................................................................................................................	113	Main	Concern	.........................................................................................................................................................................	115	Reconciling	Resistance	......................................................................................................................................................	117	Gauging	.....................................................................................................................................................................................	124	Gauging	essentiality	...............................................................................................................................................................................................	124	Gauging	risk.	.............................................................................................................................................................................................................	130	Gauging	relational	risk.	...................................................................................................................................................................................	131	Risk	to	infant	needs	and	family	harmony:	Choices	around	independent	activity.	..........................................................	131	Risk	to	infant	needs	and	family	harmony:	Choices	around	time	of	day	and	scheduled	activities.	..........................	134	Gauging	personal	risk.	....................................................................................................................................................................................	136	Emotional	risk:	Choices	around	levels	of	emotional	investment.	..........................................................................................	137	x  Physical	risk:	Choices	around	levels	of	activity	intensity	and	duration.	.............................................................................	140	Group-based	risk:	Choices	around	physical	activity	milieu.	....................................................................................................	146	Gauging	accessibility	.............................................................................................................................................................................................	149	Choices	around	recreational	centres	and	fitness	programming	..................................................................................................	149	Choices	around	outdoors	activity	..............................................................................................................................................................	155	Gauging	Summary	................................................................................................................................................................	157	Engaging	...................................................................................................................................................................................	157	Pushing	through.	.....................................................................................................................................................................................................	157	Holding	back.	............................................................................................................................................................................................................	160	Holding	still	...............................................................................................................................................................................................................	163	Engaging	Summary	.............................................................................................................................................................	166	Adjusting	..................................................................................................................................................................................	166	Adjusting	for	women	who	were	pushing	through.	..................................................................................................................................	169	Adjusting	for	women	who	were	holding	back.	..........................................................................................................................................	172	Adjusting	for	women	who	were	holding	still	.............................................................................................................................................	175	Adjusting	Summary	.............................................................................................................................................................	176	Retaining	or	Shifting	Engagement	Strategies.	.........................................................................................................	177	Retaining	pushing	through.	................................................................................................................................................................................	177	Retaining	holding	back.	........................................................................................................................................................................................	178	Retaining	holding	still	...........................................................................................................................................................................................	182	Shifting	strategies	...................................................................................................................................................................................................	184	Outcomes	of	shifting	strategies	and	unreconciled	resistance	.......................................................................................................	186	Chapter	Summary	................................................................................................................................................................	189	Chapter	6:	Discussion	............................................................................................................................................	192	Introduction	...........................................................................................................................................................................	192	xi  Contribution	of	Symbolic	Interactionism	to	the	Theory	.....................................................................................	193	Comparing	and	Contrasting	the	Theory	with	the	Extant	Literature	.............................................................	194	Resistance	..................................................................................................................................................................................................................	196	Implications	of	intersecting	resistance	...................................................................................................................................................	200	Multiple	goal	pursuit	.............................................................................................................................................................................................	204	Gauging	........................................................................................................................................................................................................................	206	Physical	activity	centrality	resistance	......................................................................................................................................................	207	Personal	embodiment	resistance	...............................................................................................................................................................	211	Relational	resistance	.......................................................................................................................................................................................	216	Environmental	resistance	.............................................................................................................................................................................	219	Engaging	strategies	................................................................................................................................................................................................	222	Adjusting	and	shifting	strategies	.....................................................................................................................................................................	225														Summary	.................................................................................................................................................................................	228	Study	Implications	...............................................................................................................................................................	229	General	postnatal	physical	activity	provider	support	.......................................................................................................................	229	Centrality-specific	postnatal	physical	activity	considerations	.....................................................................................................	233	Education	and	health	service	implications	............................................................................................................................................	235	Research	implications	.....................................................................................................................................................................................	238	Limitations	.................................................................................................................................................................................................................	243	Conclusion	...............................................................................................................................................................................	244	References	.................................................................................................................................................................	248	Appendix	A:	Recruitment	Flyer	A	..................................................................................................................	282	Appendix	B:	Edited	Recruitment	Flyer	........................................................................................................	283	Appendix	C:	Consent	..........................................................................................................................................	284	Appendix	D:	Information	letter	.....................................................................................................................	286	Appendix	E:	Diary	Template	...........................................................................................................................	287	xii  Appendix	F:	Demographic	Questionnaire	.................................................................................................	288	Appendix	G:	Women’s	Summary	...................................................................................................................	289	Appendix	H:	Transcription	Confidentiality	Agreement	........................................................................	297	Appendix	I:	Initial	Interview	Guide	.............................................................................................................	299	Appendix	J:	Interview	Guide	Iterations	Second	and	Third	..................................................................	301	Appendix	K:	Initial	Theory	Diagram	...........................................................................................................	307	Appendix	L:	Sample	Field	Note	......................................................................................................................	308	Appendix	M:	Calibrating	Resistance	Memo	...............................................................................................	310	Appendix	N:	Early	Memo	for	Calibrating	Resistance,	Ground	Rules	................................................	317	Appendix	O:	Revised	Diagram	.......................................................................................................................	333	Appendix	P:	Example	of	an	Early	Memo	.....................................................................................................	334	Appendix	Q:	Example	of	Early	Diagram	.....................................................................................................	337	xiii  List of Tables Table 1: Characteristics as a Percentage of the Sample ...................................................................................... 114 xiv   List of Figures  Figure 1: Reconciling Resistance ........................................................................................................................ 118  Figure 2: Adjusting .............................................................................................................................................. 121  Figure 3: Shifting Strategies and Unreconciled Resistance ................................................................................. 123 xv  List of Abbreviations and Conventions    … A word or a few words within sentence omitted …. Omissions between two or more original sentences [e.g.] Addition of word or phrase to increase clarity to the reader  P Participant  I Interviewer   xvi  Acknowledgements   This study would not have been possible without the women who so graciously offered their time to participate. I thank all the women who were willing to carve space and invite me into their world to share their perspectives and stories and teach me about the complex intersecting personal, social, structural, and relational spaces that affect physical activity during the postnatal year. I thank the women for their feedback on theoretical ideas and their perspectives to help improve the fittingness of the theory with their experiences. This has been an eventful journey and I am indebted to countless experiences and people who have reframed my perspectives, lifted, pushed, pulled, and walked with me along the way. I extend my deepest gratitude that I was blessed to have such an extraordinary supervisor, Dr. Wendy Hall. Thank you Wendy for being five steps ahead of me and looking out for me in the fullest sense of this phrase by helping me not only with scholarly related opportunities but through being responsive to my emotional and learning needs. I thank you for steadfast commitment to quality and integrity that has been instilled in me to my core and your unwavering ability to model excellence in every realm of mentoring and supervision. I am thankful for all the seeds you’ve sown and some that I take delight in recognizing starting to grow and your influence; surely, I will have many more experiences than I have already where I recognize the impact you’ve had on my leadership, research, teaching, and mentorship capacity. Most importantly, thank you for your kind heart and caring not just about this process but for me as a person on this journey.  I extend a sincere thank you to committee members, Professors Tanya Berry and John Oliffe, for your time and commitment to reading such lengthy working copies and patience and flexibility with timing. Thank you to John for your ability to find the threads that needed to be pulled to condense and bring the dissertation up, particularly in Chapters 5 and 6. Thank you to Tanya for helping me to re-work my thinking about the resistance concept, re-situate my discussion, and refine my literature review. I thank you both for your insights and xvii  perspectives that have pushed, broadened, and enriched my thinking.  I am very grateful for funding throughout my program, particularly through the University of British Columbia and Canadian Institute of Health Research’s (CIHR) Transdisciplinary Understanding and Training on Research – Primary Health Care (TUTOR-PHC) and Better Days Better Nights Programs. I am also grateful to have been a part of the Canadian Obesity Network’s Summer School in 2015, which along the with the CIHR programs provided me with mentorship experiences, research connections, and support to build my research capacity.  I cannot do due diligence in the space provided to extend adequate thank yous to my family and friends who provided endless childcare hours and multi-dimensional support during this process, though every extension of support helped me take another step forward. Most importantly, I thank my beautiful husband Aaron who has been my foundation; thank you for your theoretical insights, deep-seated resilience, and for being brave enough to find gold in the trenches with me. And finally, I thank my children Ethan, Kalen, and Brayden for grounding  and teaching me that letting the balls drop is more important than keeping them up; Brayden you will never know the support you provided during our 2:46 PM after-school chats.     xviii                                                                                                                               For my grandfather 1  Chapter 1: Introduction  Background  The World Health Organization (2009) lists physical inactivity as the 4th highest risk factor for global mortality, and the primary cause of a quarter of breast and colon cancer (21-25%), diabetes (27%), and heart disease (30%) worldwide. Adults, including women within a year of childbirth (i.e., postnatal1), are encouraged to engage in 150 minutes of moderate-vigorous activity per week to support weight maintenance and reduce risk of obesity and complex chronic diseases (World Health Organization (WHO), 2010). Childbearing women represent a significant population in Canada; in 2012/2013, nearly 400,000 women gave birth (Statistics Canada, 2013). Recent Canadian estimates based on accelerometry (motion sensor physical activity measures) suggested only 14% of women in their childbearing years achieve sufficient weekly moderate-vigorous activity (Colley et al., 2011; Human Resources and Skills Development Canada, 2013). Childbearing can negatively affect physical activity levels and increase long-term obesity and related disease risk (Adamo, Langlois, Brett, & Colley, 2012; Larson-Meyer, 2002; McIntyre & Rhodes, 2009; Rooney, Shauberger, & Mathiason, 2005). A longitudinal study found childbearing women who did not return to pre-pregnant weight by 6 months were at increased risk for obesity and chronic disease 15 years following childbirth, although this relationship was moderated by physical activity at six months postpartum (Rooney et al., 2005). Bellows-Riecken and Rhodes’ (2008) systematic review suggested women with children are less physically active than women without children and men. Women with infants may be particularly at risk for physical inactivity; an American study based on accelerometry suggested only 3% of their postnatal sample were meeting physical activity recommendations (Evenson, Herring & Wen, 2012). In another multiethnic American postnatal sample, based on self-report, 65% were not meeting recommendations and 43% reported  1	Postnatal refers to women within 12 months of giving birth. 2  reducing their physical activity following childbirth (Albright, Maddock, & Nigg, 2006). People tend to overestimate their levels of physical activity when self-reporting because it is difficult to recall physical activity accurately; difficulty recalling suggests the possibility that higher proportions of women reduced their physical activity and were not meeting recommendations (Ainsworth, 2010). Research Problem  In North America and Europe, the postnatal period presents a unique opportunity for health care providers to support physical activity because women have frequent visits to public health nurses and their primary care providers in the first year after birth (Kinnunun et al., 2007; National Institutes of Health, 2008; Public Health Agency of Canada, 2013). Research suggests most women do not receive any physical activity advice from health care providers within the year following childbirth (Ferrari & Siega-Riz, 2010), which may be related to providers’ lack of knowledge and/or beliefs they do not have sufficient ability or time to counsel women about physical activity (Parker, Steyn, Levitt, & Lombard, 2011). There is a lack of research about women’s decision-making processes around physical activity in the postnatal period, whereby women assess and judge, in the context of other competing decisions, whether and how much they will engage in physical activity (Eccles & Wigfield, 2002; Reisberg, 2006). Knowledge about women’s postnatal physical activity decision-making processes can inform health care providers about the feasibility of existing interventions to promote physical activity. Our lack of understanding about how women make decisions about engaging in physical activity in the year following childbirth prevents effective endeavours to assist women with targeted interventions. Lack of clarity around effective strategies to promote physical activity.  There has been a modest amount of physical activity intervention research both with general and postpartum populations. In the general physical activity literature, researchers have critiqued the multifaceted nature of physical activity interventions (i.e., more than one strategy, such as education, counseling, and 3  exercise classes) because they limit understanding about which intervention components contribute to effectiveness (Conn, Hafdahl, & Mehr, 2011). In Conn and colleagues’ meta-analysis of 99,011 healthy adults, moderator analysis revealed behavioural-type interventions (e.g., goal-setting), face-to-face delivery models (e.g., counseling), and targeting individuals were more effective intervention strategies than cognitive strategies, mediated delivery (e.g., telephone), and community targets, respectively. Those findings are not definitive; because of the multifaceted nature of the interventions, the authors had to group intervention strategies generally to perform their meta-analysis (Conn et al., 2011). Other meta-analyses and narrative reviews are similarly vague; while individual face-to-face behavioural strategies may be effective, exactly what types of behavioural interventions are effective is still unclear (Davies, Spence, Vandelanotte, Caperchione, & Mummery, 2012; Dishman & Buckworth, 1996; Heath et al., 2012). Although, in general, the quantity of physical activity intervention studies is significantly greater in adult physical activity literature than in the postnatal literature, no existing studies have developed understandings about which intervention components support long-term increases in physical activity (Rhodes & Nigg, 2011). Types of interventions that best support women’s postnatal physical activity remain poorly understood, largely because foundational insights to women’s decision-making are lacking. A small number of studies have tested postnatal physical activity interventions for women but results have been mixed and researchers have generally reported low attendance (Hartman, Hosper, & Stronks, 2010). While authors of small pilot studies2 and studies sampling clinically depressed populations have reported that their postpartum physical activity  2	Here I refer to intervention studies tested exclusively in postnatal populations and those testing for changes in physical activity rates (frequency, intensity, and duration). There are other physical activity intervention studies conducted with mothers of young children < 5 and studies that do not include post-test physical activity levels I have not included here (Aittasalo, Pasanen, Fogelholm, Kinnunen, Ojala, & Luoto, 2008; Armstrong & Edwards, 2004; Clarke, Freeland-Graves, Klohe-Lehman, Milani, Nuss, & Laffrey, 2012; Fahrenwald, Atwood, Walker, Johnson, & Berg, 2004; Leermakers, Anglin, & Wing, 1998; McIntyre, Peacock, Miller, Koh, & Marshall, 2012; Miller, Trost, and Brown, 2002). 4  interventions of 2 to12 weeks duration have small-large effect sizes (Albright, Maddock, & Nigg, 2009, Albright et al., 2012; Cramp & Brawley, 2009; Da Costa et al., 2009; Daley, Winter, Grimmett, McGuinness, McManus, & MacArthur, 2008; Lewis, Martinson, Sherwood & Avery, 2011; McCrory, Nommsen-Rivers, Mole, Lonnerdal, & Dewey, 1999), larger studies testing interventions between 6 to12 months in length have found their interventions to have no effect on physical activity (Kinnunen et al., 2007; Ostbye et al., 2009; Taveras et al., 2011). Such results suggest short-term intervention effects are not sustained. Similarly, physical activity interventions in the general adult literature have been associated with positive short-term increases in physical activity; only limited studies have found a sustained increased in physical activity beyond a year post- intervention (Silva et al., 2011). The lack of positive findings resulting from interventions and low attendance may be explained by these studies failing to adequately take into account how the postpartum context affects women’s decision-making about physical activity. For example, Ostbye and colleagues (2009) used interventions commonly used in the general population (e.g., goal-setting, group-based classes, and printed self-help material), but those interventions did not result in a significant difference in rates of physical activity between control and intervention groups at 10 months postpartum. Women living in the postpartum period have reported different barriers to physical activity than women without children, suggesting that different intervention strategies may be needed to help women engage in physical activity in the postpartum period (Collins, Marshall, & Miller, 2007; Taveras & Plotkinoff, 2008). Andajani-Sutjahjo, Ball, Warren, Inglis, and Crawford (2004) surveyed 462 women aged 18-32 and reported lack of motivation (74%), lack of time (58%) and cost of exercise (51%) were women’s most commonly reported physical activity barriers. When Evenson, Aytur, and Borodulin (2009) surveyed women within a year of childbirth (N= 530) they found different barriers; rather than motivation and cost, women reported their most common barriers were lack of time (47%), lack of childcare (26%), and fatigue (12%). Other authors have reported these same exercise barriers, as well as lack of social support, as the most 5  common problems for women during the postpartum period (Groth & Davids, 2005; Miller & Brown, 2005). Women within a year of childbirth appear to have more concerns with fatigue, childcare, and support than women without children when trying to engage in physical activity. Importance of seeking women’s perspectives on decision-making within the postnatal period.  Notwithstanding differences in women’s perceived barriers during the postnatal year, most postnatal barrier research has been survey-based; thus, there is limited understanding about what reported barriers mean to women living in the postnatal context. Symons-Downs and Hausenblas (2004), in a study surveying women about postnatal exercise barriers, interpreted the barrier ‘lack of time’ as related to poor time management skills and suggested time management counseling as a way to increase physical activity. However, Tavares and Plotkinoff (2008) suggested, from their Canadian qualitative study of full-time employed women, that women with children lacked time because they felt pressured to uphold gendered social norms around motherhood and employment by being present at all their children’s extracurricular activities and working unpaid hours. Varying interpretations about reasons for women’s lack of time emphasize the importance of seeking the perspectives of women who are living within the postnatal context about their decision-making to help account for these meanings; we cannot develop effective interventions without understanding women’s underlying reasons for engaging or not engaging in physical activity during the postpartum period. Furthermore, knowledge about women’s perspectives of physical activity decision-making during the postnatal period might augment existing physical activity literature that examines correlates, moderators, and mediators of physical activity behaviour by contextualizing how women understand these factors as interrelated and important for their decision-making (Rhodes & Dickau, 2013; Schuz, Ziegelmann, Wolff, Warner, Schwarzer, & Romer, 2012). Ecological physical activity research has illuminated the complexity of factors influencing physical activity decision-making by demonstrating significant relationships between personal (e.g., demographic), psychosocial (e.g., cognitions), policy, and environmental factors and behaviour patterns 6  (Burton et al., 2005; Sallis et al., 2006). However, moderation and mediation studies that have explored interrelationships between these factors and physical activity behaviour have produced mixed findings that have not provided clear understanding about how these factors interrelate in decision-making (e.g., Rhodes & Dickau, 2013; Rhodes & Pfaeffli, 2010; Schuz et al., 2012). In the postpartum context, perceptions of more control over time, available social support, and access to childcare, and perceptions of decreased fatigue seem important in women’s physical activity decision-making because they have been identified as physical activity enablers and associated with continued physical activity in parenthood (McIntyre & Rhodes, 2009). Studies seeking women’s perspectives about associations among such factors and their relationships with broader influences, such as environment, sociocultural norms, and life experiences with physical activity, will support understandings about the postnatal context influences on physical activity decision-making and identification of ways to support women’s physical activity during the postpartum period. The symbolic interactionist view of human behaviour as variable also points to the importance of developing understandings about women’s physical activity decision-making processes that precede behaviour (Blumer, 1969). Behaviour can be considered an end product of a decision (Evans, 2008), with physical activity behaviour reflecting decisions around physical activity. Physical activity literature has consistently reported that behaviour is difficult to predict; for example, a recent meta-analysis suggested physical activity intention (considered a strong predictor of behaviour) only related to physical activity approximately 50% of the time (Rhodes & de Bruijn, 2013). That finding suggests that it might be useful to explore how people make decisions around physical activity and why they believe their intentions are not always enacted. Decision-making theories, such as Eccles’ expectancy-value theory (1991), have suggested the importance of considering peoples’ expectancies and values beyond those around physical activity (e.g., around motherhood), while taking into account the influence of personal, social, and societal contextual factors, as 7  specified in ecological physical activity models on decision-making. Exploring women’s perspectives about physical activity decision-making in the context of other values and expectations might help illuminate the complexity of physical activity decision-making so that health care professionals can consider tailored ways to support women’s physical activity engagement during the postpartum period. Postnatal physical activity literature.  Currently, research about postnatal physical activity decision-making is limited and may not reflect the complexity of the postnatal population. Feminist work seeking women’s perspectives around physical activity decision-making during the postpartum period has been limited to examining mothers with young children (< 5 years) and presented a narrow view because it largely linked decision-making around physical activity to gender socialization and gendered social norms around motherhood (e.g., Thompsson, 1999). Such a narrow approach is inconsistent with ecological models suggesting consideration of a breadth of factors relevant in decision-making (Dworkin & Wachs, 2009; Sallis et al., 2006). Moreover, women living in the postnatal period may have more diverse processes of decision- making around physical activity than women with older children (i.e., < 5 years old) because they make decisions in a different context; women during the first year post birth are less likely to be working (i.e., on maternity leave) and more likely to be coping with infant sleep and breastfeeding concerns (Hunter, Rychnovsky, & Yount, 2009; Sadeh, Luedtke &Wiegand, 2009). Physical activity feminist literature has approached motherhood from the standpoint of women’s inactivity (e.g., Miller & Brown, 2005). Lewis and Ridge (2005) suggested, from their qualitative study of 40 Australian mothers, that motherhood creates opportunities for reconstructing meanings around physical activity. The postnatal period, despite its challenges, may be an important time for women to change or regain past physical activity patterns but it is important to understand how women make physical activity decisions in order to consider how to promote physical activity in this population. 8  Purpose, Definitions, and Significance  The postpartum period is defined as the first 12 months after birth; postnatal and postpartum are synonymous terms. The purpose of this research was to explore women’s decision-making processes around physical activity within the postnatal period and generate a grounded theory to explain their decision-making. The theoretical perspective guiding the study is symbolic interactionism, which posits that people constantly interpret and act on the basis of meanings objects have for them (Blumer, 1969; Mead, 1934). Grounded theory fits this perspective because it explains conditions that create variability in decision-making and supports understanding of decisional processes (Charmaz, 2006). The primary research question was: What are women’s decision-making processes about physical activity in the postnatal period? Physical activity was defined as any bodily movement involving energy expenditure above resting, which includes any activity listed on the Ainsworth compendium of activities, and leisure time physical activity (Ainsworth et al., 2011; WHO, 2010).3 Processes are defined as lines or ways of thinking, in this case, about the engagement in physical activity (Reisberg, 2006). Decision-making is defined as involving a subjective assessment of the available options for behaviour and judgment about an appropriate line of behaviour (Reisberg, 2006). Decision-making is not a wholly reasoned and controlled process, whereby people consciously weigh options about the costs and benefits of engaging in particular behaviours, because automatic thoughts (e.g., habitual ways of thinking about physical activity) and emotions also influence the decisions people make (Evans, 2008; Reisberg, 2006). Because behaviour is ongoing, decision-making is dynamic (Evans, 2008; Reisberg, 2006). Daily decision-making involves the intersection of conscious (e.g., short-term conscious planning, within the day, within the half hour) and automatic thought processes (e.g., emotions) (Evans, 2008; Wood & Neal, 2007). I am interested in  3Resting is defined in terms of 1 metabolic unit or 1 unit of expenditure of energy (Warbuton, 2010). 9  conscious decision-making processes in the postnatal period from daily to longer term (over the postnatal year), but not moment-to-moment decision-making (e.g., spontaneous, deciding to increase the pace of a run) because women might not access the cognitive processes influencing spontaneous decision-making (Evans, 2008). I intended to explore the decision-making factors women perceive and can consciously reflect on as influencing their decision-making. The study was intended to increase understanding about how women comprehend complex personal, social, policy, and environmental factors as interrelated in their decision-making. It explores how women make decisions about physical activity during the postpartum period in the context of other competing demands. By examining the meanings women associate with personal, social, policy, and environmental factors and emotions and habits in physical activity decision-making and their relationships, this study potentially contributed to nuanced understandings about how women make decisions around physical activity during the postpartum period that may inform the development of tailored interventions that account for women’s perspectives. Summary and Thesis Organization  In this chapter, I have provided background information, described the research problem and its significance, the research question, and study purpose. I have highlighted potential study contributions towards understanding postnatal physical activity decision-making. In Chapter 2, I critically review the general and postnatal physical activity literature. In Chapter 3, I provide an overview of symbolic interactionism and grounded theory methodology. In Chapter 4, I describe the methods used to conduct the study, including: ethical considerations, recruitment, data collection, theoretical sampling, analysis, and rigour. Chapter 5 provides a description of the sample, an overview of the core category, and detailed descriptions of the related categories that explain how women make physical activity decisions during the postpartum period. In Chapter 6, I discuss the findings in relation to the existing literature and implications of this research for health care 10  providers and the health care system. I conclude by describing the implications arising from the study findings. 11  Chapter 2: Literature Review   Introduction  In this chapter, I present a critical analysis and synthesis of general and postnatal physical activity literature. I begin the chapter by interrogating the use of a literature review in the context of a grounded theory study. Use of the Literature.  The purpose of this study is to generate a grounded theory; however, there is debate around the use of literature with this approach. Glaser (1978), one of the founders of grounded theory, argued against conducting a literature review prior to data collection because he believed that reading literature could introduce preconceived ideas that would bias the researcher’s theory development. Other authors have argued that it is ethically imperative to situate the research problem in the context of current literature and provide justification for the study through the use of a literature review (McGhee, Marland, & Atkinson, 2007; Walls, Parahoo, & Fleming, 2010). Moreover, they argued that it is unlikely a researcher would be unfamiliar with the literature in his/her field. I take the perspective that a broad literature review is necessary to position the contribution of my research and enhance my theoretical sensitivity, which enriches the researcher’s ability to handle the data theoretically (Glaser & Holton, 2004). Given the narrow base of the postnatal physical activity decision-making literature, this review incorporates general physical activity literature, as well as decision-making research. I used a range of databases: PsycINFO (1880-present), Web of Science (1900-present), Pub Med (1946-present), Sport Discus (1949-present), and Sociological Abstracts (1952-present) and searched, with no search restrictions, using combinations of keyword and/or MeSH heading terms including: decision-making, judgment, grounded theory, postpartum, postnatal, physical activity, exercise, mother*, barrier*, and behavio*r. 12  The following review is organized around the general and postnatal physical activity literature. I discuss the limitations of general physical activity literature and critique postnatal physical activity literature around two main areas of inquiry: barrier/facilitation studies and feminist/gendered research. My discussion of decision-making asserts the utility of exploring emotions and goals around physical activity decisions. I conclude by applying my critical analysis of the literature towards a potentially useful approach to develop knowledge about women’s postnatal physical activity decision-making that incorporates ecological physical activity and expectancy-value perspectives. General Physical Activity Literature and Critique  The general physical activity literature has had a strong emphasis on using social cognition and behavioural theories to explain patterns of physical activity and hypothesize/test strategies to influence change in physical activity behaviour, based on findings that a wide range of cognitions4 (e.g., self-efficacy, outcome expectations) are associated with physical activity behaviour (Bauman et al., 2012; Heinrich, Maddock, & Bauman, 2011; Mack, Sabiston, McDonough, Wilson, & Paskevich, 2011; Pan et al., 2009; Plotkinoff, Costigan, Karunamuni, & Lubans, 2013; Rhodes & Nigg, 2011). Beyond cognitions, many demographic and environmental factors have been associated with physical activity (e.g., Bauman et al., 2012). For example, Bauman and colleagues (2012) comprehensively reviewed physical activity literature around correlates of physical activity and found that a breadth of variables across individual and social domains have been associated with physical activity behaviour. Specifically, increased age, poorer health status, and increased body weight correlated with decreased physical activity, while heightened self-efficacy, income, perception of social support, and behavioural intention related to increased physical activity. In the Canadian context, these findings were largely supported by Pan and colleagues (2009). They sampled 2,854 Canadian women and found that  4	Cognition is the overarching term for mental activities involving how people process knowledge (acquire, store, transform, use); cognitions are a product of cognitive processing (Matlin, 2002). 13  increased education, income, behavioural intention, and self-efficacy, and lower perceived barrier scores predicted higher physical activity levels. Additionally, theory-based intervention studies aimed at enhancing the level of cognitions that support physical activity motivation (e.g., self-efficacy) had varying degrees of success at increasing physical activity levels in participants (e.g., Cramp & Brawley, 2006; Hardeman, Kinmouth, Michie, & Sutton, 2011). Unfortunately, the success of socio-cognitive and correlational research to explain factors important for physical activity decision-making and behaviour has been limited; correlations between predictors and physical activity are low to moderate on average and the amount of physical activity variance accounted for by socio- cognitive constructs is approximated as fluctuating between 30 and 50% (Plotkinoff et al., 2013; Rhodes & Nigg, 2011). Recent research suggests that only 50% of people who indicate behavioural intention to exercise actually follow through with their intention (de Bruijn & Rhodes, 2013). There are strategies (i.e., creation of implementation intentions or planning around when and where the activity will be performed) that have improved the strength of the relationship between intention and behaviour (Gollwitzer & Sheeran, 2006); however, Belange-Gravel, Godin and Amireault’s (2013) meta-analysis suggested that the effect of these strategies is limited in the domain of physical activity. While socio-cognitive theories and demographic predictors have helped develop theoretical understandings about cognitions and demographic characteristics relevant to physical activity decision-making there remains a gap in explaining decision-making processes that lead to physical inactivity or activity. Mielewczyk and Willig (2007) identified a limitation of socio-cognitive research as the underlying assumption that health behaviours exist as a distinct entity when, in reality, health behaviours are part of a web of other behaviours that are enacted differently across contexts. For example, cognitive constructs, such as self- efficacy, may not consistently predict behaviour because self-efficacy is not stable in different contexts. A 14  person quitting smoking may be able to abstain when by his or herself, but at a social event (where smoking might be associated with socializing), self-efficacy to abstain from smoking decreases (Mielewczyk & Willig, 2007). Likewise, an individual’s level of physical activity behavioural intention may vary; in a research setting, a person may have high level of intention to exercise, but across other situations, intentions may decrease. Such differences suggest that research is needed to explore peoples’ differing behaviours and how they might affect decisions around physical activity. As an extension to this thinking, socio-cognitive health behaviour research has been critiqued as reducing the complexity of health behaviour decision-making by only examining relationships between variables and the behaviour of interest (Mielewycyk & Willig, 2007). Decision-making around physical activity occurs in the context of other behavioural decisions (Mielewycyk & Willig, 2007) but there remains limited understanding about how other behaviours affect cognitions and intersect with physical activity decision-making. For example, one finding of a systematic review of physical activity intention-behaviour relationship moderators was that behavioural intention stability increased the relationship strength between intention and behaviour and that intention instability largely accounted for the inability of people to translate intentions into behaviour (Rhodes & Dickau, 2013). While this interpretation explains intention stability’s influence on physical activity decision- making it does not explain the other competing daily demands (e.g., coping with fatigue) that affect intention stability. To understand decision-making around physical activity it is important to consider interrelationships between cognitions across a range of behaviours that women use. Authors have critiqued the socio-cognitive health behaviour literature for not adequately explaining why particular variables are related to health behaviour given the contexts of peoples’ lives (Frolich & Poland, 2007; Mielewycyk & Willig, 2007; Shim, 2002). In the context of physical activity socio-cognitive models explain that variables are related to or mediate physical activity but they do not sufficiently explain how living in 15  particular social contexts and circumstances (e.g., having a high or low income) influence physical activity decisions and perceptions about physical activity (Frolich & Poland, 2007). For example, Schuz and colleagues (2012) found that the socioeconomic status of participants’ neighbourhoods moderated the relationship between health attitudes and physical activity. They interpreted their findings in relation to social cognitive theory by suggesting that increased physical activity opportunities and role models in more affluent neighbourhoods positively influenced peoples’ ability to translate physical activity attitudes into behaviour. While their interpretation offers some suggestions about facilitators that might have positively affected physical activity decision-making it provides a limited account about how the social context of affluence might also influence the breadth of general cognitions (e.g., values and goals), which intersect in physical activity decision-making (i.e., in how valuable physical activity is perceived to be given that social context) (Eccles & Wigfield, 2002; Mielewycyk & Willig, 2007). Some authors have argued that measurement error across various theoretical constructs and physical activity outcomes partially accounted for limitations in explaining physical activity, although, improvement in measurement might only partially aid in explaining decision-making processes (Brug et al., 2005; Collins et al., 2007; Rhodes & Nigg, 2011). For example, behavioural intention is usually measured by asking participants 1-3 questions about their physical activity intention (e.g., intention in the next month, or next 7 days) (Ajzen & Cote, 2008). Rhodes and Horne (2013) argued that this form of measurement does not take into account how other competing goals and intentions might influence physical activity and suggested replacing the measurement of intention with behavioural resolve. Behavioural resolve measures whether participants would: make exercise a priority; exercise even if they had other demands on their time; and exercise even if they were tired (Rhodes & Horne, 2013). Nevertheless, while measuring behavioural resolve might enhance the explanation of physical activity behaviour patterns, it does not explain what other priorities are, what demands 16  time, or what makes individuals tired. Thus, improvements in the measurement of socio-cognitive constructs might not extend explanations of the processes of physical activity decision-making because the measurements do not adequately explain the contexts in which people make decisions. Mielewczyk and Willig (2007) suggested that research is needed to explore contexts of decision-making and, more importantly, what behaviours mean to people so that we can develop more effective health behaviour interventions. Following their line of reasoning it is important to extend explorations beyond physical activity cognitions and demographics, as measured by self-report surveys, and their relationships to behaviour. Researchers need to learn about how people understand their cognitions, and how they operate in different contexts and in the context of other decisions. By understanding meanings and context, researchers can consider how cognitions vary in different contexts and how women understand physical activity in relation to other decisions. Furthermore, researchers can begin to identify ‘when’ and in what situations beliefs relate to behaviour. Only then can they develop effective interventions that work with peoples’ beliefs. Using qualitative research approaches can augment existing physical activity literature by developing understanding about contexts of decision-making and personal meanings around physical activity.  For example, exploring with women how the context of the postpartum period affects their goals and values and how those goals/values intersect with physical activity decision-making can illuminate how inactive women explain their inactivity. Unfortunately, qualitative research has been limited by being undertaken with particular groups in the context of physical activity decision-making. Weed (2010) and others (Holt & Tamminen, 2010; 2011) have systematically reviewed major exercise science and psychology journals. Most grounded theories from those reviews were about elite athletes and their sports (N = 11); none addressed women’s physical activity decision-making during the postnatal period. 17  Ecological physical activity research.  There has been a shift toward consideration of ecological theoretical models in understanding engagement in physical activity. Ecologically-guided research has advanced findings from socio-cognitive research (Sallis et al., 2006) because it has explicitly recognized effects of contextual influences on physical activity decision-making, beyond personal factors (i.e., demographic characteristics and physical activity constructs), that include interpersonal, environmental, and policy contexts. The ecological approach has guided physical activity researchers to explore these influences and their relationships to physical activity (Sallis, Owen, & Fisher, 2008). For example, ecologically-guided physical activity research would explore relationships between built environment (e.g., the presence of walkways) and policies around land-use and physical activity decision-making (Sallis et al., 2008). Key to socio-ecological understanding is the presence of interrelationships between all factors that influence women’s decisions about physical activity behaviour (Sallis et al., 2008). Despite showing promise in considering broader contextual influences ecological research has usually been based on developing understanding about relationships between multilevel factors and physical activity, and not on personal meanings and decision-making processes around physical activity (Frolich & Poland, 2007; Shim, 2002). On the other hand, ecological research can provide a useful framework for researchers to consider personal, social, policy, and environmental factors in decision-making. In the following section, I provide brief definitions of personal, social, policy, and environmental influences using Sallis and colleagues’ (2006, 2008) descriptions, explain how research has explored these areas in relation to physical activity, and speculate about effects of these influences on women’s decisions about physical activity within the postpartum period. The purpose of this section is to explain my thinking about the types of factors that might be relevant for postnatal physical activity decision- making and to highlight a range of possible influences on physical activity decision-making rather than to provide an exhaustive explanation of all research in these areas. 18  Intrapersonal factors.  Sallis and colleagues (2006) collectively referred to demographic, biological, and socio-cognitive factors as intrapersonal influences (herein referred to as personal influences) on physical activity. The influence of personal influences on physical activity behaviour has received the most attention in general physical activity literature (e.g., psychology and sport science research). It has generally been accepted that individuals’ previous accomplishments in sports/physical activity predicted more favourable cognitions to physical activity engagement (e.g., increased self-efficacy and behavioural intentions) (Bandura, 1997; Mack et al., 2011). From a demographic standpoint increased income has been positively associated with physical activity, and age has been inversely associated with physical activity, while increased numbers of children and lower ages of children have predicted lower levels of physical activity (Adamo et al., 2012; Bauman et al., 2012; Pereira et al., 2007; Withall, Jago, & Fox, 2011). From a biological and psychological stance women’s previous experiences with physical activity might also affect how they make decisions about physical activity during the postnatal period. A life course perspective has increasingly been used to guide health and sport science research (Braveman & Barclay, 2009; Dixon, Warner, & Bruening, 2008; Hertzman & Power, 2006). Traditionally, life course perspectives have been concerned with biological exposures to risk (e.g., low birth weight) and their relationships to health but these perspectives have been expanded to incorporate the intersection of social context with the development of cognitions, health, and behaviour across the life span (Braveman & Barclay, 2009; Giele & Elder, 1998). Life course perspectives have suggested that experiences have a cumulative effect on peoples’ beliefs and decision-making around health behaviours. For example, in an Australian qualitative study, a life course perspective expanded understanding about why low-income women in their sample had low levels of physical activity (Ball, Salmon, Giles-Cortia, & Crawford, 2006). Woman with moderate to high incomes typically 19  recalled physical activity as enjoyable and related to team sports as children but women with lower incomes described physical activity as being more negative, having usually been associated with walking for transportation purposes. Different meanings associated with physical activity for low and high-income women related to both current and previous experiences. Although women’s current contexts of having low, moderate, and high income can affect decisions about physical activity (Frisby & Millar, 2002) negative early associations with physical activity might also be influential. It might be important when considering postnatal physical activity decision-making to examine not only women’s current context but also their past experiences with physical activity. Interpersonal influences: social and media influences.  Sallis and colleagues (2006) described interpersonal influences on physical activity decision-making behaviour as situated in the sociocultural environment. In brief, interpersonal influences on decision-making about physical activity are those occurring in relation to interactions among people, and so have included, but were not limited to, the influence of culture, peer networks, family structure and support, and social norms (Sallis et al., 2006). Although Sallis and colleagues positioned media in the ‘environment’ portion of their model media could also be considered an interpersonal influence because social interaction has influenced how people interpret media sources (e.g., magazines, newspapers, social media, television) (Dworkin & Wachs, 2009). Peoples’ beliefs and understandings around physical activity are influenced by socialization processes, which are lifelong processes of knowledge, value, and goal acquisition occurring through interactions with other people and groups (Stockard, 1999). Socialization processes have influenced how people think, act, and interpret their own and others’ actions because people have learned socially acceptable ways of behaving and thinking through interactions. Role socialization has described how people come to understand their roles (e.g., 20  mother, friend) through interactions with others; it is the process by which people can acquire similar beliefs as others in these roles concerning behaviours and role expectations (Stockard, 1999). Gender socialization processes might have influenced women’s perceptions about physical activity. From childhood, women might have had stronger inclinations to view physical activity as more negative and related to weight and appearance than men. Vertinsky (1998) argued that girls have been socialized to understand appearance and attractiveness as highly important and have learned to associate undertaking physical activity with increasing attractiveness and reducing weight. Spencer, Rehman, and Kirk’s (2015) scoping review about gender norms and physical activity in adolescent girls found that girls had complex relationships with physical activity because they enjoyed physical activity but also experienced pressure to be perceived as both feminine and athletic in their physical activity engagement. Consideration of women’s perceptions about how they have been socialized about physical activity could enhance understandings about how women make decisions around physical activity in the postnatal period. Media is how information is communicated (e.g., newspapers, computer, magazines, and advertisements) to a wide range of people; media might also have influenced women’s beliefs about physical activity and affected their decision-making (McGannon & Spence, 2012; Pankratow, Berry, & McHugh, 2013). Because the body of literature about the influence of media on beliefs is extensive I limit my discussion to how media has been related to goals around physical activity for women. Media may have reinforced the main goals of physical activity to weight loss and achieving an ideal feminine body, rather than achieving health and well-being. Dworkin and Wachs (2009) argued that, in general, advertising has positioned the ideal female body as one that is toned and slim. Likewise, magazines or images of women in the postpartum period have tended to emphasize the need to return to pre-pregnancy body weight (Dworkin & Wachs, 2009; Fox, Hefferson, & Nicolson, 2009). Moreover, Dworkin and Wachs argued that pregnancy magazines position the return to pre-pregnant body as a responsibility of women to achieve 21  feminine ideals of a controlled body. Because exercising to reach health goals may be more intrinsically motivating than exercising to achieve weight loss or an ideal feminine body, the tendency for media to have emphasized women’s weight loss in terms of attractiveness over health might have negatively influenced women’s decisions about physical activity within the postnatal period if they believed such goals were not within their reach (Segar, Eccles, & Richardson, 2011). Women’s perceptions of media messages might illuminate how women develop goals and expectations around physical activity in the postnatal period. Policy and environmental influences.  In the ‘policy environment’ Sallis and colleagues (2008) included policies that have influenced recreation and transit service delivery, the accessibility of recreation services, school, and health care, and land development. They used policy to explain influences on physical activity because they argued that policy structures the delivery, quantity, and type of recreation programs, and number and structure of indoor and outdoor facilities (Sallis et al., 2008). Concerning health care and the school systems they argued that policy has influenced programs related to physical activity and education provided to health care providers and educators (Sallis et al., 2008). They situated policy as influencing where recreation facilities are located, when programs are delivered, the cost of programming, and the existence of transit and road systems to access programs or facilities (Sallis et al., 2008). All of the factors they have attributed to emanating from policy decisions could influence women’s decisions about engaging in physical activity. Cureton and Frisby (2011) conducted a Canadian study with 20 recreational staff from diverse recreation facilities exploring leisure access policy implementation. The leisure access policy ensured that citizens below the Canadian low-income cut-off line could access a facility at reduced cost or free. Cureton and Frisby found that residents had to provide a number of documents to “prove poverty,” which staff reported 22  could be a deterrent to their decisions to attend (p. 12). Revenue-generating recreation programs were highly advertised while free programs were not because it was assumed participants would find out about them from their social workers. In these examples, the presence of a leisure access policy that complicated physical activity access or absence of policies supporting accessible transport to leisure activities (e.g., a shuttle that picked women up) influenced women’s perceptions of barriers to physical activity. Exploring women’s experiences with recreation services and leisure programming might help to explain how policies around recreation and leisure enter into physical activity decision-making. Policy and environmental effects on physical activity have been linked because policy has influenced how land is used. Sallis and colleagues (2006) generally characterized environmental factors that could affect physical activity as the presence of walkways, traffic volume, aesthetics, parks, trails, facilities, neighbourhood safety (i.e., the presence of crime), transit, transportation systems, weather, air quality, and topography. Research has suggested that the built environment (i.e., increased presence of aesthetically pleasing facilities and trails compared to traffic congestion and roadways) might influence individuals’ perceptions of facilitators to exercise while increased perceptions of crime and poor weather have been associated with barriers to physical activity (e.g., Wolff & Fitzhugh, 2011). Although environmental factors are intuitively attractive to explain women’s physical activity decision-making during the postnatal period Evenson and colleagues (2009) found that only 25 of 530 American women, surveyed at 1 year post birth, reported environmental factors (i.e., poor weather, lack of facilities) as barriers in their decisions to engage in physical activity; that finding has been supported by general physical activity literature (Duncan, Spence, & Mummery, 2005; Pan et al., 2009). In contrast, in the preceding study, 426 of 530 women reported personal or social factors as barriers in their decisions to exercise (Evenson et al., 2009). In addition to environmental influences on physical activity decision-making it may be salient to examine interpersonal and intrapersonal factors. 23  Description of Postnatal Physical Activity Literature  The postnatal physical activity literature has centred on how interpersonal and intrapersonal factors affect physical activity behaviour. Most postnatal physical activity research is intervention or survey-based and has tested the effects of a physical activity intervention or explored barriers to and facilitators of postnatal physical activity (e.g., Evenson et al., 2009). I identified a small body of qualitative articles that explored mothers’ beliefs around physical activity. None of those studies exclusively sampled women within a year of childbirth (Collins et al., 2007; Hamilton & White, 2010a; Lewis & Ridge, 2005; McGannon & Schinke, 2013; Miller & Brown, 2005; Tavares & Oken, 2008; Thomsson, 1999). Postpartum Physical Activity Research Areas and Critique  The postpartum physical activity literature has explored postnatal physical activity barriers/facilitators and women’s beliefs about postnatal physical activity behaviour. Collectively the literature provides some insight about how women with young children view physical activity and their beliefs about physical activity barriers/facilitators but it does not describe processes involved in physical activity decision-making. This is particularly the case for women who have been physically active during the postpartum period. Barrier/facilitator research.  Physical activity barrier/facilitator research that has informed postnatal physical activity interventions has typically consisted of cross-sectional survey-based studies. In a large postnatal barrier/facilitator survey Evenson and colleagues (2009) recruited 530 American women who were < 20 weeks gestation and measured their beliefs about physical activity barriers and facilitators via surveys at 3 and 12 months postpartum. The women’s most commonly reported barriers to exercise at 3 and 12 months postpartum were lack of time (47% and 51%, respectively), issues with childcare (26% and 22%), and fatigue (12% and 13%), while their most commonly reported facilitators were partner support (16% and 10%), desire to feel better (14% and 9%), and 24  social support (11% and 11%). Lack of time and childcare, and fatigue have also operated as major barriers, and social support has been an important facilitator of physical activity in reports of smaller studies of postpartum barriers/facilitators (Doran & Davis, 2011; Groth & David 2008; Symons-Downs & Hausenblas, 2004). Although prior research has been helpful to understand women’s beliefs about barriers to physical activity it has not explained complex decision-making processes that have influencing their physical activity behaviours. Studies such as Evenson et al.’s (2009) and others (Doran & Davis, 2011; Groth & David, 2008; Symonds-Downs & Hausenblas, 2004) have provided descriptions of barriers but have not examined the relationships between barriers and facilitators in women’s decision-making. Hamilton and White (2010a) identified 5 major themes and 30 subthemes from their qualitative study of Australian parents’ perceptions about physical activity. They found that parents described a range of decisional factors that motivated (e.g., to model behaviour for their children, support a happy upbringing for their children) or demotivated (e.g., perceptions they could make up for lost activity later, guilt when exercising, and the need to attend to domestic duties) their physical activity engagement. Because the themes were not integrated it was difficult to understand how the different decisional factors were related in parents’ processes of physical activity decision-making. Qualitative strategies that facilitate integration of themes, such as grounded theory, can be used to account for multiple factors affecting physical activity decision-making (Charmaz, 2006; Glaser, 1978). What barriers/facilitators mean to women, in the context of their lives, is not fully understood. The barrier labelled ‘fatigue’ (e.g., Evenson et al., 2009) could carry diverse meanings for different women. Fatigue might have represented emotional fatigue, physical fatigue, and perceived fatigue from efforts to organize and plan physical activity, or a combination of these meanings; those meanings may differentially affect women’s decision-making processes around physical activity engagement. 25  Meanings of other barriers/facilitators have received more research attention in parents with older children but further research in the postpartum period that considers interrelationships between these barriers/facilitators would be valuable. Hamilton and White (2010b) explored the meanings that Australian parents of children aged 8 months-21years associated with social support for physical activity. Parents described types of support across five domains (instrumental (i.e., tangible support, watching children while parents exercised), informational, emotional, reciprocal, and autonomy support) that helped facilitate physical activity. While Hamilton and White highlighted the complexity of social support in relation to supporting physical activity they did not explain how other barriers/facilitators intersected with parents’ decision-making around physical activity or how perceptions of social support beyond exercise (i.e., general perceptions of their relationships with others) influenced parents’ physical activity decisions. It is important to explore how women understand physical activity and related barriers/facilitators in order to explain influences on their decisions around physical activity. Reported barriers may reflect women’s views about physical activity and the value they place on it rather than an actual lack of time, fatigue, or unavailable childcare. Watson, Milat, Thomas, and Currie (2005) tested an exercise intervention where mothers walked in a group setting with their infants (childcare was not a barrier) but they reported a 20% attendance rate. Studies about leisure-time activity for women within a year of childbirth suggested that women spend on average 2 hours per day on social media websites and text messaging and that this social interaction improves maternal mood (Taveras et al., 2010). Other studies have suggested that women within a year of childbirth watch approximately 2 hours of television per day (Oken, Taveras, Popoola, Rich-Edwards, & Gillman, 2007). Mothers may have viewed such activities as important to improve their moods or to maintain social relationships with friends/partners but may not have attributed such benefits to physical activity. 26  Feminist and gendered perspectives.  A small body of qualitative research, guided mostly by feminist perspectives and conducted primarily in Australia, has examined the contexts in which women with young children make physical activity decisions (Appleby & Fisher, 2009; Choi, Henshaw, Baker, & Tree, 2005; Hamilton & White, 2010a; Hamilton & White, 2010b; Lewis & Ridge, 2005; McGannon & Schinke, 2013; Miller & Brown, 2005; Thomsson, 1999). This literature has collectively linked women’s inactivity to gender socialization processes. A number of authors have argued that the ‘ethic of care’ and ‘supermom’ discourses5 have helped to explain why women have found it challenging to decide to undertake physical activity when they have children. Because women feel morally compelled to care for others at the expense of themselves (i.e., adopt an ethic of care) women have appeared to feel a lack of entitlement to seek leisure pursuits, such as physical activity (Choi et al., 2005; Lewis & Ridge, 2005; McGannon & Schinke, 2013; Miller & Brown, 2005). From a feminist perspective, the lack of entitlement to leisure has arisen from the differences in social norms around domestic responsibilities and leisure between men and women (Thomsson, 1999). For example, McGannon and Schinke (2013) conducted a discourse analysis of a Canadian couple’s interactions around physical activity; the partner explained that his wife needed to exercise during her work lunch hour to keep time for child care and domestic responsibilities (e.g., laundry) because he was not responsible for helping with childcare or domestic responsibilities so that his wife could exercise outside of work. Miller and Brown (2005) interviewed 12 women with young children; some women viewed finding time for physical activity as an additional responsibility, after their childcare and domestic responsibilities. The women described little reprieve because their   5 Discourses are ways of thinking that people take up either verbally, in media or print used in everyday social life (Malpas & Wake, 2006). Discourses represent the available and acceptable language people have given their particular context and environment and shape what people can say as well as the knowledge they can generate (Malpas & Wake, 2006). 27  husbands/partners were often at work or at leisure (e.g., socializing). For those women, a “double standard” existed, whereby their partners’ leisure time was justified but theirs was not (p. 413). The influence of the ethic of care should be considered within the good mother and supermom discourses. Choi and colleagues (2005) argued that a prevailing ‘supermom’ ideology, the idea that motherhood is natural and fulfilling, has negatively influenced how new mothers interpret postnatal experiences and their behaviour. In their qualitative study of 24 Australian mothers, they found women were mostly disappointed with their postnatal experiences. Most of the women believed they should have been able to cope with their infants and domestic responsibilities; when they found themselves struggling they perceived themselves as not meeting the expectations of motherhood. Rather than asking for help, the women in their study tried to hide their struggles so as not to appear inadequate in terms of sacrificing themselves to care for others. In the context of physical activity decision-making, it is possible some mothers might feel they are not meeting expectations of motherhood if they have to ask others for help (i.e., watching their children or helping with domestic responsibilities) so that they can be physically active. Lewis and Ridge (2005) interviewed 40 Australian mothers about physical activity and highlighted the effects of the good mother and supermom discourses in creating challenges for mothers to be physically active; they argued that these discourses partly explained what the barrier ‘lack of time’ might mean to mothers. Some women in their study explained they were not physically active because their children’s needs came before their own (i.e., they took up the good mother discourse). Women used the supermom discourse to explain that they believed they should be able to handle their multiple responsibilities (e.g., taking children to sports, putting children to bed, and cleaning). The adoption of the supermom and good mother discourses appeared to reduce the value that the women had placed on physical activity. Lewis and Ridge argued that some mothers positioned physical activity as a non-priority in the context of their other responsibilities; in combination with the ethic of 28  care ideology the supermom and good mother discourses might have influenced women to put needs of others and domestic responsibilities before their own needs in their decision-making. While the feminist and gendered qualitative literature highlights sociocultural influences on physical activity decisions authors have bound physical activity decisions to gender and gendered norms (McGannon & Schinke, 2012; Miller & Brown, 2005; Thompssen, 1999) and presented broad societal change about social beliefs about motherhood and the role of a mother as ways to change women’s activity patterns (Lewis & Ridge, 2005; McGannon & Schinke, 2012; Miller & Brown, 2005). Social constructionist gender theorists’ have problematized these types of interpretations by arguing against set constructions of gender and sex roles (i.e., the role of mother) and the over-simplification of relationships between gender and health behaviours (Courtenay, 2000; Kimmel, 1986). Similarly to gender, there are a plurality of ways women take up and construct motherhood because motherhood is enacted in diverse social contexts of daily living, which has further informed ways of thinking about motherhood (Courtenay, 2000). Thus, constructions about gendered ways of thinking in relation to motherhood (i.e., supermom discourse) cannot singularly account for decision-making around health because complex intersections of sociocultural, personal, and structural contexts have influenced women’s perceptions about health and health behaviour, such as physical activity (Courtenay, 2000). A striking bias in the feminist and gendered research literature is its emphasis on reasons for inactivity; only three studies explored women’s beliefs about the processes of being physically active and only one did not sample elite athletes (Appleby & Fisher, 2009; Lewis & Ridge, 2005; Palmer & Leberman, 2009). Albright and colleagues (2006) found variation in women’s patterns of physical activity pre and postnatally; 26% were active before and after birth and 13% became active post-birth. Their findings emphasized the importance of understanding not only reasons women are inactive during the postpartum period but also reasons they give for deciding to be active. Lewis and Ridge (2005) found that active women in their study justified their activity as something that made them better mothers and partners; they argued that there is opportunity for mothers to 29  contest the ethic of care and supermom discourses by reframing physical activity as something contributing to family wellbeing. Appleby and Fisher (2009) interviewed 10 elite athletes (marathon-runners) about their return to physical activity postnatally and described women who believed returning to running involved a difficult renegotiation of identity. Their findings are similar to those of Palmer and Leberman (2009) who explored nine elite athletes’ perceptions of returning to physical activity post-birth. A greater emphasis on physical activity engagement could support understanding about how women make decisions to take up or maintain physical activity postnatally. Physical Activity Literature Summary  My analysis about general and postnatal physical activity literature has revealed its lack of emphasis on processes of decision-making. Authors have criticized lack of attention to women’s meanings around physical activity, contexts in which decisions are made, and explanation of relationships between factors/facilitators/and barriers to physical activity (Frolich & Poland, 2007; Mielewczyk & Willig, 2007; Shim, 2002). The focus has been on traits/circumstances of individuals who are more physically active and how cognitions related to behaviour rather than the process of choices people made around physical activity and the contexts in which they made those choices. Physical Activity and Decision-Making  While the physical activity literature has generally been positioned as explaining behaviour patterns it also indirectly identifies factors that might be relevant in postnatal physical activity decision-making processes. In the context of the general physical activity literature, cognitive, personal, and environmental factors that are associated with physical activity behaviour might be considered decisional factors. Physical activity cognitions may affect the level of motivation people have to be physically active and the personal value they associate with 30  engagement in physical activity (Mack et al., 2011). Demographic factors (i.e., level of income) may be related to social contexts that influence cognitions that people have about physical activity (Mielewycyk & Willig, 2007). Physical activity researchers have advocated for further integration of cognitive decision-making perspectives into socio-cognitive models, which include perspectives about the importance of emotions and habits (Ekkekakis, Hargreaves, & Parfitt, 2013; Rhodes & Nigg, 2011). In the following section, I critically analyze the implications of habit and emotion research in relation to physical activity decision-making and conclude by considering one framework that might be applicable to understanding postnatal physical activity decision-making. The influence of habit.  There has been mixed support for the concept of physical activity (or inactivity) as a habitual behaviour.  It is believed that habits (i.e., repeated behaviours) form when individuals have repeated experiences with behavioural decisions (such as exercising) and receive positive feedback reinforcing the decision (Aarts Paulussen, & Schaalma, 1997). When people encounter contextual cues related to the activity (e.g., seeing a gym bag in relation to exercising) cognitions might automatically activate that reinforce or initiate habitual behaviour (Wood & Neal, 2007). Gardner, de Bruijn, and Lally’s (2011) meta-analysis found that the correlation between physical activity habit strength6 and physical activity behaviour was .44 over 7 studies. When authors have included other variables that might influence decision-making in the prediction of physical activity, the influence of physical activity habit strength is decreased. Rhodes, de Bruijn, and Matheson (2010) tested the influence of habit and the theory of planned behaviour variables on physical activity and found that habit strength only accounted for 7% of variance in behaviour, which was similar to Verplanken and  6	The habit strength tool measured automaticity of physical activity behaviour, participant’s identity with physical activity, and frequency of physical activity (Gardner, de Bruijn, & Lally, 2011). 31  Melkevik’s (2008) findings. Nonetheless, studies have supported the relevance of considering habits in physical activity decision-making. Habit research brings to the forefront the importance of goals and expectations in decision-making. Wood and Neal (2007) argued, in their theoretical article about goal-habit relationships, that habits represent traces of the initial goals that formed the behaviour. For example, a person’s habit of taking walks may have been initiated related to health goals. Wood and Neal broadened conceptualizations about goals because they argued that goals need not be oriented to the longer term; driving to work or getting the newspaper could be legitimate goals in the context of daily living. Moreover, behaviours performed to achieve goals have not been consistently viewed as positive; sitting on the couch (i.e., being sedentary) or smoking a cigarette could achieve the goal of relaxation (Wood & Neal, 2007). Wood and Neal also suggested that habits have functioned to deliver the “cached value” or the expectations that people perceive with the performance of the behaviour  (p. 847). Habitual television viewing or walking might function to deliver an expected ‘cached value’ of relaxation. Habits might be initiated by goals but sustained, in part, because people have held expectations about what performance of the habitual behaviour will deliver. Exploring peoples’ habits, expectations, and goals in relation to physical activity and their day-to-day lives might be highly relevant in the context of understanding peoples’ decision-making. Wood and Neal (2007) contended that people may not be able to holistically explain their habitual behaviour but by reflecting on habits they can readily infer goals and expectations that support their habits and their personal dispositions (e.g., I am an active person). Likewise, Verplanken and Melkevik (2008) argued that people have consciously assessed their habits as evidenced by their ability to complete habit questionnaires. Regardless of activity levels, asking people to reflect on their day-to-day habitual behaviour in relation to underlying goals and expectations might be useful to develop understanding about factors influencing decisions around physical activity. 32  Emotions and decision-making.  Cognitive psychology perspectives have highlighted the relevance of emotion in decision-making.  Emotion research has emphasized values and expectations and the cyclical and cumulative nature of decision- making. Emotions are states of mind that form in response to a stimulus; they can have various effects on the body, including physiological (e.g., increased heart rate), cognitive (e.g., priming the brain to process information in particular ways), physical (e.g., facial expressions), and behavioural (e.g., people have acted in ways to change their emotional state) (Winkielman, Knutson, Paulus, & Trujillo, 2007). Researchers have argued that decision-making cannot occur without the influence of emotions (van Woerkom, 2010). Paulus and Yu (2012) provided a useful model about the cumulative nature of decision-making and how emotions impact that process, which is consistent with other emotion-behaviour theorists’ views (Baumeister, Vohs, DeWall, & Zhang, 2007). Paulus and Yu indicated that people assess their options, assign value to each option, and decide how to act to make decisions; following the action (behaviour) people evaluate and process the outcome of their behaviour for use in future decisions. They suggested that emotions are intertwined with each of these steps. Paulus and Yu explained that peoples’ perceptions and valuing of their options have been based, in part, on their emotional responses to particular options. That is, a person’s emotional response to an option affected the weighted value of that option. The emotional response is closely connected to expectations of the outcome of that activity; expectations have been informed by previous experiences and the current context (Paulus & Yu, 2012). They claimed that peoples’ processing and evaluation of the outcomes of their behaviour affected their emotional responses in their next experience involving that decision. For example, a person’s negative experience with jogging (e.g., experiencing an injury as a result of jogging) gets processed and information about that experience might manifest in the future as a negative emotional reaction to jogging. Of course, the actual decision to go for a jog is influenced by multiple factors, but knowledge of the negative 33  jogging experience has the potential to affect the decisional balance towards jogging or not. A model that has emphasized how previous experiences influenced future decision-making is consistent with the life course perspective and highlights the cumulative nature of decision-making (Braveman & Barclay, 2009). Expressed perceptions about physical activity-related emotions might explain how expectations relate to physical activity decisions. An important consideration has been people misattributing their feelings to situations when they are actually linked to their internal emotional state (Baumeister et al., 2007; Laird & Lacasse, 2014). For example, a woman may feel guilty about leaving her children to go exercise but attributes her negative emotions to exercising itself even though she enjoys the actual experience of exercising. This point highlights the importance of exploring other influences in decision-making beyond actual experiences with physical activity. Eccles’ Expectancy-Value Decision-Making Framework  In consideration of the literature reviewed, one framework that could increase the sensitivity of researchers to factors influencing women’s physical activity decision-making during the postpartum period is Eccles and Wigfields’ (2002) expectancy-value decision framework because it links ideas about barriers, socio-ecological factors, role socialization, expectations, and goals. The framework has been applied to explain children’s decision-making around physical activity (Chiang, Bird & Molin, 2011), relationships between sport and gender, and decisions related to women’s education and career choices (Eccles & Harold, 1991; Eccles, Wigfield, & Byrnes, 2003; Eccles, 2008; 2009; Wigfield & Eccles, 2000). The framework has integrated concepts, such as expectations, values, goals, and personal, social, and policy/environmental influences on physical activity by acknowledging them as factors that influence physical activity decision-making. Other socio-cognitive theoretical frameworks could be considered decision-making models but they have had limitations in their abilities to incorporate the relevant factors discussed in this literature review and to account for decision-making in the context of other competing decisions (Mack et al., 2011). Because of the complexity 34  of Eccles and Wigfields’ expectancy-value framework and their extensive consideration of decision-making as cumulative across the lifespan and in the context of other competing decisions the framework could sensitize researchers to many elements of women’s physical activity decisional factors during the postpartum period (Eccles & Wigfield, 2002; Eccles, 2009). The expectancy-value framework has posited that people have positioned their values against their expectations for success in particular activities to determine whether or not they would engage in that activity (Wigfield & Eccles, 2000). For physical activity, the framework has suggested that expectations are informed by past experiences, self-belief of physical activity ability, and beliefs about the difficulty of physical activity (Wigfield & Eccles, 2000). Moreover, the value of physical activity is informed by a person’s long and short- term goals (i.e., how useful is physical activity to achieve his/her long and short-term goals), a person’s identity beliefs (i.e., does the person believe being a physically active person is important and consonant with who he/she believes him/herself to be), and the negative consequences of engagement (i.e., cost) (Wigfield & Eccles, 2000). Developers of the framework argued that the more strongly a person values an activity, the more likely he/she will persevere to engage in the activity regardless of her/his expectations for success (Eccles, Wigfield, & Schiefele, 1998). The framework has extended our thinking about habits to the exploration of goals in relation to behaviour. The framework has brought together many factors influencing physical activity because it has accounted for how expectations and values are layered and contextual and being influenced by culture, gender, socialization, personal memories, and particular situations. It has theoretically incorporated media, gender, and cultural socialization, as well as interaction with others in the immediate environment, as influences on beliefs about physical activity ability and value (Segar et al., 2011). For example, a partner’s beliefs about the value of 35  physical activity could influence a woman’s decision-making by affecting her beliefs about whether engaging in physical activity was valuable. The expectancy-value framework has also positioned decisions about physical activity in the context of other competing values (Eccles & Wigfield, 2002). Women may value behaviours because they are important for their long and short-term goals and self-identity, including behaviours around being a ‘good’ mother, partner/spouse, friend, employee, and family member. The framework has supported the view that behaviours required to successfully meet a woman’s expectations of fulfilling those roles may compete with her desire to be physically active. For example, women who are living in the postpartum period have reported experiencing fatigue from engaging in the activities they believe are necessary to be a good mother and employee (e.g., working long hours, dropping children off at activities, making dinner, cleaning the house, and putting children to bed) (Taveras & Oken, 2008). For those women, meeting what they hold as role expectations for mothers might have been more valuable than getting sufficient physical activity because such behaviours are met with more social approval. Thus, the expectancy value framework has the potential to sensitize a researcher to explore not only women’s expectations and experiences with physical activity but also women’s values across a range of roles and activities and how their values relate to decisions about physical activity. Chapter Summary  Despite inactivity being related to increased risk of chronic disease and obesity research has suggested that women decrease physical activity following childbirth and do not return to pre-pregnancy levels (Bellows- Riecken & Rhodes, 2008). Researchers’ interventions to increase women’s postnatal physical activity have often been ineffective (Hartman et al., 2010). Postnatal physical activity research has explored women’s beliefs about reasons for inactivity and how gender socialization affects physical activity decision-making; however, it 36  has provided limited accounts of the contexts in which women make decisions about physical activity and the meanings women ascribe to physical activity during the postpartum period. Physical activity research has focused on using factors related to physical activity to explain physical activity behaviour. I argue that studies attempting to explain physical activity behaviour can be extended by undertaking qualitative research to explore women’s processes of decision-making while attending to varying contexts potentially affecting women’s beliefs about physical activity. Approaching exploration of women’s physical activity decision-making when the researcher is sensitized by an expectancy-value framework offers promise for explaining women’s decision-making about physical activity in the postnatal period. In this chapter, I have critically reviewed the general/postnatal physical activity, habit, and emotion literature and situated the current study in the context of this research. In the next chapter, I critically review symbolic interactionism, the theoretical perspective underpinning the current study. I describe the strengths and limitations of grounded theory and explain my ontological and epistemological positioning, and theoretical and methodological considerations. 37  Chapter 3: Theoretical Perspectives and Methodology  Research methods should flow from the underlying epistemology/ontology, theoretical perspectives, and methodological positions from which the researcher is operating. Epistemology is concerned with the nature of knowledge, primarily, who can provide it, how to generate it, and the meaning of the knowledge generated (Harding, 1987). While researchers tend to conflate methodology with theoretical perspectives, I consider a theoretical perspective as a philosophical stance about the nature of the operation of the world, while I consider methodology to be a theory about how research should be conducted, given the guiding theoretical perspectives embedded in a particular epistemological stance (Crotty, 1998; Schreiber, 2001). The term method is reserved for the literal research techniques (e.g., surveys, interviews) (Crotty, 1998). The theoretical perspective informing this study was symbolic interactionism, while the guiding methodology was grounded theory. In the following sections, I describe symbolic interactionism, discuss critiques of symbolic interactionism in the context of developing knowledge about physical activity decision-making, and position myself in terms of the various traditions of grounded theory (i.e., Charmaz, 2006; Glaser, 1978; Strauss & Corbin, 1998;). I conclude with a summary of my epistemological assumptions and ontological position given my theoretical perspectives and methodology. Symbolic Interactionism  In this section, I describe the general perspective of symbolic interactionism, its critiques, and the direction that this theoretical perspective provided for this study. Symbolic interactionism was a particularly valuable theoretical perspective in the context of developing a theory about women’s physical activity decision-making during the postpartum period because it emphasizes the influence of context on decision-making and that peoples’ decisions are based on perceptions about what those decisions will mean for them (Blumer, 1969). Such a perspective fits with my definition of physical activity decisions as subjectively justified and influenced 38  by personal meanings and multiple contextual factors (Reisberg, 2006). Symbolic interactionism is usually credited to philosopher George Herbert Mead and his student, sociologist Herbert George Blumer. Mead (1934) used several key concepts in his thinking: object, act, symbol, mind, self, and society; key to Mead’s thinking is the view that these concepts emerge through social interaction and are dependent on social interaction. He was deeply concerned about how people make meaning, their behaviours, communication, consciousness, the development of self and mind, and how meanings inform behaviour. The object.  Symbolic interactionism suggests that what people can know depends on their interactions with others. Mead (1934) considered objects as essentially any ‘thing’ (p.71) that people can act towards. Blumer (1969) more precisely explained objects as anything that someone could make reference to, whether it is physical (i.e., a house), social (i.e., a teacher), or something abstract, such as “moral principles, philosophical doctrines, or ideas, such as justice, exploitation, or compassion” (p. 11). Essential to a symbolic interactionist perspective is the position that individuals can only learn what objects are through interaction with others (Blumer, 1969). Meanings emerge from interaction. For example, a person can only know what an exercise personal trainer is because he/she has interacted with others and learned what a personal trainer is and associated meanings attached to the object personal trainer. Because objects are socially constructed Mead (1934) argued that objects can vary in their meaning for different people and in different situations. For example, for a competitive athlete, physical activity may be associated with prestige and his or her career, whereas for a person who swims recreationally it could be associated with enjoyment or relaxation. Meaning might also vary by context; for the athlete in a competition situation physical activity may carry different meanings than it does when the athlete is going for a recreational run. 39  The meanings a person has toward an object are important because they influence peoples’ decision- making and how they act toward the object. Blumer (1969) argued that peoples’ behaviours towards objects are guided by the meanings those objects have for them. For example, rather than do an indoor physical activity (e.g., yoga) that she enjoys, a mother might nap when her infant is napping because she believes it will give her energy but physical activity will not. In that context, her association of physical activity with inducing fatigue affects her decision to be physically active. Symbols, the act, and symbolic interactionism.  The term symbol in symbolic interactionism relates to the commonality in understanding between people that arises through social interaction; symbols are how people represent ideas in thought and how they communicate them to others (Mead, 1934). Mead explained that humans think in symbols; a person’s representation of an object (i.e., a car) in thought is a symbol, something that represents that idea. Likewise, he argued that people communicate ideas to others using symbols, by using words, body language, etc. to represent collections of ideas. Symbols become significant when they invoke a response in the person using it as well as the person they are directed towards because there is shared meaning about the symbols; however, symbols are not significant if the person does not understand the intention of the other (Mead, 1934). For example, if a person did not understand a fitness instructor’s hand cue used to encourage the person to move in a particular direction, the cue would be non-symbolic. There is a lack of mutual understanding of what that cue represents and so the individual does not respond accordingly. Blumer (1969) argued that people engage in an ongoing process of interpretation of their own and others’ actions to make meanings that influence decision-making and behaviour. He suggested that these processes of interpretation occurred during acts, originally conceptualized by Mead (1934). Mead called an act between two people a social act, comprised of symbolic and non-symbolic gestures, and interpretation. Mead 40  defined a gesture7 as something a person does that makes the other respond (e.g., raise her voice, move towards the person) and symbolic gestures, as those having the function of indicating what the other person should do (e.g., using a hand motion to indicate the person should sit), what the gesturer is planning to do, and what is happening. Mead explained that the social act occurs as follows: a person gestures to another; the other interprets that gesture and responds with another gesture, and so on. Blumer likened Mead’s description of processes to fitting lines of action during the social act; a person interprets a gesture and acts according to his/her interpretation of what is going on in the situation, whether or not his/her interpretation was accurate. Thus, Blumer explained that personal interpretations substantially influence decision-making and subsequent behaviour. Blumer (1969) also stipulated that a social act was not only between two people, but could also be between organizations or social groups. He indicated that the interpretation at that level involved group discussion to develop a line of action. For example, in a business merger, the parties involved would interpret their own and the other party’s behaviour to discern how to act. Crucially, each party’s interpretation of his/her own behaviour and that of others is influenced by some set of common understandings about appropriate lines of action in that particular situation. Notwithstanding commonality in understanding about appropriate ‘merger’ norms, Blumer argued that social action was not constrained to these normed patterns of behaviour. Emergence of the act.  Blumer (1969) and Mead (1934) were careful to avoid suggesting that the meanings people attach to certain objects necessarily direct behaviour because they viewed behaviour (the act) as variable and emergent; they believed behavioural decisions depended on a particular situation. For example, in an interaction between  7	The gesture was a novel concept Mead (1934) developed from philosopher William Wundt that helped address the standstill of parallelism existing in psychology/sociology at the time, which Mead described as the ontological problem of an object and a person’s interpretation, but nothing that binds the person’s interpretation to the object. 41  two people during a job interview, each would person bring his/her behavioural tendencies and understandings but how those influence the interaction is unpredictable. If a person who normally associates interviews with stress is interviewed by a manager who turns out to be an old friend the interviewee may be more outgoing and at ease during that particular interaction because he/she interprets the manager as friendlier or more likely to give him/her the job. Likewise, a manager normally stern during interviews might be less so because he/she is interviewing an old friend. Hence, the behaviours of both the interviewee and manager emerge during the interview based on their interpretations of the situation. Not only does behaviour emerge during interaction but also meanings people attribute to a situation arise from the situation itself (Mead, 1934). How a person understands and makes meaning of a situation depends on what they believed happened in that situation. For example, how a track and field athlete interprets a loss in competition depends on any number of things: the reaction and interaction with her/his coach, family, and other competitors following the loss; what happened in the weeks leading up to the competition; what happened on that particular morning; and whether he/she had injuries, etc. If his/her coach believed that there was something unfair about the race, this view could influence the athlete’s construction about how valid the loss was; the athlete may construct the meaning of the loss as not related to lack of skill per se but the unfairness of the race. Hence, the athlete’s construction of the loss emerged from the situation itself, not simply because of how the athlete was preconditioned to understand it. The athlete’s construction of the situation represents a new object that will influence the athlete’s decision-making around competition in the future; this reflects a final important point about the act; through the act and use of symbols new objects are created (Mead, 1934). Mind.  The notion of emergence is important to the symbolic interactionist perspective in general, and can be found in Mead’s (1934) thinking about the mind, self, and society. Mead argued that a person’s mind only 42  exists in relation to others because a person can only think about those significant symbols that he/she has internalized through interaction. For example, a person can only feel sad about a competition loss because of internalized social meaning about what a loss represents. Rather than signifying people as incapable of unique thought Mead contended that unique combinations of social experiences allowed individuals to generate unique thought. Thus, he positioned human thinking as constrained by social interaction but also enabled by the diversity of social interactions. Mead (1934) depicted humans as rational, thinking animals but also acknowledged the role of the unconscious in behaviour, albeit somewhat abstractly. Blumer (1969) downplayed the role of the unconscious in human behaviour; although he acknowledged that people could behave habitually, he also constructed humans as rational actors that intentionally selected lines of action. Mead argued that reflexiveness, or the ability of humans to rationally analyze their behaviour and experiences, is what set humans apart from other animals and the primary condition that made mind possible. However, Mead also regarded the mind as implicitly producing and testing lines of action to determine how to act, with action being possible without the actor (person) being conscious of it. For him, the unconscious central nervous system was a part of the act because it provided information about past experiences in the present. Unfortunately, Mead did not clearly explicate how he understood human behaviour as both rationally and implicitly guided. It may be that he assumed people sometimes reflected and determined lines of action thoughtfully and, at other times, decision-making around behaviour was guided more implicitly, which is consistent with decision-making models that individuals incorporate both implicit and conscious reasoning during decision-making (Evans, 2008). Self.  For Mead (1934) and Blumer (1969) the self was co-dependent with the mind and continually developing through social experience. Mead described the self as an object to itself; he believed that people 43  could consider themselves and act toward themselves (i.e., people can reflect on their own behaviour and decide to change their behaviour in future). Mead also explained that people could never experience themselves directly rather he believed that it was through taking the standpoint of others (or groups of others) that people experienced themselves. For Mead, a person would interpret and understand his/her behavioural decision and self given how she/he understood others would interpret it. For example, a mother might experience feeling guilty when exercising because she believed others might view her as selfishly taking time to exercise. Thus, Mead and Blumer’s perspectives about the co-dependence of thinking and interpretation in relation to others emphasized the importance of social interaction in developing perceptions. Mead (1934) also distinguished between the ‘I’ and the ‘me’; two concepts that highlighted the emergent nature of the mind and self. The ‘me,’ he explained, represented the organized self, the self an individual has become through interaction with themselves and others. The ‘I’ represented the in-the-moment act of the individual in response to others and the self. The ‘I’, as such, was not accessible to the individual because the ‘I’ is always in a state of becoming the ‘me’ that one can reflect on (Mead, 1934). The notion of the ‘I’ always integrating into the ‘me’ highlights the perspective of symbolic interactionism that reality for individuals is not fixed; the self (me) is not fixed, but changing through interaction (Blumer, 1969). Society.  The perspective, about self as an emerging process, extends into Mead (1934) and Blumer’s (1969) conception of society. Society, for Mead and Blumer, was composed of ongoing interaction between groups and individuals. Structures, such as social norms, social roles, culture, and institutional organizations, were viewed as having an influence on ways of thinking and behaviour but, rather than acting as determinants of behaviour, they were merely acting as frameworks for behavior (Blumer, 1969). Structures were not denied in a symbolic interactionist perspective; both Blumer and Mead agreed that structure influenced patterns of thinking and 44  behaviour but Blumer argued that humans interacted with each other rather than interacting with structure. Critical to their view was that humans were seen as actors constantly dealing with the situations they were in rather than behaving and thinking as though they were preconditioned. Their viewpoints contrasted with the traditional structuralist view of society, which presupposed society was composed of static structures that shaped thinking and determined behaviour (Blumer, 1969). Blumer (1969) argued that the greatest problem with the structuralist view was that it ignored the role of the self as a reflective being who could choose his/her own behaviour (and to that end acknowledged Mead’s (1934) explication of self as his most important intellectual contribution). Thus, Blumer argued that people have agency to change existing patterns of behaviour with themselves and with others. The appearance of stability in society, Blumer (1969) explained, was due to common understandings between people within and between groups. He argued that, while a person might engage in interpreting his/her particular situation, often there is mutual understanding between groups and individuals (developed from a history of interactions) that helps guide interpretations and action with relative ease. However, for any number of situations (e.g., novel situations) mutual understanding may not be present and new lines of action are constructed that change these patterns; his statements suggest that behaviours or patterns of decision-making are not necessarily predictable (Blumer, 1969). Application of symbolic interactionism to the study.  Examining the alignment of Blumer’s (1969) perspectives with current critiques of physical activity research reveals the exceptional fit of symbolic interactionism as a guiding theoretical perspective for this study. There has been a lack of empirical investigation about how women understand and make physical activity decisions. The emphasis of the theoretical framework of symbolic interactionism on meaning making and the 45  importance it places on social interaction guided me to explore how the women understood physical activity and their physical activity decision-making in relation to others. Another critique of the physical activity literature is the assumption that physical activity behaviour exists as a static entity; in contrast, physical activity behaviour can be unpredictable and is likely influenced by the context in which the person makes the decision (Mielewycyk & Willig, 2007). Mead’s (1934) major contribution was that behaviour is not necessarily predictable; that concept was further developed by Blumer (1969) who suggested that people constantly engage in thinking about their current situations and decide how to act given their understanding of their situations. These perspectives supported my research assumption that understanding context is important. The symbolic interactionist perspective guided me towards understanding the context of decision-making for postnatal physical activity behaviour and how varying contexts related to variation in the women’s beliefs and values about physical activity behaviour. Blumer (1969) also argued against reality being fixed and the predominant methodology of ‘sociological determinism,’ used during his time. He defined sociological determinism as the idea social interaction was a medium for pre-conditioned attitudes, roles, norms, values, etc.8 to be played out in a predictable pattern. Given my guidance by symbolic interactionism I was cautious about presuming that particular structures influenced behaviour because I assumed that behaviour is emergent and based on participants’ interpretations of their situations. The emphasis on emergence and social processes in symbolic interactionism enhanced my sensitivity to processes of decision-making around physical activity behaviour. Because Mead (1934) believed that the act involved both a conscious and unconscious element, and the unconscious element was informed by previous experiences his thinking is consonant in some ways with current thinking around decision-making. However, neither Blumer (1969) nor Mead commented on peoples’  8 By structure, I refer to that which can have an effect on our cognitions and behaviour (Akram, 2012; Elder- Vass, 2007). Structure can be social, relational, institutional, or embodied (Elder-Vass, 2008). 46  ability to access their unconscious in explaining their behaviour; therefore, it is not possible to make assumptions about their understandings of unconscious access. Based on decision-making literature I assumed there were limitations to peoples’ abilities to access their unconscious to reflect on their decisions (Greenwald & Banaji, 1995; Greenwald et al., 2009). Blumer’s (1969) major contribution to the symbolic interactionist perspective (beyond creating the term itself) was an explanation of how to apply Mead’s (1934; 1938) philosophical position towards research. His position concerned the need for synchrony between epistemological/ontological assumptions and research methodology. Because he believed people acted on the basis of the meanings things had for them he also believed that the primary task of the researcher was to seek understanding of participants’ perspectives (i.e., understand how they understood their actions and the meanings they attributed to them) to develop theories. Blumer viewed the participant as the expert knower rather than the researcher. He argued for an exploratory approach to research, whereby lines of inquiry changed and evolved as the researcher began to understand the participants’ perspectives. He encouraged use of a broad number of research strategies, including interviews, observation, diaries, letters, public records, and focus groups. Importantly, he argued against the use of predetermined scientific protocols (e.g., using hypotheses, and statistical analyses) because he viewed these approaches as failing to develop adequate knowledge about the meanings and contexts of the participants under study. In answering the research question I posed, I took Blumer’s recommendations into account by using grounded theory methodology, which emphasizes an exploratory approach. Symbolic Interactionism Critique  Criticism about symbolic interactionism has varied over time but has primarily focused on symbolic interactionism’s astructural bias, lack of attention to power, and subjective ontology (Fine, 1993). Investigators defending symbolic interactionism have argued that many critiques may not accurately represent symbolic 47  interactionism but rely instead on how it has been interpreted by individuals (Dennis & Martin, 2005; Maines, 2001; Snow, 2001). While it is important to consider these critiques, it is also important to examine their legitimacy in relation to the works of Mead (1934) and Blumer (1969) and pragmatism, a philosophical perspective underpinning of symbolic interactionism. A recurring critique involves the view that symbolic interactionism aligns with subjectivist ontology, which posits the existence of multiple realities based on different individuals’ perceptions (Crotty, 1998). Burbank and Martins (2010) contended that symbolic interactionism held an objective relativist stance; reality was relative to the person interpreting it. Critics have suggested that the emphasis of symbolic interactionism on understanding peoples’ subjective experiences and their interpretations conflicts with legitimate knowledge and theory development about human behaviour (i.e., if the knowledge sought is subjective it is difficult to apply it objectively to theory development) (Burbank & Martins, 2010; Fine, 1993). Critiques of symbolic interactionism that emphasize its subjectivism appear problematic. Mead (1934) obviously aligned with social constructionism. Social constructionism is a philosophical stance that represents knowledge as created and passed on through social interaction and claims that people can only interpret their world based on understanding derived socially (Crotty, 1998). Reality is neither objective nor subjective but it is constructed through interaction; there is a reality but it is a reality that is only viewed through socially derived understandings (Crotty, 1998). Mead clearly did not think reality was subjective because he wrote: “one individual has one experience and another has another experience, and both are stated in terms of their biographies; but there is in addition that which is common to the experience of all” (p. 33). Mead’s perspective indicated that, while he thought people had their own interpretations of their situations, the nature of self as a social being that emerged through social interaction reinforced people’s experiences of commonality in interpretation. In other words, a person’s reality was bound to others, rather than existing independently. 48  The weight of the subjectivist critique has lessened as critical and post-structural perspectives have become more commonly accepted theoretical perspectives used to guide inquiry; in those perspectives, knowledge is accepted as partial. For example, critical perspectives generally hold reality can never be known because knowledge is always value-laden, contextual, perspectival, and historical (Kincheloe & McLaren, 2005). People’s knowledge is acquired given their experiences, which makes particular knowledge available to them (Kincheloe & McLaren, 2005). The purpose of critical inquiry is not to determine an objective reality but to understand participants’ reality in the best possible ways so as to develop meaningful strategies to address inequities and power imbalances (Kincheloe & McLaren, 2005). Similarly, the purpose of symbolic interaction is not to develop objective knowledge but to understand the perspectives and the behaviour patterns of those under study to develop useful theories about their beliefs and behaviour (Blumer, 1969). More recently, symbolic interactionism has been viewed as in opposition to critical perspectives; hence a major critique has concerned its astructural bias and lack of attention to how power shapes behaviour because structure and power are major tenets of critical inquiry (Bilton, 2002; Fine, 1993). Those critiques about symbolic interactionism have also been extended to grounded theory methodology, which is often guided by symbolic interactionism (Milliken & Schreiber, 2001). In the nursing literature, critical inquiry9 is proposed to diverge from symbolic interactionism in terms of its purposes, goals, and concepts of interest; that divergence is argued to reveal the weaknesses of symbolic interactionism (Burbank & Martins, 2010; Martins & Burbank, 2011; Sundin & Fahy, 2008). In brief, critically guided health research is concerned with exposing how societal structures (e.g., political, social, and economic), ideology (taken-for-granted assumptions that structure everyday life), power, and intersecting social locations (e.g., culture, gender) shape and/or perpetuate health inequities to consider necessary changes to equalize  9	Critical inquiry includes, but is not limited to the broad streams of research, such as feminist, masculinity, critical social theory, post-structural, post-modern, and post-colonial (e.g., Reimer-Kirkham et al., 2009). 49  inequities (Kinchloe & McLaren, 2005; Lyons & Chamberlain, 2006, Martins & Burbank, 2011). Critical perspectives typically assume that people’s behaviour is constrained by structures and power and focus on the need for societal level changes (i.e., macro-focused) (Kincheloe & McLaren, 2005). With its emphasis on individual meaning-making, interpretation, and concepts, such as the self and mind, symbolic interactionism has been positioned as individually focused, and criticized as downplaying or even ignoring the effects of structure and power on behaviour (Bilton et al., 2002; Burbank & Martins, 2010; Dennis & Martin, 2005; Fine, 1993; Snow, 2001; Stryker, 1980; Sundin & Fahy, 2008). Fine (1993) argued that there are limits to peoples’ agency because actions have consequences; people do not have free will to act and behave according to their interpretation as suggested by proponents of symbolic interactionism. Others have indicated that Blumer’s (1969) insistence that institutional structures do not determine behaviour results in symbolic interactionist researchers having limited understanding of the effects of structure on behaviour (Bilton et al., 2002). Moreover, symbolic interactionism has been criticized because the goals and purposes are not clearly delineated beyond an understanding of people’s perspectives and the development of theory (Bilton et al., 2002). Thus, from a critical inquiry perspective, the claim is that symbolic interactionist research is limited; it may lead to understanding about peoples’ behaviour but it will do nothing to change existing structures and power imbalances that lead to health inequities (Sundin & Fahy, 2008). Response to the astructural bias critique.  Some authors have suggested that the astructural bias criticisms lodged against symbolic interactionism reflect a limited portrayal of the perspective. Snow (2001) argued that the oft-quoted assumptions of Blumer (1969) (i.e., meaning is learned through interaction, people interpret situations and act according their interpretation) have been taken up as representing the perspective when, in actuality, symbolic interactionism encompasses much more content than these assumptions; it includes an emphasis on structure. Both Blumer and 50  Mead (1934) argued against structure as a behavioural determinant rather than denying its influence on behaviour; thus, symbolic interactionist researchers have accepted the influence of structures (e.g., social norms, policies, and economic context) and ideology on behaviour. Using the symbolic interactionist perspective; however, requires the researcher to avoid preconceptions about which structures and ideologies might be influencing behaviour; those assumptions narrow what the researcher can understand. For example, physical activity research guided by a postcolonial critical perspective, which is a complex perspective, but involves consideration of intersecting oppressions, such as race and class, for example, is likely to find these to be the relevant influences in physical activity decision-making (Reimer-Kirkham, Baumbusch, Schultz, & Anderson, 2009, p. 353). Nonetheless, symbolic interactionist researchers are expected to be aware of a range of possibly relevant influences on behaviour, including structural factors. Response to the power and agency critique.  Dennis and Martin (2005) argued that using symbolic interactionism as a theoretical approach reveals power relationships; they objected to the reification of power by critical traditions as a distinct ‘entity’ that influences behaviour. Drawing on Foucauldian depictions of power as omnipresent, they suggested that studying social interaction between groups and individuals is as likely to uncover power relationships and constraints on behaviour as macro-sociological approaches. Using research examples of deviant behaviour and alcohol and marijuana use, they described how symbolic interactionist work has revealed processes by which laws and legislation are “established, enforced, challenged, or broken” and explained the work of people enforcing or reacting to those laws and legislation (Dennis & Martin, 2005, p.198). Furthermore, the work that they described has illuminated the complexity of the term ‘deviance’; they found that understanding of the term varies among social groups who construct their own interpretations of behaviour considered deviant. They 51  argued that symbolic interactionist work explicitly explains how power is created, enacted, contested, and reproduced through social interaction. By using this theoretical framework, I anticipate that my work about women’s physical activity decision-making will develop my understanding about power and agency through women explaining their beliefs about constraints affecting their decisions about physical activity. These constraints might extend into social, political, and environmental contexts that would reveal how power affects women’s agency beyond interpersonal interaction. Pragmatism and the goal of symbolic interactionist work.  Dennis and Martin (2005) also claimed that power and structure have limited bearing on work guided by symbolic interactionism given its pragmatist underpinnings. Authors often refer to a pragmatist-symbolic interactionism link but they have rarely clearly delineated how this philosophy grounds the symbolic interactionist perspective (Denzin, 1994; Martins & Burbank, 2011). Pragmatism, a philosophy believed to have informed Mead’s thinking, is typically associated with the 20th century philosophers John Dewey, William James, Charles Pierce (Scheffler, 1975), and more recently and prominently, Richard Rorty (1999). The primary interest of pragmatists is the practical significance of theories and concepts (James, 1907). Specifically, they are interested in whether theories and concepts help in understanding ways people deal with and manage problems (James, 1907; Kallen, 1954). Pragmatists see theories and science as tools to help solve problems, not as means to help develop a truer understanding of reality (James, 1907; Kallen, 1954; Rorty, 1999). Truth is not seen as something that exists per se; rather ideas are true insofar as they work to help solve problems (James, 1907; Rorty, 1999). When a truth no longer works, it is discarded (James, 1907; Rorty, 1999). 52  To that end, for pragmatists, concepts such as structure and power are judged by their utility, rather than their validity as real structures that affect behaviour (Dennis & Martin, 2005). Pragmatists also dispense with theoretical dualisms, such as structure-agency and macro-micro because what is relevant to pragmatists is not the application of predetermined dualisms about structure or agency but how relevant those views are to people in real-world situations (Dennis & Martin, 2005; Rorty, 1999). Power dynamics, viewed by people as constraints or facilitators, may be highly relevant to decision-making about physical activity but their degree of importance is only understood through engagement with people about their physical activity decision-making; they are only as important as they are to the people dealing with their situations. Thus, pragmatism encourages a critical stance of engagement with commonly accepted theories and terms to consider their relevance (Doane & Varcoe, 2009). In essence, it grounds the researcher to participants’ real-world interpretation and situations rather than allowing for theorizing out of context about the nature of participants’ experiences by researchers (Dennis & Martin, 2005). Pragmatist thinking is evident in symbolic interactionism and relieves the burden of the criticism that symbolic interactionist work does nothing to change underlying structures and power. Symbolic interactionists seek to understand how participants in their real-world situations act in their situations so that interactionists learn and develop ways with participants to solve their problems (Blumer, 1969). Furthermore, in the context of physical activity decision-making, symbolic interactionist work would not simply be focused on development of individual level interventions (e.g., counseling); participants’ understandings may also reveal structural and power concerns that influence their behaviour (e.g., lack of funds for transportation access to fitness facilities). Introduction to Grounded Theory I used grounded theory methodology to guide data collection and analysis. Grounded theory is a relatively new methodology. Explicitly credited to Barney Glaser and Anselm Strauss; it has undergone 53  significant shifts in methodological thinking since its emergence in the 1960s (Charmaz, 2006; Bryant & Charmaz, 2007a). In the process of conducting their (now iconic) sociological work about the death and dying, Glaser and Strauss (1967) developed grounded theory as a novel qualitative methodological approach to understand peoples’ experiences and explain behaviour. In brief, grounded theory involves identifying a core category that helps to explain the most variation in participants’ behaviour (Glaser, 1978). Unique to the method was constant comparative analysis, which involved using participants’ language and understandings to build codes that were then continuously compared to each other to develop more abstract codes and concepts to create a core category (Glaser, 1978). The resulting theory that developed from this process included a set of interrelated categories that helped explain how participants manage their main concern (Glaser & Holton, 2004). Participants’ main concern can be likened to participants’ primary goal or ‘prime mover’ that, in this case, drives physical activity decision-making (Glaser, 2004, Getting Started, para.1). For example, Hjamhult and Lomborg (2012) conducted a grounded theory study with new Swedish mothers (6-12 postnatal weeks) to explore how they managed newborn care. They described mothers’ main concern as preserving control and integrity; the core category was prioritizing newborn care. Prioritizing newborn care helped them manage the concern of preserving control and integrity while the categories, developing competence, changing focus in relationships, and stretching to the critical level represented interrelated strategies mothers performed to prioritize infant care. Bryant and Charmaz (2007b) described how the research environment of positivism in the 1960s and 1970s (characterized by a belief in truth and science as a means to develop objective knowledge about the world) accounted for many of the initial procedures of grounded theory. Glaser and Strauss (1967) provided a qualitative method that, in many ways, mirrored quantitative research approaches (Walker & Myrick, 2006). Charmaz (2006) argued that grounded theory emphasized rigour and a systematic data collection approach to 54  approximate quantitative procedures. Glaser and Strauss’ method of using constant comparison during data analysis was considered by researchers as a way to decrease the researcher’s influence on theory development (i.e., reduce bias) and generate a theory as close to participants’ perspectives as possible (Bryant & Charmaz 2007b; Charmaz, 2000). Incorporating elements of the positivist research environment was necessary for the grounded theory methodology to gain traction and legitimacy in that era (Bryant & Charmaz, 2007b). Grounded theory was distinguished from quantitative approaches that involved constructing, testing, and verifying hypotheses because Glaser and Strauss (1967) emphasized its inductive nature (Bryant & Charmaz, 2007b; Walker & Myrick, 2006). Glaser and Strauss (1967) argued that the grounded theorist researcher does not enter the study with preconceived hypotheses and sampling strategies; the theory is constructed using participants’ perspectives to develop generalizations about their behaviour. More recently, it has been argued by several authors (Bryant & Charmaz, 2007b; Richardson & Kramer, 2006) that grounded theory more accurately uses an abductive10 reasoning approach (specifically, Strauss’s later grounded theory approach); although, neither Glaser nor Strauss used the term, Glaser (1992; 2012) consistently referred to grounded theory as inductive, and others have argued against the association of grounded theory with abduction11 (Reichartz, 2007). Notwithstanding the debates, grounded theory avoids primarily testing hypotheses derived from pre- existing theories (i.e., an initial deductive approach) as the means to develop knowledge.    10 Abduction is an integrated inductive and deductive approach to reasoning first introduced by pragmatist Charles Peirce. It involves inductively generating novel ideas and then testing those ideas deductively (Peirce, 1957).  11 Actually, Glaser (1978) mentioned that deductive reasoning is used in grounded theory in the service of inductive logic. He explained the researcher uses deductive reasoning to consider where to sample next and hypothesize about the developing theory; however, he emphasized this is driven by the inductively derived codes. Rather than starting with a preconceived framework and hypotheses, the framework is built from the data, so Glaser considers the approach inductive rather than deductive. 55  Grounded theory’s alignment with symbolic interactionism.  The theoretical perspective of symbolic interactionism is almost always positioned as underlying grounded theory methodology likely because grounded theory is a logical extension of symbolic interactionist epistemology. One of Blumer’s (1969) main theses was that researchers should be explicit about and align their epistemological perspectives with their research methodologies. He argued that a study undergirded by symbolic interactionism would best fit with an exploratory research methodological approach. Grounded theorists do not adhere to a prescribed framework of inquiry and dispense with preconceived explanatory frameworks and a priori hypotheses (Glaser, 1978). In a grounded theory approach, the types of data, the nature of the sample, and the questions asked of the data are left open and become a product of the research process itself, arising from the researcher’s ongoing analysis of the data and theorizing about the types of data required to build the grounded theory (Glaser, 1978). Hence, grounded theory methodology emphasizes a quality of emergence, much like Blumer (1969) and Mead’s (1934) perspectives about the emergent nature of human action, mind, and society. Symbolic interactionism’s assertions that people interpret their situations and construct meaning align with grounded theory’s goal to develop a theory that explains how people understand their situations (Milliken & Schrieber, 2001). Milliken and Schrieber (2001) argued that the constant comparative method, which ensures that the researcher integrates the perspectives of participants in his/her construction of the theory, is an extension of Blumer’s (1969) perspective about researchers learning the meanings people attach to their behaviours through exploration and personal interviews. The core category developed in grounded theory is based on an assumption of commonality of meanings between participants that fits with Mead’s (1934) perspective of commonality of experiences. Symbolic interactionism is a fitting theoretical perspective for grounded theory methodology. 56  Sensitizing concepts.  Blumer (1969) also introduced the idea of sensitizing concepts in his critique of sociological theory, which is consistent with grounded theory methodology12. He argued that the approach of arbitrarily hypothesizing/testing using abstract concepts was problematic because the concept definitions were predetermined and might not fit a particular research context. While he believed, as did Glaser and Strauss (1967), that researchers entered their research studies with prior knowledge of particular concepts (i.e., social norms, attitudes) he considered their definitions ‘fuzzy’ and open to change. In other words, he suggested that while researchers might be aware of particular concepts they should not search for them but wait to integrate them in a study when they were discussed by participants and seemed relevant. Thus, sensitizing concepts represent only ideas that a researcher has about factors that may influence the research area under study. Their use is consistent with Charmaz’s (2006) grounded theory approach (that I use to guide this study). She contended that researchers enter a research study with disciplinary sets of knowledge and awareness of different concepts that might influence the research area. She suggested that researchers identify sensitizing concepts to enhance theoretical sensitivity, defined as the ability of the researcher to draw abstract connections between codes to develop a theory about how participants resolve their main concern. Glaser (1978) also supported the use of sensitizing concepts to enhance theoretical sensitivity but he recommended they be used after the researcher had decided on the core category rather than at the outset of the study. Glaser’s (1992) position has been criticized as embracing a naïve perspective by suggesting that researchers can avoid being sensitized to concepts before they start their research (MacDonald & Schrieber, 2001). My review of the literature sensitized me to how personal, social, policy, and environmental factors   12	Concepts are socially constructed collections of ideas that in some cases are theory components (e.g. self- efficacy in relation to social cognitive theory) (Rodgers & Knafl, 2000). 57  might influence women’s decision-making. In chapter 4, I explicitly describe how these concepts sensitized me during my data collection and analysis. Navigating a Grounded Theory Methodological Course and Epistemological/Ontological Stance  Despite the fit between symbolic interactionism and grounded theory, there has been divergence in grounded theory protocols and the development of the methodology since its inception, which made it necessary for me to navigate a course for the procedures used and to articulate the epistemological/ontological perspective that underpinned my grounded theory approach. In the next section, I describe the divergence in grounded theory approaches and the shifting location of grounded theory and explain my epistemological/ontological position, the approach I used, and associated methodological/theoretical considerations. Divergence in grounded theory approaches.  A commonly discussed issue in grounded theory is the schism between Glaser and Strauss in the 1990s that centred on data analysis (Charmaz, 2006; Schreiber, 2001). Corbin and Strauss (1990a; 1990b) introduced new and complex coding methods that Glaser (1992) viewed as forcing grounded theory, and antithetical to grounded theory methodology. Specifically, they introduced the axial coding technique, another step in the coding process that followed the first step of open coding and emphasized deconstructing the data and testing hypotheses to verify categories and their relationships (Corbin & Strauss, 1990b). Axial coding emphasized how context and conditions changed relationships between categories (groups of related codes) and the consequences of participants’ actions (Corbin & Strauss; 1990a; 1990b). Corbin and Strauss also introduced a conditional matrix diagram that incorporated consideration of structural conditions, such as economics, cultural values, political trends, and history, to explain relationships between categories (Corbin & Strauss, 1990a; 1990b). 58  Researchers have mixed opinions on the value of Corbin and Strauss (1990a) coding methods because they have argued that coding has been demystified but that researchers may force the data (MacDonald, 2001; Charmaz, 2006). Glaser’s (1992) claim that Corbin and Strauss’s emphasis on data verification (checking ideas for their plausibility in the data), axial coding, and the conditional matrix actually compromised grounded theory development is convincing. Glaser (1992) insisted that constant comparative analysis, whereby the researcher constantly compared ideas (i.e., codes) in the data to each other, naturally allowed for them to be verified as opposed to forcing particular codes into the theory. Others have suggested that, if the researcher only used ideas from the data, there was no need to go back and test whether they were accurate (Walker & Myrick, 2006). Corbin and Strauss’s (1990b) more explicit concern with structure, contexts, conditions, and consequences of actions are important; however, such concerns might be addressed in grounded theory by using these as sensitizing concepts. Corbin and Strauss’s (1990a; 1990b) approach addresses some of the criticism about grounded theory’s astructural bias by forcing researchers to consider the complexity of behaviour and how structures might be influencing participants’ beliefs and behaviour. On the other hand, even those who agree that, in principle, structures should be explored have argued against using the conditional matrix and/or axial coding because they view those approaches as unwieldy (Charmaz, 2000; 2006; MacDonald, 2001). Glaser’s (1978) approach did not ignore analysis about how contexts and consequences affect behaviour; he explicitly encouraged such analyses when developing the core category and articulating its relationships with other categories. Rather than deciding to analyze the data via Corbin and Strauss’s approach, grounded theory researchers might effectively consider using social, economic, cultural, and environmental contexts as sensitizing concepts or ideas that they are attuned to listen for or analyze (Kushner & Morrow, 2003), which is the approach I used for this study. Since the major differences between Corbin and Strauss’s 59  and Glaser’s grounded theory approaches are in coding strategies I largely adhered to Glaser’s (1992) grounded theory approach (as well as that of Charmaz, 2006). The shifting location of grounded theory.  Despite authors’ depictions of grounded theory as strictly positivist (Bryant & Charmaz, 2007b; Charmaz, 2006; MacDonald, 2001), Glaser seemed to align more with a post-positivist epistemology. Post- positivists believe in the existence of reality but realize researchers’ methods are always flawed and truth is never obtainable (Lincoln & Guba, 2000). Using methods to enhance objectivity is important; however, post- positivists regard all scientific knowledge as representing flawed versions of truth (Lincoln & Guba, 2000). Glaser’s (1978) language about grounded theory’s role to uncover “what’s actually happening in the data” (p.57), and find a theory that “rings true,” coupled with his suggestion researchers can remove bias and get “closer to objectivity” (p.8) seemed to suggest Glaser believed in obtainable truth. Moreover, Glaser’s language around emergence and discovery of theory suggested he viewed a real theory as existing out there that was accessible (Bryant & Charmaz, 2007b). At the same time, Glaser (1978; 2012) has noted that a grounded theory is endlessly modifiable; it is not an accurate depiction of the substantive area per se or verifiable but it is a plausible depiction. Specifically, he has stated the theory is not intended as “proof;” categories are just grounded in the data “not proven, they are only suggested” (1978, p.135). Notwithstanding Glaser’s mixed messages and his epistemological position not fitting squarely with post-positivism, the suggestion Glaser was a strict positivist is misleading. It would be difficult to claim Glaser as a social constructionist. In a more recent article, Glaser (2012) explicitly argued against the social construction of knowledge, stating researchers have little impact on theory development; they simply ask neutral questions and allow the participants to speak. He suggested that any construction on the part of the theorist is simply bias that can be recognized and acknowledged in analysis. 60  Furthermore, he insisted that sampling and theory construction is driven by the data rather than the researcher’s biases and positions. Glaser’s strict position, which denies researchers’ influences on the research process, seems untenable. It is widely acknowledged in current social scientific literature that researchers affect the research process in ways that cannot completely be avoided, thus reflexivity, defined as the continual assessment of and reflection on the researcher’s influence during the research process, is a necessary component of undertaking qualitative research (Charmaz, 2006). Although grounded theory initially was aligned with a positivist/post-positivist stance, its application has broadened. In the last 20 years, authors have argued for the integration of grounded theory methodology with diverse theoretical and philosophical perspectives, such as feminism, critical theory, and postmodernism (Kushner & Morrow, 2003; MacDonald & Schrieber, 2001; Wuest & Merritt-Gray, 2001). Authors’ major criticisms, which parallel critiques about symbolic interactionism, suggested that Glaser’s grounded theory approach neither considered how power and structure influenced behaviour nor provided ethical guidance to conduct research. Wuest and Merritt-Gray (2001) argued that a feminist grounded theory approach allowed for a more ethically-derived and relevant theory because the feminist perspective encourages increased sensitivity to how factors, such as gender, social class, and age, affect behaviour and the researcher’s influence on the process of knowledge generation. Kushner and Morrow (2003) suggested that adding a critical theoretical perspective to feminist grounded theory helps researchers develop theories that can create meaningful social change because they will attend to less explicit behavioural influences (i.e., power, structure, and ideology). While these authors have thoughtfully articulated the limitations of grounded theory I argue that adherence to a pre-specified approach could be limiting. For example, in Wuest’s (2001) feminist grounded theory about women’s caring she directed scant attention to effects of women’s relationships with others on their beliefs about caring but emphasized effects of structures and resource availability. In her study rationale 61  she stated that she wanted to develop a theory to consider higher-level (i.e., policy) changes to improve women’s caring experiences. Her aim arguably skewed her interests towards understanding the types of programming, policy, or structural changes that could/should take place to improve women’s caring experiences to the detriment of understanding how immediate social context and relationships influenced women’s experiences. Preconceived lenses or foci of analysis may limit understanding about the many layers of influence that affect peoples’ decision-making. Charmaz (2000; 2006) has introduced a more generic open-ended approach to grounded theory analysis, which she has called constructivist grounded theory. The major thrust of this approach is that grounded theory methodology is amenable to change; she emphasized the co-construction of knowledge, the researcher’s influence on the research product, and ethical concerns about the conduct of inquiry. Thus, her approach has been flexible, addressed Glaser’s limitations, and avoided using predetermined lenses by which to guide analysis, such as power, oppression, emancipation, structure, and gender. My grounded theory location.  The epistemological location of grounded theory has bearing on how grounded theory is evaluated, conducted, and interpreted; given the variations in grounded theory application, it is important for researchers to be explicit about their approaches. Briefly, I regard my social constructionist view of knowledge construction as consistent with symbolic interactionism, and largely consistent with Charmaz’s (2006) epistemology/ontology, which I will describe briefly. Charmaz (2006) argued that all knowledge is perspectival and partial because knowledge is a product of co-creation between the researcher and the participants; objectivity is neither desirable nor attainable. She viewed the goal of the grounded theory researcher as trying to advance the most plausible and useful grounded 63  theory, not a true objective theory13. Charmaz has described how researchers and participants have values and particular perspectives that shape what knowledge can be generated and viewed knowledge as shaped by particular contexts and power relationships. She has claimed a moderate position about reality; viewing reality as socially constructed through interaction and between realism (true reality) and postmodernism (idealism or no reality) (Charmaz, 2006; Bryant & Charmaz, 2007b). While I generally agree with Charmaz’s (2006) epistemological/ontological position I consider her perspective closer to social constructionism than constructivism. Charmaz (2000; 2006) has claimed that she is a constructivist grounded theorist but she seems to have conflated the two philosophical perspectives of constructivism and social constructionism under the heading of constructivism. The difference between constructivism and constructionism is primarily about differences in views of reality; a constructivist believes in multiple realities while constructionists believe in a reality that is both relative and real (Crotty, 1998). That is, reality exists, but only as it is socially constructed. Social constructionists make an explicit claim that people share common understandings while constructivists claim knowledge is relative and local (Blumer 1969; Lincoln & Guba, 2000; Mead 1934). Charmaz (2006) has consistently stated that she disagrees with a relativist view of knowledge and has described constructivist grounded theory as the perspective whereby data and analyses are social constructions. Such claims are more consistent with constructionism than constructivism. My epistemological/ontological position is grounded in Crotty (1998), Blumer (1969), and Mead (1934). Crotty cited only three epistemological perspectives (objectivism, social constructionism, and subjectivism) and excluded constructivism14. For Crotty, social constructionism naturally aligned with the theoretical perspective  13	Which actually Glaser also supported (2004).  14	This is contrary to Lincoln & Guba (2000) who excluded constructionism, but included constructivism. Their position on constructivism is contradictory; they claimed reality is multiple, yet also asserted some knowledge is commonly shared between people; a social constructionist view avoids this issue because it assumes knowledge  64   of symbolic interactionism because it is based on the understanding of socially constructed meanings that emerge from interactions. He reserved the epistemology of subjectivism for postmodern research perspectives. My perspective incorporates elements of Charmaz’s (2006) description but I do not subscribe to multiple realities. Given my symbolic interactionist perspective, I view reality as linked to meanings developed in social interactions (Blumer, 1969; Crotty, 1998; Mead, 1934). My view of reality is bound to the commonly understood socially constructed language that exists. I consider that, through processes of social interaction, some knowledge is common between people (Crotty, 1998; Mead, 1934). I also view the assumption that knowledge is historically bound, partial, contextual, and value-laden as consistent with social constructionism. My knowledge is partial; by engaging with others, I can develop an understanding of their perspectives (Blumer, 1969). I agree; however, that it is only through socially constructed understandings we can communicate each other’s perspectives (Mead, 1934). My knowledge is value-laden; I hold particular beliefs and values that shape how I understand others and myself (Mead, 1934). It is also historically bound and contextual; knowledge I have about a particular situation emerges from the situation itself and my previous experiences (Mead, 1934; Blumer, 1969). Theoretical and methodological considerations for my proposed approach.  Although Charmaz’s (2000; 2006) articulation of her epistemological/ontological position could be refined, she has advanced grounded theory methodology by providing methods to analyze texts in grounded theory and articulating how a belief in the co-construction of knowledge shapes the grounded theory research process and interpretation. She also has increased the accessibility of Glaser’s grounded theory method by expanding explanations about coding processes. Charmaz (2006) has been influenced by Glaser because she   is shared, while also acknowledging people can have unique perspectives (Mead, 1934).65  follows his methodological steps with little variation but she interprets the resulting work differently than Glaser. Thus, I navigated both perspectives in the conduct and interpretation of this study. Both Charmaz (2006) and Glaser (1978) distinguished a grounded theory from general qualitative work because it is beyond thematic description. My findings must represent my interpretation of women’s decision-making about physical activity within the postpartum period organized as a set of integrated relationships around a core category that explains how women try to manage their main concern (Glaser, 1978). Charmaz argued that the term ‘grounded theory’ connoted a positivistic view whereby a theory predicts behaviour. Rather than view the grounded theory product as a predictive model I view the theory as a way to explain behaviour. A constructionist view of knowledge places me in a position where I consider my findings as having some applicability to women living in the postpartum period as a group. My view of the findings concurs with Blumer’s (1969) view about how to generate knowledge from a symbolic interactionist perspective; he suggested that the researcher explore the perspectives of those under study to develop understanding about their meanings and perspectives. Additional rigour and ethical considerations.  As suggested by Charmaz (2006), to produce a plausible theory about women’s decision-making, I needed to consider my influence on the research process, how the research process affects the knowledge generated, and how the participants contribute to the theory. I took a reflexive stance. Reflexivity acknowledges the researcher’s influence on the research process by recognizing knowledge as co-constructed and a combined product of the researcher’s beliefs, participants’ perspectives, and the research process (Mruck & Mey, 2007). Reflexivity needs to be considered at all stages of the research process (Mruck & Mey, 2007). Initially, it was important to reflect on my assumptions and rationale for conducting the study. As a mother who 66  exercises regularly, I assumed that, while it is difficult to obtain sufficient physical activity, it is possible if one values physical activity. I recognized that merely locating my assumptions and position did not remove my biases because my assumptions and knowledge still influenced who I sampled, the questions asked, my analysis, and how I presented my results (Mruck & Mey, 2007). Therefore, I constantly reflected on how my position as a researcher with a unique social location was influencing the research process (Harding, 1987). I view qualitative research as a relational process; hence, it was also imperative that I took steps to reduce power differentials between myself and the participant, and to enhance relationality and reciprocity. While reducing power and enhancing relationality and reciprocity are considered ethical principles associated with feminist research, I view them as a logical extension of my belief in the co-construction of knowledge and grounded in current research ethics board criteria (Hesse-Biber, 2006). Researchers can be in a position of considerable power during interviews and over participants’ information; therefore, it is important for them to be sensitive to the effect of the research process on participants and respectful of participants’ information and their willingness to provide it (Hesse-Biber, 2006). This includes finding balance between making space for participants to freely provide the information they want and the researcher asking questions relevant to the developing theory, providing participants with the opportunity to refuse to answer questions, and responding according to participants’ reactions during interviews (e.g., stop asking questions making participants uncomfortable) (Wuest & Merritt-Gray, 2001). It could also include following through with a planned interview, even when the researcher feels no further interviews are needed (Wuest & Merritt-Gray, 2001). The concept of relationality reflects that, in an interview context, the information a participant provides is connected to the relationship between the researcher and participant (Thayer-Bacon, 2009). Researchers can enhance trust with participants by authentically demonstrating caring and empathy and by responding openly and honestly to participants’ questions directed at them, which can facilitate more meaningful and honest responses from participants (Hall & Callery, 2001). Such honest and meaningful responses are important in the 67  context of my pragmatist perspective that suggests that research should generate knowledge that can be usefully applied for the benefit of participants (Rorty, 1999). Reciprocity reflects the view that a participant’s information should not be taken; rather, the information should be meaningfully used and something should be given back to the participant (Harrison, MacGibbon, & Morton, 2001). There is no assurance a participant will gain something meaningful or positive from participating in research; however, a researcher can take measures such as inviting the participant’s feedback, conducting more than one interview, involving the participants in the analysis, and offering a summary of the findings to facilitate reciprocity (Wuest & Merritt Gray, 2001). Chapter Summary  In this chapter I have described my theoretical perspective and methodological approach. My study was guided by symbolic interactionism, which directed me to an exploratory research approach and understanding meanings and perspectives of women around decision-making about physical activity during the postnatal period. While symbolic interactionism has been critiqued for its astructural bias and lack of attention to power I have argued a symbolic interactionist perspective addresses power relationships, albeit obliquely, by exploring individuals’ perspectives on constraints and facilitators in decision-making. Moreover, my sensitizing concepts that were drawn from physical activity literature took into account personal, social, policy, and environmental factors and sensitized me to structural influences. Beyond structural factors, I was sensitized to how emotions and expectancies and values across a range of decisions influence physical activity behaviour. My symbolic interactionist perspective aligns with social constructionism and my belief that knowledge is socially constructed and reality is both relative and real. While there are multiple approaches to grounded theory inquiry, I have argued against Corbin and Strauss’s (1990a; 1990b) overly complex methods that can force grounded theory and used Charmaz’s (2006) grounded theory approach, which largely aligns with my social constructionist stance. Charmaz’s approach also adheres to most of Glaser’s grounded theory 68  methodology but adds the notion of co-construction of knowledge. Incorporating Charmaz’s recommendations made it imperative for me to incorporate reflexivity into the research. In the next chapter I detail the methods used to conduct the research study, including: ethical considerations, approaches to sampling, data collecting, analysis, theoretical integration, rigour, and the study sample and limitations. 69  Chapter 4: Methods  Introduction  In this chapter, I state my research question, define terms, explain the study design and ethical considerations, and describe the procedures that I used for data collection, sampling, data analysis, theoretical integration, and rigour. Study Design and Research Questions  This study was qualitative, using grounded theory methodology. The aim of this study was to develop a theory that integrates relationships around a core category to explain women’s physical activity decision-making processes during the postnatal period. The research question was: What are women’s decision-making processes about physical activity in the postnatal period? To explore decision-making I used participant diaries, individual interviews, and document analysis. Grounded theorists presuppose that participants have a main concern in a phenomenon of interest that is the primary mover driving their actions (Glaser & Holton, 2004). A grounded theory helps explain decision-making processes by explaining the main concern and how participants’ try to resolve it through their actions (Charmaz, 2006). I used the data to develop a theory in the form of a core category, with related subcategories, that helps to explain how participants tried to resolve their main concern, which was to minimize discord between their physical activity desires and actual physical activity patterns. Definition of terms.  In the next section I describe how I defined research question terms, although the women themselves defined physical activity in various ways. 70  Physical activity: Any bodily movement involving energy expenditure above resting, which includes any activity listed on the Ainsworth compendium of activities and leisure time physical activity (Ainsworth et al., 2011; WHO, 2010). Postnatal: The period from 0-12 months post-birth.  Decision-making: A subjective assessment of the available options for behaviour and judgment about an appropriate line of behaviour (Reisberg, 2006). Decision-making involves both automatic and conscious cognitive processing (Reisberg, 2006). Postnatal decision-making incorporates daily and longer-term decisions in the context of the postnatal year. Processes: Lines or ways of thinking, in this case, about the engagement in physical activity (Reisberg, 2006).  Ethical Considerations  I received ethical approval for this study through the Behavioural Research Ethics Board (BREB) at UBC in May 2014 to recruit participants through flyers posted in public venues (e.g., coffee shops, libraries) and online. In October 2014, I received ethical approval for an addendum to edit the recruitment flyer (recruitment flyers: Appendix A and, revised, Appendix B) and expand the venues through which I could recruit participants to include magazines, newspapers, and centres with family programming (e.g., West Coast family centres). Most interested participants contacted me by email; in a few cases women phoned me. I responded to women’s emails by providing a brief synopsis of the study and the time involved and attaching the consent and information letters (Appendix C and D, respectively). I asked that they read the materials and contact me if they were still interested in participating in the study. For women who contacted me by phone, I discussed the study with them, sent all information via email, and asked them to re-contact me if still interested.  71  I met with the women twice at a place of their choosing. The initial meeting with participants occurred 3-7 days before the interview to review/sign the consent, explain the diaries (Appendix E), and give a coffee card as a thank you for participating. The women were informed they were not under obligation to complete the interview after our initial meeting. I interviewed participants at the second meeting and took steps to reduce power differences. Before the interview I aimed to increase the participant’s comfort by sharing my general background as a nurse and giving an overview of what to expect during the interview. I reminded the women that they could stop their interviews at any time, refuse to answer any questions, or change the subject. Participants were also reminded that they could request any information they shared during the interview to be omitted from their transcribed interviews and that they had the option not to withhold their diaries for analysis. I was attentive and responsive to the participants’ receptiveness and engagement in the interviews. For example, I ended one interview when a participant claimed she was comfortable continuing the interview after receiving multiple text messages from her family but her body language suggested increasing disengagement from my questions. The interview content about physical activity decision-making could raise participants’ personal concerns; in three interviews women became visibly upset when discussing frustrations about weight gain and limited finances. I validated and empathized with their concerns and debriefed participants after the interview; I also offered them links and phone numbers to connect with resources. I sent follow-up emails to these participants, thanking them for sharing their story and providing appropriate links to health services and phone numbers should they experience distress following the interview about the information discussed. Many of the participants wanted to debrief after the interview to understand how their story about postnatal physical activity ‘fit’ with others and whether their story was relevant. I encouraged the mothers by explaining how 72  many of the aspects they described were similar to other mothers’ depictions and how their particular story helped me further understand aspects of postnatal physical activity decision-making. All but two participants gave me their diaries following the interview (one had forgotten to bring the diary and one did not complete the diary). None of the women requested that anything be omitted from their transcripts. The women completed the demographic questionnaire following the interview (Appendix F). I gave the participants the option of receiving the study results. All participants asked for the study results and the chance to review the theory. They were all given a brief summary of the study findings in January 2016 and invited to comment (Appendix G). A transcriptionist completed 14 of the interview transcriptions while I completed the other 16. The transcriptionist signed a confidentiality agreement (Appendix H) and confidentiality was maintained by files that were exchanged with the transcriptionist being encrypted and password protected. Participants’ names were changed into numbers and identifying information (such as names, cities, or locations) was altered on the transcripts. Data were stored on my personal password protected computer and the paper copies of the physical activity diaries and demographic data, along with the digital recorders were kept in a locked research cabinet in my home office (when not in use). The data will be kept for up to five years for the possibility of a secondary analysis in a secured office at UBC; participants were informed about arrangements for data storage at the time of consent. Sampling  In the section on sampling I describe inclusion and exclusion criteria, recruitment, and sampling approaches. 73  Inclusion and exclusion criteria.  Healthy mothers between 2.5-12 months postpartum15 were eligible to participate, regardless of parity (i.e., the number of children they have) and mode of delivery (cesarean, assisted, vaginal). Because some conditions could be associated with different decision-making processes about physical activity than those of the general population I intended to exclude mothers with chronic conditions that might affect their daily physical activity decision-making (e.g., postnatal depression, multiple sclerosis, paraplegia, cystic fibrosis, type 1 diabetes) and mothers with severe birth complications requiring extended hospital length of stay, (e.g., a hysterectomy or sepsis). None of the participants who contacted me had been diagnosed with these types of conditions or experienced severe birth complications. Two participants had been treated for depression in the past, either in the postpartum period of one of their older children (n=1), or generally (n=1), and one participant indicated receiving ongoing treatment for chronic depression but these women indicated they did not have diagnoses of postpartum depression at the time of the interview. While I intended to exclude participants with infants who were still in hospital due to complications, those whose infants had severe conditions (e.g., congenital heart defect), or those whose infants had developmental delays none of the participants who contacted me had infants with those conditions. I received an ethical amendment to include women with infants who were born aged 34-36 weeks and who were healthy at the time of interview. I also included women who had a healthy birth of multiples (e.g., twins) and those experiencing common postnatal concerns (e.g., breastfeeding issues). Approaches to sampling.  Grounded theory sampling begins by purposefully sampling participants believed to have relevant information to the field under study (Adolph, Hall, & Kruchten, 2011); thus, I started by sampling  15 To account for the fact that women having a cesarean will not be significantly physically active until at least 6 weeks post-birth. 74  women within the postnatal period who met the recruitment criteria. At the end of each interview, I used snowball sampling (Polit & Beck, 2008) by asking participants to pass on information about the study to other women they knew who might be interested. Two women participated after learning about the study from another participant. There were two rounds of recruitment: September-December 2014 and July-September 2015. I recruited participants from the Lower Mainland, with the majority of participants recruited from the Greater Valley Regional District Tri-Cities area (Coquitlam, Port Coquitlam, and Port Moody). In 2014 and 2015 I posted flyers in all of the coffee shops and libraries within the Tri-Cities, as well as in three physician offices and five Churches (either online or on the community board) in these areas. Online, my flyers were posted on Pre/Postnatal fitness websites and Urban Baby Magazine Facebook®	pages. In print, my flyer was posted in the Urban Baby Magazine.  As I coded, analyzed, and developed categories I began theoretical sampling wherein I sampled women based on my perceptions of their contribution of incidents to the developing theory (Glaser, 1978; 2004). Participants are sampled theoretically to obtain knowledge to help develop the theory; specifically, they provide incidents that might help refine categories and theoretical relationships (Adolph, Hall, & Kruchten, 2011; Charmaz, 2006). For example, after my first round of recruitment, I found none of participants had used childcare at recreation or fitness centres and only one had allowed other friends and family to provide childcare so that she could engage in physical activity without her baby. I wondered whether women who used childcare held different perspectives about motherhood and physical activity. My awareness that some women used fitness centre childcare or received childcare from others to attend fitness centre programming assisted my theoretical sampling because I re-posted my flyer at fitness centres and online Facebook®	pages of fitness centres where childcare was offered to locate women with potentially different perspectives. After posting these 75  flyers I recruited seven participants who either used fitness centre childcare or received childcare from family and friends to engage in activity without their babies. The use of theoretical sampling helped develop my theoretical coding; for example, the women I recruited who used childcare so that they could engage in physical activity without their children regarded physical activity as central to their lives and I was able to learn more about the perspectives of women indicating higher physical activity centrality. Previously, I had primarily interviewed participants who indicated that they regarded physical activity as more peripheral to their lives. Data Collection Sources and Strategies  Glaser (1978) indicated that grounded theory can be informed through multiple sources of data. Interviews are a primary data collection strategy in grounded theory, although other data sources are also important, such as observation, field notes, and documents (Glaser, 1978). Glaser (2004) actually considered all as data, meaning anything could be considered data if it helped develop the theory. Therefore, data can be derived from many sources, such as media, online documents, or diaries. Charmaz (2006) distinguished between extant data (i.e., data obtained from sources the participant was not involved in making, such as a magazine article or policy document) and elicited data (i.e., data obtained from participants in the form of surveys, diaries, or interview data). Using multiple forms of elicited data can enhance the researcher’s grasp of participants’ beliefs and understandings, while analyzing extant data can help sensitize researchers to how structural conditions, power relationships, and sociocultural media messages might affect participants’ beliefs (Charmaz, 2006). In this study, I used both elicited data (diaries and interviews) and extant data (textual documents). Participant diaries.  Participant diary data can be considered a form of elicited data to help develop the theory (Charmaz, 2006). Before each interview, I gave each woman a three-day diary. I created a physical activity diary for them 76  to fill in on an hourly basis to describe their general activity. Because I was particularly interested in the contexts for and complexity of decision-making I also provided a ‘context’ column for participants to add their thoughts, emotions, or environmental considerations (e.g., noise levels) beside the activity column. Along with the blank diary pages, I included a face sheet with a diary template and mock diary entries to provide the participants with a general idea about how they could choose to fill in their diaries. However, I did not provide specifications about what I wanted the participants to include in the context columns, which was consistent with my goal to understand participant’s activities in relation to contextual factors that they believed were salient for them. I explained to participants that they could fill in as much or as little of the diary as they wanted and were able to provide. I emphasized that the diaries were not just for listing physical activity and they could include all forms of daily activity. I used the diaries to start our conversations in the interviews. They were valuable for gaining a deeper appreciation of the contexts of the participants’ lives. I typically said little as the women talked about the context of their diaries; most women either talked about some of the thoughts they had when they were completing their diaries or described each day, for a few minutes to up to 20 minutes during the interviews. Most took 7-8 minutes to describe their diaries. Having the participants write about their activities while describing concurrent contexts was helpful because context comments prompted women to elaborate (e.g., in the activity column, a woman would write ‘prepared for birthday party’, but in the context column, she might write, ‘was very tired and frustrated about having people over’). Women also talked about their contexts in terms of difficulties they had with social support and having to ‘perform’ perfect motherhood. 77  Our discussion of the diaries identified concerns for the mothers that I could explore later in the conversation. For example, one participant repeatedly talked about her lack of sleep when talking about her diary entries. When I explored this further, she described how sleep-seeking throughout the day shaped her overall daily structure. Having the participant highlight the importance of sleep when describing her diary helped my understanding about the effects of her efforts to achieve sleep on her decision-making around physical activity. The three-day diary facilitated conversations about participants’ typical physical activity patterns and whether they felt the diary was a fair representation of their activities. Reviewing the diaries helped me understand what the women perceived as physical activity because they discussed their physical activity in relation to instances in their diaries. By highlighting instances of physical activity the diaries also facilitated conversations with the women about their physical activity decision-making processes and experiences with physical activity (e.g., what women felt contributed to their ability to be physically active in that instance, as well as their emotions, and satisfaction associated with the activity). Interviews.  Because discussion around physical activity could also include discussion about sensitive issues, such as diet and weight, I interviewed participants individually (Gray, Williamson, Karp, & Dalphin, 2007). Interviews ranged in length from 45 to 90 minutes. I conducted many interviews (n = 10) outside the participants’ homes, e.g., at coffee shops (n = 7), a park (n = 1), and libraries (n = 2), while 20 were conducted at their homes. At most interviews (n = 26), the participant’s children were present and, for some, children’s siblings (n = 3) or women’s partners (n = 5) were also present or in the home. Norlyk, Haahr and Hall (2015) highlighted possible ethical and methodological concerns with conducting interviews when partners are present, since participants may defer to or tailor their responses in the presence of their partner. However, these concerns at the time of my interviews were of lower risk as three of the five partners were not in the same room as the participants and were 78  occupied doing other tasks (e.g., child care and work). The two partners who came into the room and participated in the conversation did so only briefly (e.g., <5 minutes). The partners’ contributions to the interview might have enhanced the data collected as one partner elaborated on the participant's discussion about lack of childcare, which prompted further discussion about environmental resistance. The other partner engaged in a short conversation with the participant about social norms of motherhood and then left to care for the couple’s infant. Interviews were audiotaped with a digital recorder and transcribed. Fourteen of the transcripts were prepared professionally through a transcription company; I transcribed the remaining 16 interviews. I checked the audiotapes against the transcription company’s transcripts and corrected them as necessary. I followed a naturalized transcription approach, incorporating information about: paraverbal language (i.e., intonation, pacing – e.g., silences and pauses), errors of speech, ‘ums, ahs’, and information about background noise and context (Davidson, 2009). The hired transcriptionist was also instructed to transcribe as the participants spoke (i.e., including ‘ums’, ‘ahhs’, pauses, and silences) and context (i.e., the participant was interrupted by infant, information about background noise). When I conducted interviews with children present it influenced the interview flow and duration.  Interviews were stopped at least once and some up to twelve times when mothers had to care for their infants or their siblings; most of the time there were 2-3 stoppages during an interview. I used mothers’ engagement in the questioning and body language (e.g., frustrated, tired, and having difficulty thinking) and my questions about whether mothers were able to continue as children’s needs became more explicit to suggest interview closure. That evidence helped me to determine the appropriateness of continuing interview questions. Due to mothers’ needs to focus on my questions and attend to their child(ren) the interviews mostly ended between 55 minutes and 1 hour, at which point the mothers were often either breastfeeding or walking their babies as they talked, (i.e., using some measure to soothe their babies). Interviews were generally fairly relaxed with women indicating that they were fine to continue when asked about their comfort levels. The participants were also 79  flexible about starting and stopping the interviews as needed. Mothers I interviewed without their children (n=4) indicated they had a specific amount of time that they could commit to the interview, usually one hour to one and a half hours; these interviews were actually less relaxing in nature because we had to keep track of the time and the mothers often received text messages or phone calls during the interviews. Because the goal of grounded theory is to understand the participants’ main concern and how they resolved it (Glaser & Holton, 2004) I was particularly attuned to participants’ discussions about their concerns around physical activity, how they managed them, and how their behaviours and perceptions changed over time. In particular, I listened for language indicating possible conflict or challenging situations, such as ‘frustrating, but, should be, difficult,’ goal orientated language, such as ‘I want,’ and positive language, such as ‘opportunity and role modeling’ to develop understanding about the women’s main concerns and decision-making. I also listened for language that indicated change over time, such as ‘before, but now, used to, different.’ I was sensitive to women’s perceptions of their feelings in regard to physical-activity decision-making. Importantly, as guided by my symbolic interactionist perspective, I listened for how participants’ created lines of action in relation to their understandings about their situations (Mead, 1934). My commitment to co-construction of knowledge heightened my awareness about interview context, paraverbal behaviour (such as voice inflection and volume), and non-verbal behaviour (such as hesitancies and silences) (DeVault & Gross, 2007; Hesse-Biber, 2006). Although I generated interview questions in the form of an interview guide for the purposes of highlighting my thinking about possible areas of inquiry (Appendix I) I kept the interviews broad and aimed at understanding the participants’ experiences of physical activity and the meanings they attached to them. Despite bringing the interview guide, I typically did not look at or use it during my interviews. The open-ended nature of my interviews aligned with both Glaser and Holton’s (2004) and Charmaz’s (2006) recommendations that the researcher should simply let the participants explain their experiences in the beginning. 80  Having participants initiate the interview by describing their diaries created opportunities for them to lead conversations and discuss their daily decision-making contexts and decrease the power differences between us. Participants’ descriptions of their diaries, in addition to my sensitizing concepts, helped guide interview flow and questions. After participants had described their diaries I asked them about their thoughts regarding physical activity when writing the diary and their general physical activity perceptions. I also asked general questions about how they defined physical activity; physical activity changes since having their children; their ideal physical activity patterns; and what they felt affected their engagement in physical activity. To increase my depth of understanding about participant’s perspectives, I asked participants to elaborate, (e.g., ‘I’m wondering if you could tell me more about why you think it would be weird to let your friends watch your daughter?’). I noticed that, after a few interviews, from reading the transcripts and my reflexive journaling I was moving too quickly between ideas and missing opportunities to summarize and check-in with what I thought I was hearing, (e.g., ‘I’m hearing that it sounds as though sleep was really important for you, but that it has improved recently?’). I started to incorporate these types of summary statements during my interviews, which usually led to more in-depth conversation about the women’s perspectives and decision-making. I developed several interview guides in response to shifts in my foci during the interviews. I developed two iterations of the interview guides (see Appendix J). These foci evolved as I undertook my analysis and developed the concepts. As I conducted more interviews I sought to elaborate and further develop categories. I would sometimes preface my questions with other women’s experiences that indicated specific categories I was interested in developing. I would ask whether these perspectives fit for the individual I was currently interviewing (e.g., ‘Some women have talked about a process of increasing physical fitness, does that