UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Emotion regulation and pain : a daily process study of individuals with rheumatoid arthritis Welsted, Alison Carol. 2007

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2007-0637.pdf [ 1.88MB ]
Metadata
JSON: 831-1.0100967.json
JSON-LD: 831-1.0100967-ld.json
RDF/XML (Pretty): 831-1.0100967-rdf.xml
RDF/JSON: 831-1.0100967-rdf.json
Turtle: 831-1.0100967-turtle.txt
N-Triples: 831-1.0100967-rdf-ntriples.txt
Original Record: 831-1.0100967-source.json
Full Text
831-1.0100967-fulltext.txt
Citation
831-1.0100967.ris

Full Text

E M O T I O N R E G U L A T I O N A N D P A I N : A D A I L Y P R O C E S S S T U D Y O F I N D I V I D U A L S W I T H R H E U M A T O I D A R T H R I T I S b y A L I S O N C A R O L W E L S T E D A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A R T S i n The Facu l ty o f Graduate Studies (Psychology) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A Augus t 2007 ©Alison Carol Welsted, 2007 11 Abstract Whereas the l i n k between emotions and pa in is wel l -establ ished (e.g., C r a i g , 1999), the impact o f managing emotions is uncertain. T h e current study used a da i l y process methodology to investigate the prospect ive relationship between da i ly emot ion regulat ion and evening pa in intensity. One hundred and seventeen indiv iduals w i t h rheumatoid arthritis took part i n an in i t ia l background interview, fo l l owed b y seven days o f twice -da i ly telephone interviews, dur ing w h i c h participants were asked about emotions and pain . Results o f hierarchical l inear mode l i ng indicated that higher levels o f da i ly emot ion regulat ion were associated w i t h s ignif icant ly less evening pain . Further, main ta in ing or containing, as w e l l as recover ing (from), emotions were each independently associated w i t h s ignif icant ly less evening pa in . M o r e o v e r , the current study addressed gaps i n the literature b y establishing these associations above and beyond the influence o f baseline pa in intensity and baseline morn ing emotions, as w e l l as o n a subsample o f days when higher than average, morn ing pa in indicated the presence o f undesirable emotions to be regulated. T h i s prospective relationship between emot ion regulat ion and pa in intensity was largely dr iven b y regulat ing depressive emotion, whereas, i n previous research, regulating anxious emot ion was found to be more influential (Pauqet, Kergoa t , & D u b e , 2005). Th i s novel f ind ing underscores the importance o f the addi t ional controls appl ied i n the current study. In sum, these f indings h ighl ight the importance o f invest igat ing the impact o f dynamic emot ion regulat ion on the day-to-day pa in experiences o f ind iv idua ls w i t h chronic pa in and suggest potent ia l ly fruitful directions for future pa in management interventions. i i i Table of Contents Abstract i i Table of Contents i i i List of Tables iv Acknowledgements v Introduction 1 Method 3 Participants 3 Procedure 4 Background Interview Measures 5 Daily Interview Measures 5 Results 9 Overview 9 Daily Interview Completion and Descriptive Statistics 9 Bivariate Analyses 12 Multi-level Regression Analyses 12 Conclusions 28 References : 33 Appendix 38 List of Tables T a b l e 1. M e a n s and Standard Devia t ions o f D a i l y Interview Measures 10 T a b l e 2. E m o t i o n Regula t ion Frequencies 11 T a b l e 3. Assoc ia t ions o f G l o b a l E m o t i o n Regula t ion w i t h E v e n i n g P a i n Intensity 16 T a b l e 4. Assoc ia t ions o f M a i n t a i n i n g or Conta in ing , and R e c o v e r i n g , w i t h E v e n i n g P a i n Intensity 18 T a b l e 5. Assoc ia t ions o f Regula t ing A n x i o u s , Depressive, and Pos i t ive E m o t i o n w i t h E v e n i n g P a i n Intensity 20 T a b l e 6. Assoc ia t ions o f Con ta in ing or M a i n t a i n i n g , and R e c o v e r i n g (from) each o f the Emot ions w i t h E v e n i n g P a i n Intensity 22 T a b l e 7. E m o t i o n Regula t ion Frequencies: H i g h M o r n i n g P a i n D a y s ( N = 380) 25 Acknowledgements M y sincere thanks go to D r . A n i t a D e L o n g i s whose guidance and vast knowledge were invaluable to the successful comple t ion o f this work . I a m also grateful to D r . Susan H o l t z m a n and E l i Puterman for their intellectual contributions and mora l support. M y thanks also go to D r s . Jessica T r a c y and W o l f g a n g L i n d e n for their insightful comments . I a m especia l ly grateful to m y husband, D r . Geoff rey Dunbar , w h o has inspi red me to pursue m y academic interests, as w e l l as to m y son, Gareth Dunbar , whose need to laugh and p lay has helped me mainta in a balanced perspective. 1 Introduction The inextr icable l i n k between m o o d and pa in is a wel l -es tabl ished phenomenon i n the study o f pa in (Cra ig , 1999; R o b i n s o n & R i l e y , 1999). Results demonstrat ing this l i n k have been obtained over both months a n d years (e.g., Zautra et a l . , 1995) as w e l l as w i t h i n and across days (e.g., Fe ldman , D o w n e y , & Schaffer-Nei tz , 1999; N e w t h & D e L o n g i s , 2004). Recent ly , however , there has been a ca l l i n the literature to examine whether the abi l i ty to manage (or regulate) moods and emotions also influences the pa in experience (Keefe , L u m l e y , Ande r son , L y n c h , & Carson , 2001). Results o f research invest igat ing this question among indiv iduals w i t h a chronic pa in condi t ion have been equivoca l (e.g., H a m i l t o n , Zautra , & R e i c h , 2005, 2007; van M i d d e n d o r p , Geenen, Sorb i , van Doornen , & B i j l s m a , 2005; van Middendorp , Geenen, Sorb i , H o x , et a l . , 2005). A l l o f this previous w o r k used one-time, self-report measures o f emot ion regulation, al though, both research and theory suggest that emotions fluctuate cons iderably w i t h i n and across days (Stone & Shif fmah, 1994; Zautra, A f f l e c k , Tennen, R e i c h , & D a v i s , 2005) and that emot ion regulat ion is a dynamic process (Gross, 1998). Repeated and frequent assessments o f emot ion regulat ion are, therefore, best able to capture this process. In addi t ion, emot ion regulat ion is consistently defined as occurr ing both consc ious ly and subconsc ious ly ( D i a m o n d & A s p i n w a l l , 2003) 1 , thereby necessitating that emot ion regulat ion not be assessed o n l y v i a direct self-report. A methodology that addresses these l imitat ions is avai lable, but has yet to be used to investigate the impact o f emot ion regulation on pa in among ind iv idua l s w i t h a chronic pa in condi t ion (Carstensen, Pasupathi , M a y r , & Nesselroade, 2000). U s i n g this methodology, emotion is assessed at mul t ip le points both w i t h i n and across days, and emot ion regulat ion is ' Recent research suggests that unconscious emotion regulation does indeed exist and that, at least as far as anger regulation is concerned, it has beneficial consequences (Mauss, Cook, & Gross, in press), unlike deliberate anger regulation which is associated with considerable psychological and physical costs (e.g., Gross, 2002; Suarez, 2006). operat ional ized as main ta in ing or recover ing desirable emot ional states over the course o f short t ime periods (i.e., less that one day). O n l y one study that w e k n o w o f has used this methodology to investigate the impact o f emot ion regulat ion on pain , albeit among elder ly ind iv idua ls w h o d i d not have a chronic pa in condi t ion (Paquet, Kergoat , & Dube , 2005). Th i s study found that emot ion regulat ion was s ignif icant ly associated w i t h subsequent pain , al though this w o r k was not wi thout l imitat ions. Firs t , g iven the strong associations between pa in at the beg inn ing o f a short t ime per iod and pa in at the end o f it, as w e l l as the k n o w n associations between emotions and pa in , it is important to control for these factors i n order to establish the independent influence o f emot ion regulation. Moreove r , m u c h o f the emot ion regulat ion literature takes an evolut ionary perspective, according to w h i c h undesirable emotions and potential emot ion regulat ion ensue f o l l o w i n g emotional cues (e.g., Gross , 1998). It is , therefore, important to determine whether the benef ic ia l impact o f emot ion regulat ion on pa in also operates when an emotional cue is present. The current study implemented these controls and examined the prospect ive influence o f emot ion regulat ion on pa in among a sample o f ind iv iduals w i t h rheumatoid arthritis ( R A ) . It was expected that emot ion regulat ion over the course o f a day w o u l d be associated w i t h s ignif icant ly less pa in at the end o f it. 3 Method Participants A list o f potential study participants was r andomly selected f rom a database o f R A patients registered w i t h the M a r y P a c k Ar thr i t i s Society, a l oca l organiza t ion that offers treatment and education to arthritis patients throughout the p rov ince o f B r i t i s h C o l u m b i a , Canada. Individuals w h o were over the age o f 18, l i v i n g outside the L o w e r M a i n l a n d o f B r i t i s h C o l u m b i a , and w h o had sought med ica l treatment for arthritis w i t h i n the past two years, were ma i l ed an in i t i a l contact letter descr ibing the study and requesting their par t ic ipat ion. Interested indiv iduals were screened over the telephone to ensure that they had been diagnosed w i t h R A , had experienced pa in due to R A dur ing the past month , and were able to read, wri te , and speak E n g l i s h . Individuals w h o contacted our research office regarding par t ic ipat ion i n the study were entered i n a draw for C $ 1 0 0 0 . 2 In addi t ion, upon comple t ion o f the in i t i a l background interview, a l l participants were m a i l e d a smal l gift va lued at C $ 1 0 . Ini t ia l contact letters were ma i l ed to 800 indiv iduals w i t h arthritis. O f the 188 patients diagnosed w i t h R A w h o expressed interest i n the study, 28 dec l ined to participate pr ior to addi t ional e l i g ib i l i t y screening. 3 O f the 160 w h o agreed to participate i n a telephone screening interview, 120 (75%) met the inc lus ion cri teria and completed both the background and da i ly interviews. Three o f these 120 participants were dropped f rom the f inal analyses due to l o w compl iance w i t h da i ly interviews (i.e., less than 5 0 % o f da i l y in terviews completed) . O f the 20 respondents w h o were screened and d i d not meet cri teria for the current study, 17 were excluded because they had not experienced R A pa in i n the past mon th and 3 were exc luded because they 2 The initial draw was for CS300. Due to low response rate, the draw value was increased to C$1000 approximately half way through recruitment. 3 Over the course of the recruitment phase of the study, 85 initial contact letters were returned due to an incorrect mailing address. We were also informed that six of the individuals on our mailing list were deceased, and that 14 had never been diagnosed with RA. We have no way of knowing why the remainder of patients did not respond (i.e., did not receive our letter, were not diagnosed with RA, were deceased, or were not interested in participating). Twenty-eight individuals declined to participate prior to additional eligibility screening due to: having been in too many studies (n = 1), being too ill (n = 3), being too busy (n = 9), or for an unspecified reason (n = 15). had a med ica l condi t ion other than R A . O f the 20 respondents w h o met the i nc lu s ion criteria but dec l ined to participate, two indicated that they were too s ick, two said that they were too busy or that it was a bad t ime, and the remain ing 16 d i d not provide a reason. The f inal sample consisted o f 117 indiv iduals diagnosed w i t h R A . Participants were predominant ly female (88%) and Caucas ian (83%), ranging i n age f rom 26 to 84 years (M- 59, SD = 13.04). 4 The number o f years since be ing diagnosed w i t h R A ranged f rom 1 to 54 (M= 17, SD = 13.30) and R A - r e l a t e d pa in intensity ranged from 0 to 9 (M= 4 .82, SD = 1.99) out o f 10 i n the week pr io r to the in i t i a l interview. The majori ty o f participants had completed at least a h igh school education (80%). Twen ty - s ix percent o f participants were emp loyed at the t ime o f the study and the moda l f a m i l y income was between C $ 2 5 000 and C $ 4 9 999. P r o c e d u r e U p o n comple t ion o f the screening protocol , e l ig ib le participants took part i n a structured background in terview, w h i c h was conducted over the telephone and lasted approximate ly 30 minutes. B r i e f structured interviews were then administered twice da i l y for one week again, over the telephone. D a i l y interviews took place approximate ly s ix and twelve hours after participants w o k e up i n the morn ing and lasted approximately 10 minutes each. T h e purpose o f twice da i ly interviews was to a l l o w for the examinat ion o f fluctuations among study variables w i t h i n participants over the course o f each day. A t the beg inn ing o f each in terv iew, participants were asked to f ind a place where they c o u l d talk pr ivately. A l l in terviews were conducted b y trained female undergraduate and graduate research assistants. Participants were assigned the same interviewer for both the in i t i a l and the da i ly interviews. Interview sessions were tape recorded w i t h the permiss ion o f participants. Th i s a l l owed for later t ranscript ion o f open ended questions 4 The higher percentage of women than men in the current study is consistent with the greater prevalence of RA among women (i.e., the overall ratio is 3:1; Anderson, Bradley, Young, McDaniel, & Wise, 1985). 5 b y interviewers as w e l l as supervis ion o f interviewers w i t h regard to the standardized protocol . Participants also completed a b r i e f m a i l - i n questionnaire. Da ta for the current study were drawn f rom a larger prospect ive study o f psychosoc ia l factors among ind iv idua l s w i t h rheumatoid arthritis and their spouses. O n l y those measures used i n the current study w i l l be discussed here. Background Interview Measures Demographics and Disease Status. Participants were asked to prov ide demographic (e.g., gender, age) and disease status information (e.g., years since diagnosis , fatigue, functional disabi l i ty) . T w o k e y dimensions o f RA-relatedfatigue ( t iming and intensity) were assessed (Be l za , 1995). Patients were asked to indicate h o w often they had exper ienced fatigue due to their R A dur ing the past week, us ing a scale f rom 1 (never) to 5 (all the time). Intensity o f fatigue was assessed b y ask ing patients to rate h o w fatigued they were dur ing the past week on a scale f rom 0 (no fatigue) to 10 (fatigue as bad as it could be). These two measures were then combined to prov ide an overa l l fatigue rating w i t h a significant inter- i tem correlat ion o f .67. Functional disability was operat ional ized as diff icult ies per forming eight da i l y activit ies (e.g., dressing oneself, getting i n and out o f bed, wa lk ing ) on a scale f rom 1 (without any difficulty) to 4 (unable to do). These items were drawn f rom a mod i f i ed vers ion o f the Hea l th Assessment Questionnaire ( M H A Q ; P incus , Summey , Sorac i , W a l l s t o n , & H u m m o n , 1983) that is frequently used i n the assessment o f functional d isabi l i ty among patients w i t h rheumatic diseases. The M H A Q has demonstrated good re l iab i l i ty and va l id i t y i n past research (e.g., P incus , et al.) and the inter-i tem correlat ion was good i n the current study (a = .88). Daily Interview Measures D u r i n g the first in te rv iew.of the day, participants were asked to reflect o n their experiences so far that day. D u r i n g the second interview o f the day, participants were asked to reflect on their experiences since the last time they had spoken with the interviewer. M o s t o f the 6 measures used i n the da i ly interviews were brief, mod i f i ed versions o f the or ig ina l scales. Th i s is c o m m o n practice among da i ly process studies, as it reduces the otherwise p roh ib i t ive burden p laced on participants and increases the number o f constructs researchers can assess w i t h i n a single study protocol (Bolger , D a v i s , & Rafae l i , 2003; Stone & Nea le , 1984; Tennen, A f f l e c k , Coyne , Larsen , & D e L o n g i s , 2006). P a i n Intensity. Participants indicated the intensity o f their R A - r e l a t e d pa in on a numerica l rat ing scale ( N R S ) ranging f rom 0 (no pain) to 10 (pain as bad as it could be). The N R S has demonstrated good va l id i t y i n previous research, d i sp lay ing pos i t ive and signif icant associations w i t h other measures o f pa in intensity (Jensen, K a r o l y , & Braver , 1986; Pa ice & Cohen , 1997). Emot ions . Pos i t ive emot ion was measured w i t h the contentment subscale, and negative emotions were measured w i t h the depression and anxiety subscales, o f the Derogatis Affects Ba lance Scale ( D A B S : Derogat is , 1975). Participants were asked to indicate the extent to w h i c h each o f a set o f adjectives (e.g., hopeless, afraid, relaxed) described h o w they felt "so far today/since w e last spoke" on a 5-point L i k e r t scale (0 = does not apply, 1 = not at all, 2 = a little, 3 = somewhat, 4 = a lot). F o r a l l analyses, " 0 " and " 1 " responses were col lapsed into a single category. T h e D A B S has been found to have good internal consis tency i n previous research (e.g., Nor thouse & S w a i n , 1987) and, i n this sample, the internal consis tency (a) o f each o f the three subscales used was .94. E m o t i o n Regulat ion. Consistent w i t h previous research (Carstensen et a l . , 2000; Paquet et a l . , 2005) , emot ion regulat ion was operat ional ized as main ta in ing or recover ing desirable emot ional states over the course o f specific t ime periods. Des i rab le emot ional states were empi r i ca l ly defined as more intense posi t ive emot ion or less intense negative emotions than the average intensity reported b y each ind iv idua l across the seven days o f assessment. In order to obtain the emot ion regulat ion variables, first, each ind iv idua l ' s mean o f posi t ive , anxious, and depressive emotions across the seven day sampl ing per iod, was calculated, separately for 7 m o r n i n g assessments and for evening assessments. Then , for each sampl ing occas ion, posi t ive emot ion was class i f ied and d u m m y coded as either above the i n d i v i d u a l ' s mean (1) or at or be low it (0). A l s o , for each sampl ing occasion, anxious and depressive emot ion were each d u m m y coded 1 i f they were b e l o w the ind iv idua l ' s mean and 0 i f they were at the mean or above. Maintaining positive emotion was operat ional ized as pos i t ive emot ion that was above the part icipant 's mean on both measurement occasions i n the same day. Recovering positive emotion was operat ional ized as posi t ive emot ion that was at or be low the i n d i v i d u a l ' s mean on the first sampl ing occas ion o f the day and above the mean i n the evening. Containing anxious or depressive emotion was evidenced i f either o f these were b e l o w the part ic ipant 's mean on both sampl ing occasions w i t h i n the same day. Recovering from anxious or depressive emotion was assigned i f either anxious or depressive emot ion was at or above the i n d i v i d u a l ' s mean o n the first sampl ing occas ion and be low the mean b y the second sampl ing occas ion . The emot ion regulat ion types were d u m m y coded 1 i f they were determined to have taken place over the course o f a day and 0 i f they had not. In addi t ion, s ix summary scores were calculated. Global emotion regulation ranged f rom 0 i f none o f anxious, depressive, or posi t ive emot ion had been regulated o n a g iven day to 3 i f a l l three emotions had been regulated. Maintaining or containing anxious , depressive, and posi t ive emot ion ranged f rom 0 i f none had occurred to 3 i f a l l had occurred. Recovering from anxious and depressive emot ion and recovering posi t ive emot ion ranged from 0 to 3, depending on h o w many emotions recovery was associated w i t h on a g iven day. Regulating anxious, regulating depressive, and regulating positive emotion were each assigned a score o f 0 i f neither main ta in ing or containing, nor recovering, had occurred o n a g iven day and 1 i f either had occurred. These summative scores permit ted evaluat ion o f the influence o f emot ion regulat ion i n 8 general and for each emot ion , as w e l l as evaluat ion o f conta ining or main ta in ing and recover ing generally, and for each emotion. One advantage o f this measure o f emot ion regulat ion is that it avoids r e ly ing on ind iv idua l s ' self-reports o f their emot ion regulat ion sk i l l s , thereby captur ing not o n l y changes i n emot ion that participants c o u l d consc ious ly te l l us about, but also automatic emot ion regulation that m a y be unavai lable for self-report. A l s o , b y def ining emot ion regulat ion relative to the ind iv idua l ' s o w n mean o f each emot ion i n the morn ing and i n the evening, o n l y changes i n emot ion above and beyond the ind iv idua l ' s normal da i ly changes i n emot ional state are considered. Results Overview In this section, descriptive statistics are presented, f o l l o w e d b y bivariate and mul t i - l eve l regression analyses. The questions addressed us ing mul t i - l eve l analyses were: 1) do the number o f emotions regulated on any g iven day p lay a role i n evening pa in intensity, 2) do each o f main ta in ing or containing, and recover ing, have independent impacts on evening pa in intensity, 3) does regulat ing each o f anxious, depressive, and posi t ive emot ion p l ay an independent role i n evening pa in intensity, and 4) are containing or main ta in ing and recover ing (from) each o f the emotions s ignif icant ly and independently associated w i t h evening pa in intensity? F i n a l l y , these same questions were investigated i n a subsample o f days on w h i c h ind iv idua l s experienced greater than average m o r n i n g pa in intensity. The purpose o f these last analyses was to investigate whether emot ion regulat ion inf luenced evening pa in intensity, not o n l y o n an everyday basis, but also i n the presence o f a cue for undesirable emotion. Daily Interview Completion and Descriptive Statistics The comple t ion rate for the da i ly interviews was h igh . O f a poss ible 1638 da i ly interviews, 9 8 % (n = 1612) were completed. O f the 117 participants, 8 7 % (n = 102) completed a l l 14 interviews, 8% (n = 9) were mis s ing one interview, 3 % (n = 4) were mi s s ing two interviews, 1% (n = 1) was mi s s ing four interviews, and 1% (n = 1) was m i s s i n g f ive interviews. M e a n s and standard deviations were calculated for da i ly in te rv iew study variables, aggregated across a l l t ime points (see Table 1). Participants reported an average leve l o f pa in intensity o f 4.21 (SD = 2.04) i n the morn ing and 3.92 (SD = 2.03) i n the evening. Average scores for anxiety and depression were 1.33 (SD = 0.42) and 1.25 (SD = 0.39) i n the m o r n i n g and 1.27 (SD = 0.41) and 1.22 (SD = 0.40) i n the evening, respectively. A v e r a g e scores for posi t ive emot ion were 2.74 (SD = 0.58) i n the morn ing and 2.84 (SD = 0.58) i n the evening. Pa i red t-tests compar ing mean m o r n i n g and evening levels o f study variables revealed that pa in intensity, 10 anxiety, and depression were a l l s ignif icant ly higher i n the morn ing , f(800) = 5.57,p < .001, /(799) = 4.17,p < .001, and.7(799) = 2.24,p < .05, respectively. Pos i t ive emot ion was s ignif icant ly higher i n the evening, ?(799) = -4.65, p < .001. Table 1. M e a n s and Standard Devia t ions o f D a i l y Interview Measures ( N = 117) Var i ab l e" Mean (SD) P a i n Intensity (0-10) A M pa in 4.21 (2.04) P M pa in 3.92 (2.03) A n x i e t y (1-4) A M anxiety 1.33 (0.42) P M anxiety 1.27 (0.41) Depress ion (1-4) A M depression 1.25 (0.39) P M depression 1.22 (0.40) Pos i t ive E m o t i o n (1-4) A M posi t ive emot ion 2.74 (0.58) P M posi t ive emot ion 2.84 (0.58) a V a l u e s for each var iable were aggregated for each participant across a l l t ime points . Frequencies were calculated for each type o f emot ion regulat ion (see Tab le 2). In terms o f the summative emot ion regulat ion scores, there were 171 (21%) days o n w h i c h there was no emot ion regulat ion, 229 (28%) days on w h i c h one emot ion was regulated, 231 (28%) days on w h i c h two emotions were regulated, and 169 (21%) days on w h i c h three emotions were regulated. W i t h respect to containing and mainta in ing emotions, there were 333 (41%) days on w h i c h no emotions were contained or maintained, 251 (31%) days o n w h i c h one emot ion was contained or maintained, 147 (18%) days on w h i c h two emotions were contained or maintained, and 69 (8%) days on w h i c h three emotions were contained or maintained. There were 470 (57%) days on w h i c h there was no recovery, 232 (28%) days on w h i c h there was recovery o f one emot ion, 80 (10%) days on w h i c h there was recovery o f two emotions, and 18 (2%) days on w h i c h there was recovery o f three emotions. A n x i o u s emot ion was regulated on 409 (50%) days, o f w h i c h 242 were instances (30%) o f containing anxious emot ion and 167 were instances (20%) o f recover ing from anxious emotion. Depressive emot ion was regulated on 382 (47%) days, w i t h 256 instances (31%) o f containing depressive emot ion, and 126 instances (15%) o f recover ing f rom depressive emot ion. Pos i t ive emot ion was regulated on 407 (50%) days, composed o f 254 days (31%) o f main ta in ing posi t ive emot ion and 153 days (19%) o f recover ing pos i t ive emot ion . 5 T a b l e 2. E m o t i o n Regula t ion Frequencies ( N = 819) T y p e o f E m o t i o n Regula t ion # % G l o b a l emot ion regulat ion N o emotions regulated 171 21 One emot ion regulated 229 28 T w o emotions regulated 231 28 Three emotions regulated 169 21 Con ta in ing or main ta in ing N o emotions contained or maintained 333 41 One emot ion contained or maintained 251 31 T w o emotions contained or maintained 147 18 Three emotions contained or maintained 69 8 R e c o v e r i n g N o recover ing 470 57 R e c o v e r i n g o f one emot ion 232 28 R e c o v e r i n g o f two emotions 80 10 R e c o v e r i n g o f three emotions 18 2 A n x i o u s emot ion Regula t ing 409 50 Con ta in ing 242 30 R e c o v e r i n g f rom 167 20 Depress ive emot ion Regula t ing 382 47 Con ta in ing 256 31 R e c o v e r i n g f rom 126 15 Pos i t ive emot ion Regula t ing 407 50 M a i n t a i n i n g 254 31 R e c o v e r i n g 153 19 5 There were 819 person-days in the sample and 19 person-days on which emotion regulation data was missing. 12 Bivariate Analyses Correla t ions were calculated among morn ing pa in intensity, evening pa in intensity, and the emot ion regulat ion variables. O n l y m o r n i n g pa in intensity was s igni f icant ly associated w i t h evening pa in intensity, r = .92, p < .001, h igh l igh t ing the importance o f con t ro l l ing morn ing pa in intensity in the mul t i - l eve l regression analyses. The lack o f significant associat ion between evening pain intensity and a l l o f the emot ion regulat ion variables is noteworthy. H o w e v e r , Pearson product moment correlations cannot be interpreted i n the context o f da i ly d ia ry data because they do not control for dependence i n the data or for w i t h i n - and between-person variance. G i v e n this, w e analyzed our data us ing mul t i - l eve l mode l ing . Multi-level Regression Analyses M u l t i - l e v e l regression analyses were conducted us ing h ierarchica l l inear mode l i ng ( H L M ) software (v6.0; Raudenbush, B r y k , Cheong , & C o n g d o n , 2004) . O n e o f the idiosyncrasies o f da i l y d ia ry studies is that there are necessari ly at least two levels o f data (days and ind iv idua ls ) one o f w h i c h (days) is nested w i t h i n the other ( indiv iduals ) . Da ta from da i ly d iary studies is also characterized b y dependence due to repeated sampl ing . M u l t i - l e v e l analyses conducted i n H L M take these characteristics o f the data into account and produce results that are unbiased b y these features, as w e l l as b y miss ing data (Raudenbush et a l . , 2004; Schwar tz & Stone, 1998). In mul t i - l eve l regression analyses, wi thin-person var ia t ion is mode led at L e v e l 1 and between-person var ia t ion is mode led at L e v e l 2, such that var ia t ion at both levels can be examined concurrently. In the L e v e l 1 specif icat ion o f wi th in-person var ia t ion , separate regression slopes and intercepts are estimated for each person. In the L e v e l 2 specif icat ion o f between-person var ia t ion, the L e v e l 1 regression parameters are used to estimate average parameter estimates across a l l subjects as w e l l as the amount o f var ia t ion around this average. 13 In the current study, variables der ived f rom da i l y interviews (e.g., pa in , emot ion regulation) were entered at L e v e l 1 and measures that were col lec ted dur ing the in i t i a l in terview, for w h i c h there was a single value for each participant (e.g., gender, funct ional d isabi l i ty) , were entered at L e v e l 2. R a n d o m intercept models were specif ied for a l l analyses (i.e., intercepts for each dependent variable were left free to vary) , such that f indings can be general ized to the popula t ion o f participants and days from w h i c h the sample and assessments respect ively were d r a w n . 6 Cont inuous L e v e l 1 predictor variables were centered on the mean o f each ind iv idua l ' s score over the course o f the study i n order to ease parameter est imation b y reducing correlations among slopes and intercepts (Nez lek , 2001). Further, for each person-centered continuous L e v e l 1 predictor var iable , the wi th in-person aggregate mean o f the var iable was entered at L e v e l 2. Th i s ensured that treatment o f the intercepts as random factors d i d not bias the coefficients o f the wi th in-person factors (Schwartz & Stone, 1998). S lope coefficients i n these models can be interpreted as the increase or decrease i n the dependent variable, at average levels o f the continuous variables and i n the condi t ion designated as 0 for the d ichotomous variables that were inc luded i n each m o d e l (Raudenbush & B r y k , 2002). Firs t , i n order to examine the proport ion o f between- and wi th in-person variance i n the dependent var iable , the f o l l o w i n g n u l l mode l was run for evening pa in intensity: L e v e l 1: Y i j ( P M Pain) = b0] + L e v e l 2: b0j = yoo + u 0j Th i s m o d e l specifies evening pa in intensity ( Y ) at t ime point i for i nd iv idua l j as a function o f 6qj and rjj. L e v e l 2 specifies that the L e v e l 1 intercept (6qj) is composed o f the grand sample mean o f evening pa in intensity across a l l participants and a l l evening t ime points ( y o o ) , plus the between-person residual parameter (UQJ), w h i c h is the difference between the grand sample mean and an 6 Both intercepts and slopes were initially modeled as random. However, in order to get models to converge, it was necessary to fix the slope coefficients (Nezlek, 2001). 14 ind iv idua l ' s o w n mean across a l l evening t ime points. The wi th in-person residual parameter (jy) represents the difference between an ind iv idua l ' s o w n mean across a l l t ime points and their level o f pa in intensity on a part icular evening. Th i s n u l l mode l indicated that participants var ied s ignif icant ly i n their average level o f evening pa in across the study per iod , % (116) = 2082.02, p < .001. The intraclass correlat ion coefficient ( I C C ) was estimated based on this mode l , and revealed that 7 1 % o f the variance i n evening pa in intensity was attributable to differences between ind iv idua ls , whereas, 2 9 % o f the variance was attributable to differences w i t h i n ind iv idua ls . In order to determine w h i c h variables to include at L e v e l 2 i n the models , the exploratory analysis ( L e v e l 2) function o f H L M was used. T h i s provides a Mo-en te r statistic for each potential L e v e l 2 var iable , based on w h i c h it is possible to determine w h i c h L e v e l 2 variables w i l l be significant i n a g iven mode l (Raudenbush & B r y k , 2002) . T h e demographic variables tested were: gender, age, ethnicity, and education. The disease status variables tested were: years since diagnosis, average fatigue i n the week pr ior to the background in terview, and functional disabi l i ty . Besides these demographic and disease status variables, between-person differences i n emot ional l ab i l i ty were tested i n order to determine whether they should be contro l led i n subsequent analyses. E m o t i o n a l l ab i l i ty was operat ional ized as the wi th in-person coefficient o f var ia t ion for each o f anxious, depressive, and pos i t ive emot ion. Consis tent w i t h recommended mul t i l eve l mode l specif icat ion, o n l y variables w i t h a Mo-en te r statistic greater than 1.96 or less than -1.96 were retained as control variables i n the f inal models predic t ing evening pa in intensity (Kreft & de L e e u w , 1998). Does number o f emotions regulated influence evening pa in intensity? Fi rs t , a mode l was specif ied predic t ing evening pa in intensity f rom g loba l emot ion regula t ion . 7 Resul ts indicated 7 The control variables included in this model were average fatigue in the week prior to the background interview, functional disability, and lability of positive emotion. 15 that the number o f emotions regulated over the course o f a g iven day was s igni f icant ly and negat ively associated w i t h evening pa in intensity, b = -0 .31, /(795) = -6.15, p < .001 (see Table 3, mode l 1). T h i s replicated previous f indings (Paquet et a l . , 2005). N e x t , m o r n i n g pa in as w e l l as morn ing anxious, depressive and pos i t ive emot ion were added to the mode l such that the f inal mode l can be expressed as: L e v e l 1: Y j , ( P M Pain) = Z?0j + Z>H(AM Pain) + 6 2 j ( A M A n x i e t y ) + 6 3 j ( A M Depression) + 6 4 j ( A M Pos i t ive Emot ion ) + ^ ( G l o b a l E m o t i o n Regula t ion) + r^ L e v e l 2: boj = Yoo + Jo I (Age) + Yo2 ( W i t h i n - P e r s o n Aggregate M e a n o f A M Pain) + Yo3(Wifhin-Person Aggregate M e a n o f A M A n x i e t y ) + yo4(Wi th in -Pe r son Aggregate M e a n o f A M Depression) + yosCWithin-Person Aggregate M e a n o f A M Pos i t ive Emot ion ) + UQJ &IJ = YIO bij = Y 2 0 &3J = Y30 &4j = Y40 &5j=Y50 A t L e v e l 1, evening pa in intensity on any g iven day ( Y y ( P M Pain)) was specif ied as a function o f the i nd iv idua l ' s average evening pa in intensity across a l l days (boj), the m a i n effects o f morn ing pa in intensity (fry), m o r n i n g anxious emotion (foj), morn ing depressive emot ion (b^), morn ing posi t ive emot ion (b^), and number o f emotions regulated (bsj), as w e l l as that day ' s deviat ion f rom the average (rjj). A t L e v e l 2, the L e v e l 1 intercept (bo) for any person (j) was specif ied as a function o f the average intercept (mean pa in intensity) across persons (yoo), age, average morn ing pa in , average m o r n i n g anxious emotion, average morn ing depressive emot ion , and average m o r n i n g posi t ive emot ion across the week, their respective regression coefficients, ( jo i , Y 0 2 , Yo3, 16 Y04, Jos), and a random component (uoi). Results indicated that higher levels o f emotions regulated on a g iven day were s ignif icant ly associated w i t h less evening pa in , b = -0.21, £(789) = -4.40, p < .001, above and beyond the significant effect o f morn ing pa in intensity, b = 0.33, £(789) = 5.82, p < .001 (see Table 3, mode l 2). T a b l e 3. Assoc ia t ions o f G l o b a l E m o t i o n Regula t ion w i t h E v e n i n g P a i n Intensity M o d e l b SE 1. N o controls at L e v e l 1 -0.31* 0.05 2. C o n t r o l l i n g baseline pa in and emotions -0.21* 0.05 3. H i g h m o r n i n g pa in days o n l y -0.30* 0.07 Note . */?<.001. A r e both types o f emot ion regulat ion (maintaining or conta ining and recovering') prospect ively associated w i t h evening pa in intensity? In this m o d e l , evening pa in intensity was predicted f rom recover ing f rom anxious and depressive emot ion and o f posi t ive emot ion, as w e l l as f rom conta ining anxious and depressive emot ion and main ta in ing pos i t ive emot ion. Consistent w i t h previous findings (Paquet et a l . , 2005), both main ta in ing or containing as w e l l as recover ing were s igni f icant ly and negat ively associated w i t h evening pa in , b = -0.34, £(794) = -6.35,p < .001 and b = -0.24, £(794) = -3.52,p < .01, respect ively (see Tab le 4, mode l 1). N e x t , m o r n i n g pa in and morn ing emotions were added to the m o d e l such that the final mode l can be expressed as: L e v e l 1: Y i j f P M Pain) = bQi + b\j(AM Pain) + £> 2j(AM A n x i e t y ) + £> 3 j (AM Depression) + £>4j(AM Pos i t ive Emot ion ) + ^ ( R e c o v e r i n g ) + 6 6 j (Conta in ing/Mainta in ing) + ry L e v e l 2: bq/ = Yoo + Yoi(Age) + y 0 2 (Wi th in -Pe r son Aggregate M e a n o f A M Pain) + Yo3(Within-Person Aggregate M e a n o f A M A n x i e t y ) + Yo4(Wifhin-Person Aggregate M e a n o f A M Depression) + y 0 5 ( W i t h i n - P e r s o n Aggregate M e a n o f A M Pos i t ive Emot ion ) + uoj 8 The control variables included in this model were average fatigue in the week prior to the background interview, functional disability, and lability of positive emotion. 17 &ij = Yio 02) = Y 2 0 bij = Y30 &4j = Y40 &5j = Y50 &6j = Y60 A t L e v e l 1, evening pa in intensity on any g iven day ( Y j j ( P M Pain)) was specif ied as a function o f the i nd iv idua l ' s average evening pa in intensity across a l l days (boy), the m a i n effects o f morn ing pa in intensity (b\y), morn ing anxious emot ion ( & 2 j ) , m o r n i n g depressive emot ion (by^), morn ing posi t ive emot ion (b^y), the number o f emotions for w h i c h recovery occurred (b5y), the number o f emotions that were contained or maintained (b^), and that day ' s devia t ion f rom the average (r;j). A t L e v e l 2, the L e v e l 1 intercept (bo) for any person (j) was specif ied as a funct ion o f the average intercept (mean pa in intensity) across persons (yoo), age, average m o r n i n g pa in , average morn ing anxious emot ion, average morn ing depressive emot ion, average m o r n i n g pos i t ive emot ion across the week, their respective regression coefficients (yoi, Y 0 2 , Y03, Y04, Y05) , and a r andom component (uoi). Resul ts indicated that the greater the number o f emotions that were contained or maintained, and the greater the number o f emotions for w h i c h recovery occurred, were each s ignif icant ly associated w i t h lower evening pain, b - -0.22, t(788) - -3 .61,p < .01 and b = -0 .21, t(7SS) = -3.12, p< .01, above and beyond the significant effect o f m o r n i n g pa in intensity, b= 0.33, f(788) = 5.84,p < .001 (see Table 4, mode l 2). 18 T a b l e 4. Assoc ia t ions o f M a i n t a i n i n g or Conta in ing , and R e c o v e r i n g , w i t h E v e n i n g P a i n Intensity M o d e l b SE 1. N o controls at L e v e l 1 M a i n t a i n i n g or containing -0.34*** 0.05 Recove r ing -0.24** 0.07 2. C o n t r o l l i n g baseline pa in and emotions M a i n t a i n i n g or containing -0.22** 0.06 R e c o v e r i n g -0 .21** 0.07 3. H i g h m o r n i n g pa in days o n l y M a i n t a i n i n g or conta ining -0 .38*** 0.09 Recove r ing -0 .23* 0.09 Note . *p < .05, **p < .01, ***p < .001. Does regulat ing different emotions (anxious, depressive, and posi t ive) have an influence on evening pa in intensity? In order to answer this question, a m o d e l was specif ied i n w h i c h evening pa in intensity was predicted from regulat ion o f anxious, depressive, and posi t ive emot ion . 9 Consistent w i t h previous research (Paquet et a l . , 2005) , regulat ing anxious emot ion was a significant predictor o f evening pa in intensity, b= -0.28, £(793) = -2.04, p < .05. In the current study, however , regulat ing depressive emotion over the course o f a day was also s ignif icant ly associated w i t h less evening pa in , b= -0.59, /(793) = -4.39, p < .001. Regula t ing posi t ive emot ion was not a significant predictor i n this mode l (see Tab le 5, m o d e l 1). N e x t , morn ing pain as w e l l as morn ing anxious, depressive, and posi t ive emot ion were added to the mode l such that the final mode l can be expressed as: L e v e l 1: Y i j ( P M Pain) = by + ZMJ(AM Pain) + 6 2 j ( A M A n x i e t y ) + fc3J(AM Depression) + £>4j (AM Pos i t ive Emot ion ) + />5j(Anxiety Regula t ion) + ^ ( D e p r e s s i o n Regulat ion) + ^ ( R e g u l a t i o n o f Pos i t ive Emot ion ) + ry L e v e l 2: bqj = yoo + Yoi(Age) + y 0 2 (With in-Person Aggregate M e a n o f A M Pain) + Yo3(Within-Person Aggregate M e a n o f A M A n x i e t y ) + Yo4(Wifhin-Person 9 The control variables included in this model were average fatigue in the week prior to the background interview, functional disability, and lability of positive emotion. 19 Aggregate M e a n o f A M Depression) + y 0 5(With in-Person Aggregate M e a n o f A M Pos i t ive Emot ion ) + uqj = Yio = 720 hi = Y30 &4j = Y40 b5j = Y50 = Yeo * 7 j = Y70 A t L e v e l 1, evening pa in intensity on any g iven day ( Y y ( P M Pain)) was specif ied as a function o f the i nd iv idua l ' s average evening pa in intensity across a l l days (6qj), the m a i n effects o f morn ing pa in intensity (6ij), morn ing anxious emot ion (6 2j), m o r n i n g depressive emot ion (&3j), morn ing posi t ive emot ion (6 4j), regulating anxious emot ion (&5j), regulat ing depressive emot ion (6 6j), regulat ing posi t ive emot ion (&7j) and that day 's devia t ion f rom the average ( r y ) . A t L e v e l 2, the L e v e l 1 intercept (bo) for any person (j) was specif ied as a funct ion o f the average intercept (mean pa in intensity) across persons (yoo), age, average m o r n i n g pa in , average m o r n i n g anxious emot ion , average m o r n i n g depressive emotion, average m o r n i n g pos i t ive emot ion across the week, their respective regression coefficients (yoi, Y02, Y03, Y04, Y05), and a r andom component (uoi). Results indicated that after con t ro l l ing for the significant effect o f m o r n i n g pa in intensity, b = 0.33, £(787) = 5.94,p < .001, regulating depressive emot ion remained the o n l y significant predictor o f evening pa in intensity, b = -0.39, £(787) = -3.25, p < .01 (see Tab le 5, mode l 2). 20 T a b l e 5. Assoc ia t ions o f Regula t ing A n x i o u s , Depressive, and Pos i t ive E m o t i o n w i t h E v e n i n g P a i n Intensity M o d e l b SE 1. N o controls at L e v e l 1 Regula t ing anxious emot ion -0.28* 0.14 Regula t ing depressive emot ion -0 .59*** 0.13 Regula t ing posi t ive emot ion -0.34 0.10 C o n t r o l l i n g baseline pa in and emotions Regula t ing anxious emot ion -0.14 0.11 Regula t ing depressive emot ion -0 .39** 0.12 Regula t ing posi t ive emot ion -0.14 0.09 H i g h m o r n i n g pa in days o n l y Regula t ing anxious emot ion -0.25 0.17 Regula t ing depressive emot ion -0.39* 0.15 Regula t ing posi t ive emot ion -0.27 0.15 Note . *p < .05, **p < .01, ***p < .001. A r e conta ining or mainta ining, and recover ing (from), each o f the emotions s ignif icant ly and independently associated w i t h evening pa in intensity? Firs t , consistent w i t h previous research (Paquet et a l . , 2005), a mode l was specif ied i n w h i c h conta in ing anxious , recovering f rom anxious, conta ining depressive, recover ing f rom depressive, ma in ta in ing posi t ive , and recover ing posi t ive emot ion were a l l inc luded as predictors o f evening pa in in tens i ty . 1 0 Conta in ing anxious emot ion, as w e l l as containing and recover ing f rom depressive emot ion were s ignif icant ly associated w i t h evening pa in intensity, b= -0 .31, f(790) = - 2 . 1 3 , ^ < .05, b - -0.63, f(790) = -4.40,p < .001, and b= -0.49, f(790) = -3.00,p < .01, respect ively (see Table 6, mode l !)• N e x t , m o r n i n g pa in and morn ing emotions were added to the mode l such that the final mode l can be expressed as L e v e l 1: Y y ( P M Pain) = 6 0j + 6 i j ( A M Pain) + 6 2 j ( A M A n x i e t y ) + by3(Containing A n x i e t y ) + &4j(Recovering f rom A n x i e t y ) + &5j(AM Depress ion) + b^(Containing 1 0 The control variables included in this model were average fatigue in the week prior to the background interview, functional disability, and lability of positive emotion. 21 Depression) + ^ ( R e c o v e r i n g from Depression) + b^(AM Pos i t ive Emot ion) + &9j (Maintaining Pos i t ive Emot ion) + b IQJ (Recover ing Pos i t ive Emot ion ) + L e v e l 2: boj = yoo + Yoi(Age) + yo2(Within-Person Aggregate M e a n o f A M Pain) + Yo3(Within-Person Aggregate M e a n o f A M A n x i e t y ) + Yo4(Within-Person Aggregate M e a n o f A M Depression) + yo5(Wifhin-Person Aggregate M e a n o f A M Pos i t ive Emot ion ) + U O J 6ij = Yio &2j = Y20 bij = Y30 6 4 j = Y40 bsj = Yso &6j = Y60 bjj = Y70 bi] = Yso &9j = Y90 &ioj = Y100 A t L e v e l 1, evening pa in intensity on any g iven day ( Y j j ( P M Pain)) was specif ied as a function o f the i nd iv idua l ' s average evening pain intensity across a l l days (&oj), the m a i n effects o f morn ing pa in intensity (6n), morn ing anxious emot ion (Z>2j), containing anxious emot ion (63,-), recover ing f rom anxious emot ion (b$, morn ing depressive emot ion (bsf), conta in ing depressive emot ion (&6j), recover ing f rom depressive emot ion (&7j), morn ing posi t ive emot ion (6gj), main ta in ing posi t ive emot ion (£>9j), recover ing posi t ive emot ion (& IOJ), and that day ' s devia t ion from the average (rjj). A t L e v e l 2, the L e v e l 1 intercept (bo) for any person (j) was specif ied as a function o f the average intercept (mean pa in intensity) across persons (yoo), age, average m o r n i n g pain , average m o r n i n g anxious emotion, average morn ing depressive emot ion , average m o r n i n g posi t ive emot ion across the week, their respective regression coefficients (yoi, Y o 2 , Y o 3 , Yo4,Yo5)> a n d a random component (uoi). Results indicated that, above and b e y o n d the significant effect o f m o r n i n g pa in intensity, b = 0.33, f(784) = 5.90,p < .001, o n l y conta in ing and recover ing from depressive emot ion were s ignif icant ly associated w i t h decreased evening pa in intensity, b = -0.37, 4784) = -2.62,p < .01 and b = -0.40, f(784) = -2.il,p <.01, respect ively. Con ta in ing anxious emot ion ceased to be a significant predictor i n this mode l (see Tab le 6, mode l 2). T a b l e 6. Assoc ia t ions o f Con ta in ing or M a i n t a i n i n g , and R e c o v e r i n g (from) each o f the Emot ions w i t h E v e n i n g P a i n Intensity M o d e l b SE 1. N o controls at L e v e l 1 Con ta in ing anxious emot ion -0 .31* 0.14 R e c o v e r i n g f rom anxious emot ion -0.24 0.16 Con ta in ing depressive emot ion -0 .63*** 0.14 R e c o v e r i n g f rom depressive emot ion -0 .49** 0.16 M a i n t a i n i n g posi t ive emot ion -0.16 0.11 R e c o v e r i n g posi t ive emot ion -0.07 0.13 2. C o n t r o l l i n g baseline pa in and emotions Con ta in ing anxious emot ion -0.20 0.12 Recove r ing f rom anxious emot ion -0.07 0.15 Con ta in ing depressive emot ion -0 .37** 0.14 R e c o v e r i n g f rom depressive emot ion -0 .40** 0.14 M a i n t a i n i n g posi t ive emot ion -0.10 0.12 R e c o v e r i n g pos i t ive emot ion -0.18 0.12 3. H i g h m o r n i n g pa in days o n l y Con ta in ing anxious emot ion -0 .41* 0.19 R e c o v e r i n g f rom anxious emot ion -0.11 0.22 Con ta in ing depressive emot ion -0 .43* 0.20 R e c o v e r i n g f rom depressive emot ion -0 .33 # 0.17 M a i n t a i n i n g posi t ive emot ion -0.29 0.19 R e c o v e r i n g posi t ive emot ion -0.26 0.18 Note . *p = .051, *p < .05, **p < .01, ***p < .001. A r e results the same w h e n a cue for undesirable emot ion is present? T h e analyses conducted thus far are consistent w i t h a l ifespan developmental perspective o f emot ion regulat ion, w h i c h v iews emot ion regulat ion as an everyday process. Research conducted from this perspective has general ly been concerned w i t h self-supported maintenance o f desirable 23 emotional states and recovery f rom undesirable states i n the context o f no rma l da i l y situations ( D i a m o n d & A s p i n w a l l , 2003 ;Taque t et a l . , 2005). In contrast, researchers w h o take an evolut ionary approach to emot ion regulation have been more i nc l i ned to use experimental designs i n order to manipulate emot ion to ensure that there is an undesirable emot ional state avai lable to be regulated (e.g., Gross & Levenson , 1997; M a u s s , C o o k , & Gross , i n press). In order to take into account both o f these approaches, the preceding questions were re-examined i n a l imi t ed sample o f days on w h i c h morn ing pa in intensity was higher than the i nd iv idua l ' s o w n average across the seven sampl ing days. Because morn ing pa in intensi ty was s ignif icant ly associated w i t h higher m o r n i n g anxious emot ion and depressive emot ion as w e l l as w i t h lower m o r n i n g posi t ive emot ion, r ( l 15) = 29, p < .01, r ( l 15) = .36, p < .001, and r ( l 15) = - .33, p < .001, respectively, these cases represent instances dur ing w h i c h an emot ional cue was more l i k e l y to have been present w i t h respect to w h i c h emot ion regulat ion c o u l d have been employed. The advantage o f the current b lended approach is that it includes a control for undesirable emot ional state, w h i c h is par t icular ly relevant i n the context o f the f luctuating symptoms and associated emotions o f rheumatoid arthritis (Smi th & W a l l s t o n , 1992), as w e l l as mainta in ing the eco logica l v a l i d i t y o f assessing pa in and emotions i n the context o f the part icipants ' everyday l ives . A p p l y i n g this control resulted i n a reduced data set composed o f 116 participants and 380 days o f data. In this sample, the descriptive statistics for the d a i l y pa in and emot ion variables were ident ical to those i n the fu l l dataset except that average m o r n i n g pa in intensity was 4.22 (SD = 2.05) and average evening pa in intensity was 3.93 (SD = 2.03). A pai red t-test revealed that pa in intensity was s ignif icant ly higher i n the morn ing than i n the evening, f(365) = 12.58,/? < . 0 0 1 . O f the possible 380 sampl ing days dur ing w h i c h various types o f emot ion regulat ion cou ld have been exhibi ted, there were 96 (25%) days on w h i c h there was no emot ion regulation, 24 105 (28%) days on w h i c h one emot ion was regulated, 97 (26%) days on w h i c h two emotions were regulated, and 68 (18%) days on w h i c h three emotions were regulated. W i t h respect to containing or main ta in ing emotions, there were 180 (48%) days on w h i c h no emotions were contained or maintained, 110 (29%) days on w h i c h one emot ion was contained or maintained, 56 (15%) days on w h i c h two emotions were contained or maintained, and 20 (5%) days on w h i c h three emotions were contained or maintained. There were 202 (53%) days on w h i c h there was no recovery, 119 (31%) days on w h i c h there was recovery o f one emot ion , 33 (9%) days on w h i c h there was recovery o f two emotions, and 12 (3%) days on w h i c h there was recovery o f three emotions. A n x i o u s emotion was regulated on 175 (46%) days, o f w h i c h there were 97 instances (26%) o f conta ining and 78 instances (21%) o f recovering. Depress ive emot ion was regulated on 161 (42%) days, i nc lud ing 93 instances (25%) o f conta ining and 68 instances (18%) o f recovering. F i n a l l y , posi t ive emot ion was regulated on 167 (44%) days. There were 92 instances (24%) o f main ta in ing posi t ive emot ion and 75 instances (20%) o f recover ing posi t ive emot ion (see Tab le 7 ) . 1 1 1 1 There were 14 person-days on which emotion regulation variables were missing. 25 T a b l e 7. E m o t i o n Regula t ion Frequencies: H i g h M o r n i n g P a i n D a y s (N = 380) T y p e o f E m o t i o n Regula t ion _ % G l o b a l emot ion regulat ion N o emotions regulated 96 25 One emot ion regulated 105 28 T w o emotions regulated 97 26 Three emotions regulated 68 18 Con ta in ing or main ta in ing N o emotions contained or maintained 180 48 One emot ion contained or maintained 110 29 T w o emotions contained or maintained 56 15 Three emotions contained or maintained 20 5 Recove r ing R e c o v e r i n g o f one emot ion 202 53 R e c o v e r i n g o f two emotions 119 31 Recove r ing o f three emotions 33 9 Recove r ing o f one emot ion 12 3 A n x i o u s emot ion Regula t ing 175 46 Con ta in ing 97 26 R e c o v e r i n g f rom 78 21 Depress ive emot ion Regula t ing 161 42 Con ta in ing 93 25 R e c o v e r i n g f rom 68 18 Pos i t ive emot ion Regula t ing 167 44 M a i n t a i n i n g 92 24 R e c o v e r i n g 75 20 A s i n the fu l l dataset, bivariate correlations indicated that m o r n i n g pa in intensity was s ignif icant ly associated w i t h evening pa in intensity, r = .86,/? < .001, whereas, there were no significant associations between evening pa in intensity and any o f the emot ion regulat ion variables. T h i s is , again, noteworthy but not necessari ly indica t ive o f a l ack o f significant influence o f emot ion regulat ion on evening pa in intensity. The n u l l mode l for this dataset indicated that participants var ied s igni f icant ly i n their average level o f evening pa in across the study per iod, x 2 ( l 15) = 1071.30,/? < .001. The intraclass correlat ion coefficient ( I C C ) was estimated based on this mode l , and revealed that 7 2 % o f the 26 variance i n evening pa in intensity was attributable to differences between ind iv idua l s , whereas, 2 8 % o f the variance was attributable to differences w i t h i n ind iv idua l s . Consistent w i t h the results obtained i n the fu l l dataset, h igher numbers o f emotions regulated on a g iven day were s ignif icant ly associated w i t h decreased evening pa in , b = -0.30, £(355) = -4 .61, p < .001, above and beyond the significant influence o f m o r n i n g pa in intensity, b = 0.30, £(355) = 2.55, p < .05 (Table 3, mode l 3 ) . 1 2 L i k e w i s e , each o f main ta in ing or containing as w e l l as recover ing was independently associated w i t h decreased evening pain , b = -0.38, £(354) = -4.22,p < .001 and b = -0.23, £(354) = -2.54,p < .05, respect ively, above and beyond the significant influence o f both morn ing pa in intensity and m o r n i n g anxious emot ion , b = 0.30, £(354) = 2.51, p < .05 and b = -0.56, £(354) - -2.05, p < .05, respect ively (Table 4, mode l 3). A l s o , consistent w i t h previous analyses cont ro l l ing for m o r n i n g pa in and m o r n i n g emotions i n the fu l l sample, regulat ing anxious emot ion was not a significant predictor o f evening pa in intensity and regulat ing depressive emot ion was s ignif icant ly associated w i t h decreased evening pa in , b = -0.39, £(353) = -2.55,p < .05, i n this subsample. Regu la t ing pos i t ive emot ion remained nonsignif icant i n this mode l . Results concerning the effects o f regulat ing anxious , depressive, and posi t ive emot ion were above and beyond the significant influence o f m o r n i n g pa in intensity, b = 0.30, £(353) = 2.55, p < . 05 (Tab le 5, mode l 3). F i n a l l y , above and b e y o n d the significant effects o f m o r n i n g pa in intensity and morn ing anxious emot ion , b = 0.29, £(350) = 2.44, p < .05 and b = -0.65, £(350) = -2.02, p < .05, respectively, and consistent w i t h analyses i n the fu l l sample, containing depressive emot ion was s ignif icant ly associated w i t h less evening pain, b = -0.43, £(350) = -2.20,p < .05. Recove r ing from depressive emot ion dropped to a trend level o f signif icance, b = -0.33, £(350) = -1.95,p = .051. Cont ra ry to analyses us ing the fu l l sample, conta ining anxious emot ion was also s ignif icant ly associated w i t h decreased evening pa in i n the 1 2 Age was the only significant demographic factor, disease factor, or emotional lability variable in this and all other models using the high morning pain subsample. h igh m o r n i n g pa in subsample, b = -0.41, £(350) = -2.22, p< .05. N o n e o f recover ing from anxious emot ion , main ta in ing posi t ive emotion, or recover ing pos i t ive emot ion was significai associated w i t h evening pa in i n this subsample (Table 6, mode l 3). Conclusions The purpose o f the current study was to investigate the associations between da i ly emot ion regulat ion and da i ly pa in among ind iv idua ls w i t h chronic pa in . O v e r a l l , emot ion regulat ion had a prospect ive influence on pa in intensity. T h i s was the case for both number o f emotions regulated over the course o f the day, as w e l l as for each o f main ta in ing or containing, and recover ing, independently. W e also found that the influence o f emot ion regulat ion on subsequent pa in was largely dr iven b y regulating, spec i f ica l ly conta in ing, depressive emotion. The overa l l f ind ing that emot ion regulation influences subsequent pa in intensity is consistent w i t h a previous f inding that emot ion regulation, assessed us ing a cross-sectional design w i t h a self-report measure, was prospect ively associated w i t h perce ived disease act ivi ty, i nc lud ing pain , at a one-and-a-half year fo l low-up among ind iv idua l s w i t h rheumatoid arthritis (van M i d d e n d o r p , Geenen , Sorb i , van Doornen , et a l . , 2005). T h i s , and the f ind ing that each o f main ta in ing and recover ing desirable emotional states had independent influences on subsequent pain , was also consistent w i t h another study w h i c h found, us ing the same empi r i ca l l y der ived measures o f emot ion regulat ion that are used i n the current study, that d a i l y emot ion regulat ion inf luenced subsequent pain among a sample o f e lder ly ind iv idua l s w h o had not been diagnosed w i t h a chronic pa in condi t ion (Paquet et a l . , 2005). Genera l ly , these f indings support the neuromatrix theory o f pa in , w h i c h posits that regulat ion o f emot ional inputs affects the resultant pattern o f pa in outputs ( M e l z a c k , 1999). The current study goes beyond the findings o f previous research b y con t ro l l ing for baseline m o r n i n g pa in intensity as w e l l as for baseline m o r n i n g emotions per se. M o r n i n g pa in intensity is s ignif icant ly associated w i t h evening pa in intensity, and there is a wel l -establ ished l i n k between emotions and pain . So, it is c rucia l to determine that the associations o f emot ion regulat ion w i t h subsequent pa in are above and beyond the influence o f these baseline factors. Ano the r advantage o f the current study is that it is the first that w e k n o w o f to apply these 29 empi r i ca l ly der ived measures o f emot ion regulat ion to the invest igat ion o f emot ion regulat ion among a popula t ion o f ind iv iduals experiencing chronic pain . O u r findings indicate that, even after con t ro l l ing for baseline pa in and emotions, as w e l l as when the analyses were implemented on a subset o f days on w h i c h a cue for undesirable emotions was present, overa l l emot ion regulation, as w e l l as each o f main ta in ing and recover ing desirable emot ional states, were s ignif icant ly associated w i t h less subsequent pa in . Howeve r , our addi t ional controls lead to nove l f indings when emot ion regulat ion is disaggregated b y emotion, as w e l l as w h e n it is disaggregated b y emot ion and type o f regulat ion. In the first instance, when w e implemented our analyses without the controls, both regulat ing anxious emot ion and regulat ing depressive emot ion were s ignif icant ly associated w i t h decreased evening pa in intensity. H o w e v e r , both when w e control led for baseline m o r n i n g pa in and emotions, as w e l l as when w e conducted the analyses on the subsample o f days w h e n m o r n i n g pa in was higher than average, o n l y regulat ing depressive emot ion was s igni f icant ly associated w i t h subsequent pa in intensity. T h i s is inconsistent w i t h previous research w h i c h found that, wi thout implement ing our addi t ional controls, regulat ing anxious emot ion was the o n l y s ignif icant predictor o f subsequent pa in w i t h i n the same day (Pauqet et a l . , 2005). A possible explanat ion for these equ ivoca l results can be found i n the general ly stronger association between pa in and anxie ty condit ions than between pa in and depression ( M c W i l l i a m s , G o o d w i n , & C o x , 2004) . In the current study, con t ro l l ing for m o r n i n g pa in i n the fu l l dataset and conduct ing the analyses on h igh morn ing pa in days o n l y m a y have resulted i n cont ro l l ing for more o f the variance associated w i t h anxious emot ion than w i t h depressive emot ion, thereby, m a k i n g it more l i k e l y for the influence o f regulat ing depressive emot ion to emerge as significant. Th i s h ighl ights the importance o f the two pa in controls that we implemented and suggests that, for ind iv idua l s w i t h a chronic pa in condi t ion , pa in management efforts might most useful ly focus on regulat ing depressive, rather than anxious or posi t ive, emotions. 30 Previous research also led us to expect that, w h e n emot ion regulat ion was disaggregated b y both emot ion and type o f regulation, both containing and recover ing f rom anxious emotion, as w e l l as recover ing posi t ive emotion, w o u l d be s ignif icant ly associated w i t h decreased evening pa in (Paquet et a l . , 2005). In our first mode l , wi thout the addi t ional controls , w e found a significant influence o f conta ining anxious emot ion, but we also found independent influences o f containing and recover ing f rom depressive emotion. In m o d e l 2, w h i c h inc luded the morn ing pa in and morn ing emotions controls, on ly containing and recover ing f rom depressive emot ion remained significant. In the analyses conducted on the subsample o f days w h e n participants had higher than their o w n average m o r n i n g pain , conta ining anxious and conta ining depressive emot ion were each s ignif icant ly associated w i t h decreased evening pa in . Tentat ively , w e suggest that the differences between what w e found and what was expected based o n previous research m a y have had to do w i t h the differences between the two samples. L i f e span research suggests that anxiety, but not depression, decreases w i t h age ( A l e x o p o u l o s , 1990) and that older adults are re la t ively happy (Diener & Diener , 1996) and satisfied w i t h l i fe ( H e r z o g & Rodgers , 1981). Therefore, the dysregulat ion o f anxious and posi t ive emot ion m a y have a greater influence among older adults for w h o m greater regulat ion is the norm. A l t h o u g h some important improvements are made i n the current study, as compared to previous research on the topic o f emot ion regulat ion and pain , the current study is not without its l imita t ions . O n the one hand, the measure w e used captures dynamic and subconscious aspects o f emotions regulat ion that one-time, self-reports o f emot ion regulat ion do not. O n the other hand, some instances o f emot ion regulat ion m a y have been missed. F o r example , one can imagine a day o n w h i c h an event occurred that resulted i n a participant h a v i n g substantial ly higher than average negative emotions or substantially lower than average pos i t ive emot ion. One can also imagine that, on one o f those days, the participant might have somewhat regulated their emotions but s t i l l have higher than average negative emotions or lower than average posi t ive emotions. 31 The current measure o f emot ion regulation w o u l d not have counted those days as days on w h i c h emot ion was regulated, although, as is clear f rom the example , it was . T h e consequence o f this l imi ta t ion is that nonsignif icant results m a y not reflect a true lack o f inf luence o f emot ion regulat ion on pa in intensity, but rather the inab i l i ty o f the measure to fu l ly capture emot ion regulat ion when it occurred. One w a y o f addressing this l imi ta t ion i n future research w o u l d be to ask participants whether anything important had happened since their previous interview, the valence o f the event, and h o w serious the event was for them. In future research a trait measure o f self-perceived emot ion regulat ion as w e l l as da i l y self-report measures o f emot ion regulation cou ld be added to research protocols i n order to examine the independent influences o f between and wi th in - ind iv idua l differences i n emot ion regulat ion as w e l l as the extent to w h i c h self-report and empi r i ca l ly der ived da i ly measures o f emot ion regulat ion concur. A no the r useful di rect ion for future research might be to consider the temporal aspect o f affect regulat ion i n order to determine whether the impact o f m o o d regulat ion on pa in is s imi la r to that o f emot ion regulation. Th i s consideration arises out o f the different timeframes and methodologies associated w i t h the two approaches to emot ion regulat ion that have informed this study. One branch o f emot ion regulat ion research takes its departure from an evolut ionary perspective (e.g., Gross , 1998), w h i c h v iews emotions as momentary and tends to conduct experiments i n w h i c h emotions and emot ion regulat ion are manipula ted (e.g., Gross & Levenson , 1997; M a u s s et a l . , i n press). In contrast, the l ifespan developmental perspective does not make as fine a temporal d is t inct ion between moods and emotions and tends to examine everyday experiences o f emot ion, i nc lud ing regulat ion (e.g., Carstensen et a l . , 2000; Paquet et a l . , 2005). The advantage o f the first approach is methodologica l r igor, whereas, the advantage o f the second is eco log ica l va l id i ty . The methodology o f the current study is more c lose ly a l igned w i t h the second approach. Therefore, it might be useful i n future research to measure emot ion regulat ion i n close temporal p r o x i m i t y to emotional cue events, us ing , for example , event-32 triggered sampl ing (Bo lge r et a l . , 2003). Th i s w o u l d s imul taneously preserve the eco logica l va l id i t y associated w i t h the l ifespan developmental approach, as w e l l as more accurately measure emot ion regulat ion as it is defined b y the evolut ionary approach. In sum, the current study responds to recent calls i n the literature to go beyond investigations o f the effects o f static emotions on pain , to examin ing the influence o f dynamic emot ion regulat ion (e.g., Keefe et a l . , 2001). T h i s approach is consistent w i t h the dynamic mode l o f affect w h i c h asserts that emotions are not static and that their influence on phys ica l and psycho log ica l outcomes fluctuates across short timeframes (Zautra, S m i t h , A f f l e c k , & Tennen, 2001). Results o f the current study indicate that both main ta in ing desirable emot ional states and recover ing f rom undesirable emotional states have important influences on the amount o f da i ly pa in experienced b y ind iv idua ls w i t h chronic pa in . T h i s seems to be par t icular ly the case for depressive emot ion. A key impl i ca t ion o f this research for psychosoc ia l pa in management interventions is that it m a y be most important to teach ind iv idua ls w i t h chronic pa in condit ions h o w to manage their depressive emotions, especial ly h o w to a v o i d exper iencing them i n the first place. 33 References A l e x o p o u l o s , G . S. (1990). Anxie ty-depress ion syndromes i n o l d age. International Journal of Geriatric Psychiatry, 5, 351-353. Ande r son , K . , Brad ley , L . , Y o u n g , L . , M c D a n i e l , L . , & W i s e , C . (1985). Rheuma to id arthritis: R e v i e w o f psycho log ica l factors related to et iology, effects, and treatment. Psychological Bulletin, 98, 358-387. B e l z a , B . (1995). C o m p a r i s o n o f self-reported fatigue i n rheumatoid arthritis and controls. Journal of Rheumatology, 2, 639-643. Bo lge r , N . , D a v i s , A . , & Rafae l i , E . (2003). D i a r y methods: Cap tu r ing l i fe as it is l i ved . Annual Review of Psychology, 54, 579-616. Carstensen, L . L . , Pasupathi , M . , M a y r , U . , & Nesselroade, J . R . (2005). E m o t i o n a l experience i n everyday l i fe across the adult l i fe span. Journal of Personality and Social Psychology, 79, 644-655. C r a i g , K . D . (1999). Emot ions and psychobio logy . In P . D . W a l l & R . M e l z a c k (Eds.) , Textbook of pain, (pp. 331-343). N e w Y o r k : C h u r c h i l l L iv ings tone . Derogat is , L . R . (1975). Derogatis affects balance scale. Ba l t imore : C l i n i c a l Psychometr ic Research. D i a m o n d , L . M . , & A s p i n w a l l , L . G . (2003). E m o t i o n regulat ion across the l ifespan: A n integrative perspective, emphas iz ing self-regulation, pos i t ive affect, and dyadic processes. Motivation and Emotion, 27, 125-156. Diener , E . , & Diener , C . (1996). M o s t people are happy. Psychological Science, 7, 181-185. Fe ldman , S. L , D o w n e y , G . , & Schaffer-Nei tz , R . (1999). P a i n , negative m o o d , and perce ived support i n chronic pa in patients: A da i ly d iary study o f people w i t h reflex sympathetic dystrophy syndrome. Journal of Consulting and Clinical Psychology, 67, 776-785. 34 Gross , J . J . (1998). The emerging field o f emot ion regulation. Review of General Psychology, 2, 271-299. Gross , J . J . (2002). E m o t i o n regulation: Af fec t ive , cogni t ive, and soc ia l consequences. Psychophysiology, 39, 281-291. Gross , J . J . , & Levenson , R . W . (1997). H i d i n g feelings: The acute effects o f inh ib i t ing negative and posi t ive emotions. Journal of Abnormal Psychology, 105, 95-103. H a m i l t o n , N . A . , Zautra, A . J . , & R e i c h , J . W . (2005). Af fec t and pa in i n rheumatoid arthritis: D o ind iv idua l differences i n affective regulat ion and affective intensity predict emotional recovery f rom pain? Annals of Behavioral Medicine, 29, 216-224. H a m i l t o n , N . A . , Zautra, A . J . , & R e i c h , J . (2007). Ind iv idua l differences i n emot ion processing and react ivi ty to pa in among older w o m e n w i t h rheumatoid arthritis. The Clinical Journal of Pain, 23, 165-172. H e r z o g , A . R . , & Rodgers , W . L . (1981). A g e and satisfaction: Da ta f rom several large surveys. Research on Aging, 3, 142-165. Jensen, M . P . , K a r o l y , P . , & Braver , S. (1986). The measurement o f c l i n i c a l pa in intensity: A compar i son o f s ix methods. Pain, 27, 177-126. Keefe , F . J . , L u m l e y , M . , Ander son , T . , L y n c h , T . , & Carson , K . L . (2001). P a i n and emotion: N e w research directions. Journal of Clinical Psychology, 57, 587-607. Kref t , I., & de L e e u w , J . (1998). Introducing multilevel modeling. Thousand Oaks , C A : Sage Publ ica t ions . M a u s s , I. B . , C o o k , C . L . , & Gross , J . J . ( in press). Au toma t i c emot ion regulat ion dur ing anger provocat ion . Journal of Experimental Social Psychology. M c W i l l i a m s , L . A . , G o o d w i n , R . D . , & C o x , B . J . (2004). Depress ion and anxie ty associated w i t h three pa in condi t ions: Results f rom a nat ional ly representative sample. Pain, 111, 77-83. M e l z a c k , R . (1999). F r o m the gate to the neuromatrix. Pain, Suppl 6, S121 - S I 2 6 . N e w t h , S., & D e L o n g i s , A . (2004). Ind iv idua l differences, m o o d , and c o p i n g w i t h chronic pain i n rheumatoid arthritis: A da i ly process analysis. Psychology and Health, 19, 283-305. N e z l e k , J . B . (2001). M u l t i l e v e l random coefficient analyses o f event- and interval-contingent data i n socia l and personali ty p sycho logy research. Personality and Social Psychology Bulletin, 27, 771-785. Northouse, L . L . , & S w a i n , M . A . (1987). Adjustment o f patients and husbands to the in i t ia l impact o f breast cancer. Nursing Research, 36, 221-225. Paice , J . A . , & Cohen , F . L . (1997). V a l i d i t y o f a ve rba l ly administered numer ic rat ing scale to measure cancer pa in intensity. Cancer Nursing, 20, 88-93. Paquet, C , Kergoat , M . - J . , & Dube , L . (2005). The role o f everyday emot ion regulat ion on pa in i n hospi ta l ized elderly: Insights f rom a prospective wi th in -day assessment. Pain, 115, 355-363. P incus , T . , Summey , J . A . , Sorac i , S. A . , W a l l s t o n , K . A . , & H u m m o n , N . P . (1983). Assessment o f patient satisfaction i n activities o f da i ly l i v i n g us ing a m o d i f i e d Stanford health assessment questionnaire. Arthritis and Rheumatism, 26, 1346-1353. Raudenbush, S. W , & B r y k , A . S. (2002). Hierarchical linear models: Applications and data analysis methods. Thousand Oaks , C A : Sage Publ ica t ions . Raudenbush, S. W . , B r y k , A . S., Cheong , Y . F . , & C o n g d o n , R . T . (2004). HLM 6: Hierarchical linear and nonlinear modeling. L i n c o l n w o o d , I L : Scient i f ic Software International. R o b i n s o n , M . , & R i l e y , J . (1999). Models of pain. M a h w a h , N J : Lawrence E r l b a u m Associates Publ ishers . Schwartz , J . E . , & Stone, A . A . (1998). Strategies for analys ing eco log ica l momentary assessment data. Health Psychology, 17, 6-16. Smith, C , & Wallston, K. (1992). Adaptation in patients with chronic rheumatoid arthritis: Application of a general model. Health Psychology, 11, 151-162. Stone, A . A . , & Neal , J. M . (1984). New measure of daily coping: Development and preliminary results. Journal of Personality and Social Psychology, 46, 892-906. Stone, A . A . , & Shiffman, S. (1994). Ecological momentary assessment ( E M A ) in behavioral medicine. Annals of Behavioral Medicine, 16, 199-202. Suarez, E. C. (2006). Sex differences in the relation of depressive symptoms, hostility, and anger expression to indices of glucose metabolism in nondiabetic adults. Health Psychology, 25, 484-492. Tennen, FL, Aff leck, G. , Coyne, J . , Larsen, R., & DeLongis, A . (2006). Paper and plastic in daily diary research: Comment on Green, Rafaeli, Bolger, Shrout, and Reis (2006). Psychological Methods, 11, 112-118. van Middendorp, H., Geenen, R., Sorbi, M . J . , Hox. J . J . , Vingerhoets, J . J . M . , van Doornen, L. J . P., & Bi j lsma, J . W . J . (2005). Styles of emotion regulation and their associations with perceived health in patients with rheumatoid arthritis. Annals of Behavioral Medicine, 30, 44-53. van Middendorp, H., Geenen, R., Sorbi, M . J . , van Doornen, L. J . P., & Bi j lsma, J . W . J . (2005). Emotion regulation predicts change of perceived health in patients with rheumatoid arthritis. Annals of the Rheumatic Diseases, 64, 1071-1074. Zautra, A . J . , Aff leck, G. G. , Tennen, H., Reich, J . W. , & Davis, M . C. (2005). Dynamic approaches to emotion and stress in everyday life: Bolger and Zuckerman reloaded with positive as well as negative affects. Journal of Personality, 73, 1511-1538. Zautra, A . J . , Marback, J . J . , Rapael, K. G. , Dohrenwend, B. P., Lennon, M . C , & Kenny, D. A . (1995). The examination of myofacial face pain and its relationship to psychological distress among women. Health Psychology, 14, 223-231. 37 Zautra, A. , Smith, B., Affleck, G., & Tennen, H. (2001). Examinations of chronic pain and affect relationships: Applications of a dynamic model of affect. Journal of Consulting and Clinical Psychology, 69, 786-795. 38 Appendix 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0100967/manifest

Comment

Related Items