1 Copyright 2004 by the American Psychological Association, Inc. Journal of Personality and Social Psychology 0022-3514/04/$12.00 DOI: 10.1037/0022-3514.86.2.295 2004, Vol. 86, No. 2, 295–309 Gratitude in Intermediate Affective Terrain: Links of Grateful Moods to Individual Differences and Daily Emotional Experience Jo-Ann Tsang Michael E. McCullough University of Miami Baylor University Robert A. Emmons University of California, Davis Two studies were conducted to explore gratitude in daily mood and the relationships among various affective manifestations of gratitude. In Study 1, spiritual transcendence and a variety of positive affective traits were related to higher mean levels of gratitude across 21 days. Study 2 replicated these findings and revealed that on days when people had more grateful moods than was typical for them, they also reported more frequent daily episodes of grateful emotions, more intense gratitude per episode, and more people to whom they were grateful than was typical for them. In addition, gratitude as an affective trait appeared to render participants’ grateful moods somewhat resistant to the effects of discrete emotional episodes of gratitude. mons & McCullough, 2003). These findings suggest that further Throughout the history of ideas, gratitude has been defined in many ways. Adam Smith (1790/1976) defined gratitude as “the work on gratitude may foster important insights into the links sentiment which most immediately and directly prompts us to between personality, emotion, social life, and psychological well- reward” (p. 68). Similarly, Weiner and Graham (1989) defined being. gratitude as “a stimulus to return a favor to the other and thus reintroduce balance” (p. 403). In recognition of gratitude’s appre- Gratitude at Three Levels of Affective Experience ciative quality, Lazarus and Lazarus (1994) conceptualized grati- tude as one of the “empathic emotions” that reflects recognition or Gratitude, like all affects, manifests itself in several forms with appreciation of an altruistic gift. In the same vein, Emmons and distinct psychological properties. Rosenberg (1998) proposed that Crumpler (2000) wrote, “Minimally, gratitude is an emotional the common forms of affective experience could be structured into response to a gift. It is the appreciation felt after one has been the three hierarchical levels of analysis: affective traits, moods, and beneficiary of an altruistic act” (pp. 56 –57). emotions. Gratitude may serve important functions in human beings’ so- cial and emotional lives. Recent work has suggested that gratitude is a reliable emotional response to the receipt of benefits and that Gratitude as Affective Trait the experience and expression of gratitude may have important effects on behavior in the moral domain (McCullough, Kilpatrick, Rosenberg (1998) placed affective traits at the top of the hier- Emmons, & Larson, 2001). In addition, gratitude is associated archy of affective phenomena. She defined affective traits as positively with a wide variety of measures of subjective well-being “stable predispositions toward certain types of emotional respond- (McCullough, Emmons, & Tsang, 2002). Other work has shown ing” that “set the threshold for the occurrence of particular emo- that gratitude is easily cultivated and is efficacious in kindling tional states” (p. 249). For example, hostility is thought to lower positive emotions generally (Emmons & Crumpler, 2000; Em- one’s threshold for experiencing anger. In the same way, there is a “grateful disposition” that may reduce people’s threshold for experiencing grateful emotions (McCullough et al., 2002). People Michael E. McCullough, Department of Psychology, University of who score highly on measures of gratitude as an affective trait tend Miami; Jo-Ann Tsang, Department of Psychology and Neuroscience, Bay- to experience a high degree of life satisfaction and positive affects lor University; Robert A. Emmons, Department of Psychology, University such as happiness, vitality, and hope. They also experience rela- of California, Davis. tively low levels of negative affects such as resentment, depres- This research was generously supported by a grant from the John sion, and envy (see also Watkins, 2004). Finally, McCullough et Templeton Foundation. al. (2002) found that people who scored highly on measures of Correspondence concerning this article should be addressed to either gratitude also scored higher on measures of prosocial behavior, Michael E. McCullough, Department of Psychology, University of Miami, empathy, forgiveness, religiousness, and spirituality. Among the P.O. Box 248185, Coral Gables, FL 33124-0751, or Robert A. Emmons, Big Five, the grateful disposition seems related most strongly to Department of Psychology, University of California, One Shields Avenue, Agreeableness (positively) and Neuroticism (negatively). Davis, CA 95616. E-mail: [email protected] or [email protected] 295

2 CULLOUGH, TSANG, AND EMMONS C M 296 Therefore, to understand how gratitude may exert many of its Gratitude as an Emotion ’ s social and psychological lives, it presumed effects on people emotions, Gratitude can also manifest itself on the level of would be useful to understand better how people experience grate- which Rosenberg (1998) defined as “ acute, intense, and typically ful moods in daily life. However, a systematic exploration of brief psychophysiological changes that result from a response to a “ gratitude in this intermediate terrain between affective traits and s environment ” meaningful situation in one (p. 250). McCullough ’ ” emotions (Rosenberg, 1998, p. 250) is conspicuously absent from et al. (2001) reviewed data from several studies to conclude that the existing literature. people experience the emotion of gratitude most consistently and strongly when they perceive themselves to be recipients of an Personality and Daily Experience as Determinants of intentionally rendered benefit that is both valuable to the benefi- Gratitude in Daily Mood ciary and costly to the benefactor. In addition, McCullough et al. (2001) posited that the emotion of gratitude has a specific action Because daily mood occupies intermediate terrain between af- “ tendency, which is to contribute to the welfare of the benefactor fective traits and emotional reactions to discrete daily life events, (p. 252). Indeed, grateful emotions (or a third party) in the future ” it is likely that individual differences and daily events work in appear to motivate people to reciprocate the benefits they have concert to determine the extent to which people experience grate- received by rendering further benefits. This action tendency is ful moods on any given day (Rosenberg, 1998). But to which adequately distinct from the action tendencies associated with individual differences might grateful moods be related? To which indebtedness (Greenberg, 1980; see also Gray, Emmons, & Mor- daily events? Moreover, how might individual differences and rison, 2001). For example, people who are asked to describe a daily events work in concert to promote gratitude in daily mood situation in which they felt grateful report having experienced experience? In the preceding paragraphs, we proposed that the more desire to respond generously toward their benefactor and less amount of gratitude that people experience in their moods across desire to avoid the benefactor than do people who are asked to days can be decomposed into (a) stable interindividual differences describe a situation in which they felt indebted (Gray et al., 2001). s typical levels of grateful mood and (b) their ’ reflecting people day-to-day fluctuations around those mean levels. Both of these Gratitude as a Mood: Intermediate Affective Terrain components, we argue, are important aspects of mood, even though this decomposition implies that the components will have which wax and wane, ” “ Rosenberg (1998) considered moods, “ different correlates. Namely, the stable individual-differences (p. 250), as subordinate to ” fluctuating throughout or across days component can only be correlated with characteristics of persons affective traits but as superordinate to discrete emotion episodes. (i.e., personality and affective traits), whereas the day-to-day fluc- Moods comprise a stable component that is attributable in part to tuations can only be correlated with characteristics that also fluc- individual differences among persons (i.e., some people experi- tuate over time (e.g., day-to-day changes in gratitude-relevant life ence more gratitude than others as the affective background of events, or the interactions of personality traits with characteristics their day-to-day lives, just as some people are more prone to, say, that fluctuate over time). sadness or anxiety as the affective background of their daily lives). However, mood also varies across days as a function of the events that occur to people each day and their discrete emotional reactions Which Traits Are Associated With Mean Levels of to those events. Thus, mood may be an important nexus where Gratitude in Daily Mood? affective traits and short, intense experiences of discrete emotions might be expected to exert some of their most interesting effects, We anticipate that people ’ s mean levels of gratitude in their both additively and jointly. daily moods are reliably associated with measures of gratitude as Moods are subtle and less accessible to conscious awareness a disposition or affective trait (e.g., McCullough et al., 2002), than are emotions (i.e., one is less likely to be aware of gratitude because the disposition toward gratitude is presumably a as a mood than as an emotion). Despite their subtlety relative to personality-based proneness to experience grateful affect (includ- emotions, however, moods are important because they are ex- ing moodlike manifestations of affect). Although this proposition pected to have broad, pervasive effects on consciousness that may seem trivial, demonstrating a link between the affective traits emotions simply cannot because of their relatively short duration that supposedly predispose people to experience grateful moods (Rosenberg, 1998). To be sure, emotions have powerful effects on and the stable, traitlike component of those grateful moods them- consciousness, but these effects are usually goal directed and s (1998) hierarchical ’ selves is an important test of Rosenberg dissipate after the goals that are associated with the motivation `-vis gratitude). More- model of affective experience (at least vis-a (e.g., the motivation to repay or praise a benefactor, which may be over, the inability of such measures to predict gratitude in daily the primary goal of grateful emotions; McCullough et al., 2001) mood would cast doubt on the construct validity of those measures have been fulfilled. In contrast, the relatively long duration of and on gratitude theory. moods might allow them to influence information processing, ’ s mean Other personality and affective correlates of people physiological reactivity, and other psychological phenomena over levels of gratitude in their daily moods are probably a subset of the relatively long arcs of time. Indeed, many of the presumed social traits that McCullough et al. (2002) found to be correlated with effects of gratitude, such as readiness to be helpful and supportive self-reports and peer reports of gratitude as an affective trait. to others (McCullough et al., 2002), as well as psychological McCullough et al. demonstrated that people with high scores on effects of gratitude, such as better coping with stress, are probably measures of life satisfaction and trait positive affect score higher caused not by gratitude as an emotion, because the duration of on measures of gratitude as an affective trait. This may be in part emotions is far too short, but rather, by gratitude as a mood. because such traits have a common dispositional core (e.g., high

3 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 297 their own discrete emotional responses to these daily events. In Extraversion/low Neuroticism) that inclines (or, in the case of other words, individuals who score high on measures of gratitude negative affect, disinclines) people to experience positive affects as an affective trait should report many discrete events that cause of all sorts, including gratitude. McCullough et al. also found that them to feel grateful, should experience gratitude to a large number gratitude as an affective trait was correlated positively with self- of people, and should feel more intense episodes of grateful report measures of prosocial traits such as empathy and perspec- emotion in response to these gratitude-relevant daily life events. tive taking and was correlated negatively with envy, perhaps According to the conductance hypothesis, the disposition toward partially because these traits share a common dispositional sub- gratitude fosters causal dependence between emotions and moods: strate (e.g., Agreeableness). For dispositionally grateful people, their emotional reactions to McCullough et al. (2002) demonstrated further that measures of so that ” upward gratitude-relevant daily events are conducted “ religious involvement and spirituality were correlated with grati- their more pervasive and long-lasting moods throughout the day tude as an affective trait. These latter correlations may reflect the are colored by their daily emotional experience. fact that people who are highly religious or spiritual tend to Stated another way, according to the conductance hypothesis, perceive positive circumstances in their lives that are not caused by grateful emotions put dispositionally grateful people into a grateful human action (e.g., eyesight, a sunny day) as nevertheless resulting mood. Conversely, for people who are lower in the disposition from the intentional behavior of a benevolent moral agent (i.e., toward gratitude, gratitude-relevant daily life events (and emo- God or a higher power). In contrast, less religious or spiritual tional reactions of gratitude) are not effectively conducted upward individuals might attribute these same events as due to chance and to influence mood. For them, discrete emotional experiences of therefore be less inclined to feel grateful in response. Insofar as the gratitude do not lead to grateful moods as strongly or reliably. correlations between gratitude and religiousness or spirituality The conductance hypothesis is reminiscent of Larsen and Kete- reflect real differences in the affective experiences of more versus laar ’ s (1991) findings that affective traits such as Extraversion and less religious or spiritual people, these correlations should also Neuroticism have specific effects on people ’ s reactions to emo- manifest themselves when gratitude is measured at the level of tionally valenced life events. Larsen and Ketelaar found that Ex- daily mood. traversion appears to moderate the link between laboratory events known to induce positive affect and ratings of positive emotion in What Discrete Daily Experiences Are Associated With response. That is, for extroverts, the effects of positive mood Gratitude in Daily Mood? inductions were stronger than they were for introverts. Similarly, Neuroticism appeared to moderate the relationship between neg- s daily moods is built To the extent that gratitude in people ’ ative events and negative emotions. In similar work, Suls, Martin, ” ’ on the basis of people s daily gratitude- “ from the ground up and David (1998) found that highly agreeable people experience relevant experiences and their short-term emotional reactions to more negative affect in response to conflict events than do less them, people who experience more grateful mood on a given day agreeable people, although there was no difference in the affective than is typical for them should also report (a) more events on that responses of more agreeable and less agreeable people to noncon- day for which they are grateful, (b) more people to whom they felt flict events in daily life. grateful, and (c) more intense gratitude on that day in response to these discrete daily events than is typical for them. The Resistance Hypothesis How Does the Grateful Disposition Interact With Daily According to the resistance hypothesis, for people who are Emotional Experience? dispositionally prone to feel grateful, the amount of gratitude in their daily moods is determined so thoroughly by personality In her depiction of the relationships among affective traits, processes that their moods are resistant to the effects of gratitude- moods, and emotions, Rosenberg (1998) focused almost exclu- relevant daily life events (e.g., experiencing many discrete sively on the bivariate associations among these three levels of gratitude-eliciting events; experiencing gratitude to a large number affect — for example, how affective traits and emotions might of people) and their discrete emotional responses to these daily individually influence mood. However, individual differences and events (i.e., feeling intense episodes of grateful emotion in re- daily experiences might also operate interactively to determine sponse). Conversely, according to the resistance hypothesis, peo- grateful moods. Gratitude-relevant events (and fleeting emotional ple with a weaker disposition toward gratitude experience grati- reactions to them) may have a different influence on daily moods tude in their daily moods only insofar as they experience high for people who are strongly disposed to experience gratitude in psychological events that typically elicit – numbers of the social comparison with people who are weakly disposed to experience gratitude and strong amounts of grateful emotion in response to gratitude. We propose two rival hypotheses regarding how the those daily events. In this view, less dispositionally grateful peo- grateful disposition and daily experience might interact to deter- ’ s daily grateful moods are more dependent on discrete emotion ple mine grateful moods (see also Affleck, Tennen, Urrows, & Hig- episodes than are those of more dispositionally grateful people. gins, 1992, who posed a similar doublet of rival hypotheses). In support of the resistance hypothesis, Affleck et al. (1992) found that among patients with rheumatoid arthritis, people with The Conductance Hypothesis the highest levels of Neuroticism showed the weakest links be- tween their daily pain ratings and negative affect in daily mood. The conductance hypothesis states that the moods of people who However, most research to date that bears on the conductance and are highly disposed toward gratitude are particularly responsive to resistance hypotheses would seem to support the conductance the emotional effects of gratitude-relevant daily life events and

4 CULLOUGH, TSANG, AND EMMONS C M 298 The 5-item Satisfaction With Life Scale (SWLS; Satisfaction with life. hypothesis (e.g., Larsen & Ketelaar, 1991; Suls et al., 1998), Diener, Emmons, Larsen, & Griffin, 1985) assesses the cognitive compo- although few if any studies have examined these two hypotheses in In most ways my life is close to nent of subjective well-being. Items (e.g., “ the context of discrete affective traits, emotions, and moods such ” ) are rated on a 7-point scale (1  ideal ;7  strongly strongly disagree as gratitude (or anger, or fear, etc.). retest correlation – ). In previous work, the SWLS had a 2-month test agree coefficient of  r  .87 (Diener et al., 1985). .82, and coefficient alpha Overview of the Studies Campbell Well-Being Scale. The Campbell Well-Being Scale consists of nine semantic differential scales (e.g., boring – interesting, miserable – In the remainder of the present article, we describe two studies – hopeful ) that provide an overall index of general enjoyable, discouraged ’ s daily in which we explored the nature of gratitude in people well-being (Campbell, Converse, & Rodgers, 1976). The scale correlates moods. In Study 1, we began by examining the relationship of with other measures of well-being and has acceptable internal consistency ’ s typical levels of gratitude in daily mood with a variety of people reliability (Beckie & Hayduk, 1997). personality and affective traits. In Study 2, we extended Study 1 Optimism. The widely used Life Orientation Test (LOT; Scheier, Carver, & Bridges, 1994) is an eight-item scale for assessing dispositional and examined the relationship of the day-to-day fluctuations in optimism. Scheier et al. (1994) reported an internal consistency reliability grateful mood with day-to-day fluctuations in gratitude-relevant  – retest stabilities ranging from .56 to .79 across four .82 and test  of daily life events (e.g., experiencing many discrete events that cause time periods. gratitude; experiencing gratitude to a large number of people), and Participants also completed the Center for Ep- Depressive symptoms. people ’ s discrete emotional reactions (i.e., gratitude) in response to idemiologic Studies Depression Scale (Radloff, 1977). Participants used a those daily events. In Study 2, we also evaluated the conductance  Rarely or none of the time [less than 1 day] 4-point Likert-type scale (1 and resistance hypotheses. most or all of the time [5–7 days]  and 4 ) to indicate how often during In conducting these two studies, we worked with two very the last week they experienced each of 20 affective and somatic symptoms different samples of participants. The first study involved patients that characterize major depressive episodes. Total scores are the sums of at a university hospital who had neuromuscular disorders, whereas scores from all 20 items. Radloff (1977) reported test – retest stabilities  r .67 (4 weeks) to  r ranging from .32 (12 months). Internal consis- the second study involved undergraduate students. The use of these tency was estimated as  .90 (Radloff, 1977).  two samples was helpful for examining the extent to which our Positive affectivity (PA) and negative affectivity (NA). The Positive findings about the nature of gratitude in daily mood generalized and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988) across at least two interesting facets of human diversity (e.g., measures general tendencies to experience positive (e.g., ” proud “ ) and adults vs. late adolescents; people with chronic illness vs. physi- guilty negative (e.g., “ ” ) affect. Participants used a 5-point Likert-type scale cally healthy people). (1  ;5  extremely ) to indicate how well each very slightly or not at all “ of 20 adjectives described how they generally feel. ” Coefficient alphas of the positive and negative scales range in the mid- to upper .80s (Watson et Study 1 al., 1988). The Big Five Inventory (BFI). The BFI (John, Donahue, & Kentle, Method 1991) consists of 44 brief descriptive phrases that are prototypical markers of the Big Five factors of personality: Agreeableness, Conscientiousness, Participants Extraversion, Neuroticism, and Openness. Alpha reliabilities and test – 49 Participants were 96 adults (67 women and 29 men; mean age  retest reliabilities for the five subscales range from .80 to .90 (John & years, range  77) with either congenital or adult-onset neuromuscular – 22 Srivastava, 1999). diseases (NMDs). These participants were obtained from a larger group of Spiritual transcendence. The Spiritual Transcendence Scale (Pied- 153 participants drawn from a mailing list compiled by the University of mont, 1999) is a 24-item scale consisting of subscales for assessing three California, Davis, Medical Center Neuromuscular Disease Clinic. The dimensions of spirituality: prayer fulfillment, universality, and connected- majority had one of three NMDs: post-polio, Charcot-Marie-Tooth, or ness. In the present study, we used the total scale score as a measure of fascioscapulohumeral (for more information about neuromuscular disease, spirituality. see Fifty-four percent of the participants were married, 42% had college or postgraduate degrees, and their mean annual Diary Measures: Gratitude in Daily Mood income was between $15,000 and $25,000 (for more details about this sample, see Emmons & McCullough, 2003). Approximately 1 year after completing the prediary questionnaires, 98 of the original 153 participants were enrolled in a 21-day mood diary study. Measures: Prediary Questionnaires Participants rated the intensity with which they felt each of a variety of emotions each day for 21 consecutive days. Participants were instructed to Approximately 1 year prior to completing the daily diaries, the 153 “ Indicate to what extent you felt this way during the day today ” using a participants completed a 28-page survey including the self-report measures very slightly or not at all ). 5-point Likert-type scale (1 ;5  extremely  described below. They received $40 compensation for completion of this daily moods via their ’ We measured the amount of gratitude in participants 28-page survey. mean score on three gratitude-related emotion words ( grateful, thankful, The GQ-12, an early Gratitude Questionnaire: 12-item form (GQ-12). ) as in Emmons & McCullough (2003). Across the 21 and appreciative version of the Gratitude Questionnaire 6-item form (GQ-6; McCullough et days, the mean internal consistency for this three-item composite was   “ I feel a profound sense of al., 2002), consists of 12 self-report items (e.g., .92.  ) that participants endorse on a 7-point Likert-type scale (1 ” appreciation strongly agree  ;7 strongly disagree ) to indicate how well they describe In this sample, the GQ-12 had your feelings about your life as a whole. ” “ Postdiary Measure of the Disposition Toward Gratitude  an internal consistency reliability of  .83, and was correlated with the r Approximately 1 month after completing the 21-day diaries, participants  .72 GQ-6 (which participants completed 1 year later; see below) at completed a second packet of questionnaires. This packet of questionnaires .001). ( p 

5 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 299 persons differences in initial status and linear change). These models took included the GQ-6 (McCullough et al., 2002), which is a self-report the form measure of the frequency and intensity with which people experience I have so much in life to be thankful for, “ ” gratitude. Sample items include  X (2) .  u     0j 00 0j 01 01j and I am grateful to a wide variety of people. ” Items are endorsed on a “ ;7  strongly agree ).  7-point Likert-type scale (1 strongly disagree - Equation 2 specifies estimation for  , which captures individual differ 0j McCullough et al. (2002) reported internal consistency reliabilities in the ences in initial status (i.e., mean levels of gratitude in daily mood across the range of  .80. Confirmatory factor analyses showed that the measure  21-day period).  the mean initial status estimate on the three-item  00 has a robust one-factor structure and is distinct from many measures of gratitude composite for the entire sample,  the strength of the  01 subjective well-being. relationship between the between-persons differences in mean levels of gratitude in daily mood and another measured between-subjects variable X , and X is a residual reflecting  Person j ’ s score on X u , 01 0j 01j 01 Procedure between-persons differences in mean levels of gratitude in daily mood that are not accounted for by  and the between-subjects predictor variables. 00 Participants were mailed 21 daily recording forms, instructions, and Using this modeling approach, we conducted a series of univariate analyses business reply envelopes for mailing their forms directly back to the in which several personality and affective traits were used individually as researchers. They were told that their daily ratings were meant to summa- ’ between-subjects predictors of mean levels of gratitude in participants rize the day as a whole, and therefore they should try to complete them as daily moods. We expressed these relationships with t statistics and accom- late in the day as possible but before being too sleepy to complete them panying p values and also converted the t values to effect size r s, calculated accurately. The daily form took approximately 5 min to complete each as evening. Participants were asked to mail in their forms once a week. Finally, they were told that should they forget to fill out a form, it was 2 1/ 2 /  t r t  (3) n .  2   better to omit the form for that day rather than to complete it from memory. Participants were paid $20 if they completed all of the forms; they were For an explanation of the derivation of this formula, see Hunter and paid $15 if they failed to complete all 21 forms. Of the 98 participants Schmidt (1990, p. 272). initially recruited, 96 of them returned usable mood diary data. These data collection methods and compliance rates compare favorably with those Results and Discussion from other recent studies (e.g., Gable, Reis, & Downey, 2003). All but 3 of the 96 participants who returned usable mood diary data also completed the Means, standard deviations, and reliabilities for major study postdiary packet of questionnaires. For the 3 individuals who did not, we variables appear in Table 1. – maximization imputed values for their GQ-6 scores using the expectation routine. Longitudinal Trajectories of Gratitude in Daily Mood Ratings Analyses Per the linear change model specified in Equation 1, participants We used a two-level modeling strategy with the HLM 5 statistical began the study feeling above the midpoint on gratitude (coeffi- software package (Raudenbush, Bryk, Cheong, & Congdon, 2000). HLM  0.09). Because the scores on the three-item  SE 3.47, cient enabled us to fit simultaneously both (a) within-person longitudinal models and (b) between-person models that accounted for individual differences in the parameters of the within-person models. Table 1 Consistent with our view that grati- Within-person (Level 1) models. tude in daily mood comprises (a) a stable component that is consistent for Means, Standard Deviations, and Reliability Estimates for each individual across the 21-day period but that varies across persons and Major Variables, Study 1 (b) a component reflecting day-to-day variations in the amounts of grati- tude that people experience in their daily moods, we decomposed people s ’ Variable MSD Reliability Range 21 daily gratitude mood ratings into within-person (or Level 1) models that a 5 3.47 0.85 .84 Gratitude in daily mood 1 – took the form 1 GQ-12 .83 0.65 5.68 7 – GQ-6 0.96 5.93 .87 – 1 7 (1)  r   day  ,    GMR 0j ij ij 1j 1.40 4.01 7 – 1 Life satisfaction .85 0.85 Optimism 1 – 5 3.65 .83 where GMR Person j  s gratitude mood rating (i.e., their score on the ’ ij 1 – 7 Well-being 1.08 4.90 .90 three-item composite) on Day i;  ’ Person j  s mean level of gratitude 0j .87 0.63 3.47 5 – 1 Positive affect in daily mood across the 21 days (centered on the first day of the study); .90 0.72 1 Negative affect 2.09 5 –  s the effect of linear change (measured in days) on Person j  (day) ’ 1j 0 .85 0.39 1.72 Depression 4 – r - gratitude mood ratings; and reflects an occasion-specific residual vari ij 1 Spiritual transcendence 5 .91 0.47 3.60 – ance in GMR , or the extent to which Person j experienced more or less Big Five ij gratitude on Day i than was typical for him or her (Nezlek, 2001). Although 0.56 3.84 5 – .78 Openness 1 5 3.98 0.59 .78 Conscientiousness 1 – we had no theoretical reason to expect change in the amount of gratitude 0.67 2.96 5 – 1 Extraversion .81 daily moods over the 21 days, we have found in other work ’ in participants .66 0.46 4.08 5 – Agreeableness 1 s scores ’ (McCullough, Emmons, Kilpatrick, & Mooney, 2003) that people – Neuroticism 1 5 2.81 0.82 .86 on some variables decay with repeated assessments, perhaps because of habituation to the rating task over the 21-day period (the rate of change was  Gratitude Questionnaire, 12-item form. GQ-6  Note. N  96. GQ-12 also free to vary across participants). Gratitude Questionnaire, 6-item form. a Between-person (Level 2) models. The between-person (or Level 2) Initial status parameter only. Reliability calculated as percentage of true models were attempts to account for individual differences in the Level 1 parameter variance per Bryk and Raudenbush (1992). All other reliabilities beta coefficients (i.e., the random coefficients that accounted for between- ’ s alpha. estimated with Cronbach

6 CULLOUGH, TSANG, AND EMMONS C M 300 Table 2 gratitude composite ranged from 1 5, a score of zero was not – Correlations of Personality and Affective Traits With Initial possible. As a result, null hypothesis tests for this coefficient are Status Parameters for Gratitude in Daily Moods, Study 1 not informative, so we do not report them here. Over the course of the study, participants reported feeling r Coefficient SE t (94) Personality measure SE slightly less grateful each day (coefficient 0.02,   0.004),  4.55, p  .001. In other words, scores on the three-item (95) t Affective trait measures of gratitude measure of gratitude in daily mood decreased, on average, .02 GQ-12 .37 0.72 0.19 3.87*** units per day. These group trends notwithstanding, participants GQ-6 (postdiary) 0.45 0.12 .37 3.85***  manifested considerable variance in both initial status (variance Life satisfaction and affectivity .001) and linear change (variance .001).   p 0.73, 0.001, p  The two Level 1 parameters (initial status and linear change) Satisfaction With Life 0.07 3.20** .31 0.22 s reports of gratitude ’ accounted for 64% of the variance in people Well-Being 0.10 2.74** .27 0.27 Optimism 0.40 0.17 2.38* .25 in their daily moods. Because participants varied considerably on   .22 Depression 2.21* 0.68 0.31  these two parameters, it was reasonable to search for variables that .39 Positive Affectivity 0.61 0.15 4.11*** could account for individual differences in both initial status and 0.18  Negative Affectivity  1.38  .14 0.13 linear change. Almost without exception, individual differences in Religious/spiritual variables linear change were not significantly related to the affective and personality traits examined herein, so we do not report those .33 3.32** 0.18 0.60 Spiritual Transcendence associations in the present article. There was also no gender The Big Five  difference in initial status or linear change estimates ( p s .50). 0.09 Openness 0.69 .07 0.14 Conscientiousness 0.14 1.02 .11 0.14 Correlations of Dispositional Measures With Typical 0.54 4.01*** Extraversion .38 0.14 0.40 1.61 .16 Agreeableness 0.25 Levels of Gratitude in Daily Mood .19  1.87  0.11 0.21  Neuroticism First, we examined whether self-report measures of gratitude as Grati-  Gratitude Questionnaire, 12-item form. GQ-6 GQ-12 Note.  an affective trait were associated with the mean amounts of grat- tude Questionnaire, 6-item form. .001.  p .01. ***  p * p  .05. ** ’ daily moods (i.e., as indexed by their Level itude in participants 1 initial status parameter estimates, as with all analyses reported 0.72,  below). Both the GQ-12 as an affective trait (coefficient  (94) p t 0.19),  SE   .001, and the GQ-6 (coefficient 3.87, Associations of Gratitude as an Affective Trait With (94)  3.85, p  .001, were positively and SE 0.45, 0.12),  t Gratitude in Daily Mood, Controlling for Higher Order significantly related to individual differences in the mean levels of Personality and Affective Traits gratitude in people s  .37). People ’ s daily moods (effect size r who perceived themselves as highly disposed to experience grat- Given the pervasive influences of Extraversion PA and – itude did indeed experience higher levels of gratitude in their daily Neuroticism NA on social and emotional experience (Gross, Sut- – moods throughout the 21 days. ton, & Ketelaar, 1998; Larsen & Ketelaar, 1991) and their asso- ciations with gratitude herein and in previous work (McCullough et al., 2002), we estimated the associations of our two measures of Correlations With Other Personality and Affective Traits gratitude as an affective trait with mean levels of gratitude in daily mood while statistically controlling for the Big Five as well as trait Several other personality and affective traits predicted mean PA and NA. When we included these additional variables with the ’ daily mood ratings. As can be levels of gratitude in participants GQ-6 measure of gratitude as an affective trait, the association of seen in Table 2, people with high levels of life satisfaction, the GQ-6 with individual differences in mean levels of gratitude in well-being, optimism, and PA reported higher mean levels of daily mood across the 21 days remained statistically significant gratitude across the 21 days than did their counterparts who had t  (coefficient 0.33, SE  0.12),  (87)  2.62, p .009 (effect size lower scores on these traits. Conversely, people high in self- r  .26). Similarly, the association of the 12-item measure of reported depressive symptoms (but not NA) reported lower mean gratitude as an affective trait with individual differences in mean levels of gratitude in their daily moods across the 21-day period levels of gratitude in daily mood remained significant (coeffi- than did people who were lower in self-reported depressive 0.59, SE  0.18), t (87)  3.29, p  .001 (effect size r cient   symptoms. .32). Thus, the association of gratitude as an affective trait with the As expected, spiritual transcendence was positively correlated typical amount of gratitude that participants experienced in their with the mean levels of gratitude in participants ’ daily moods daily mood ratings could not be explained solely in terms of the  r .01 (  3.32,  p t 0.60),  (coefficient .33). Among the (94) Big Five, trait PA, and trait NA. Big Five (Openness, Conscientiousness, Extraversion, Agreeable- 0.54), ness, and Neuroticism), only Extraversion (coefficient  Summary .001 ( p 4.01,   (94) t r  .38) was significantly correlated with ’ the mean levels of gratitude in participants daily moods across the In Study 1 we examined the associations of several person- ality and affective traits with individual differences in the 21-day period.

7 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 301 amounts of gratitude that people typically experienced in daily Measures: Prediary Questionnaires mood. People who scored high on trait measures of positive Prior to completing the daily mood diaries, participants also completed emotion and subjective well-being experienced higher mean a packet of questionnaires containing several measures of personality and levels of gratitude in their daily moods than did people who affective traits. scored lower on such measures. Among the Big Five, Extra- Measures from Study 1. As in Study 1, participants completed the version appeared to be the strongest predictor of the typical GQ-6, the SWLS, the PANAS, and the BFI. Disposition toward gratitude. Participants also indicated their general amounts of gratitude in people ’ s daily moods. This relationship disposition toward gratitude by indicating the extent to which they gener- between the Big Five and grateful mood in some ways mirrors ally felt each of three gratitude-related emotions (using the three adjectives the relationship between gratitude as an affective trait and the grateful, thankful,  and appreciative ), using a 5-point Likert-type scale (1 Big Five. McCullough et al. (2002) found associations between  ;5 extremely ). In previous work, this three-item very slightly or not at all the Big Five dimensions Extraversion, Agreeableness, and Neu- r adjective measure was highly correlated ( .65 before correcting for  roticism and measures of gratitude as an affective trait. The measurement error; .75 after correcting for measurement error via  r additional associations of Agreeableness and Neuroticism structural equation modeling) with the GQ-6 (McCullough et al., 2002). found with gratitude as an affective trait may be due to differ- Measures of religiousness and spirituality. Participants completed sev- ences between gratitude as an affective trait and gratitude as a eral measures of religiousness and spirituality. First, they completed a mood or to the unique nature of the sample used in the present Not at all interested   ;9 single-item measure of religious interest (1 ). They also completed an eight-item measure of Extremely interested study (adults with neuromuscular diseases). general religiousness (for details on these eight items, see McCullough et People with high levels of spiritual transcendence (Piedmont, al., 2002, p. 116). This composite measure had an internal consistency 1999) also experienced higher mean levels of gratitude in their  reliability estimate of  .88. daily moods than did people lower in spiritual transcendence. s (1967) Intrinsic We measured religious orientation using Allport and Ross ’ s (2002) findings ’ These findings corroborate McCullough et al. s Quest Scale (Batson and Extrinsic Religious Orientation Scales and Batson ’ regarding the spiritual correlates of gratitude, here extended to the & Schoenrade, 1991). The intrinsic scale measures the extent to which religion realm of daily mood. In addition, Study 1 shows that two measures ” is an individual master motive ’ in life, whereas the extrinsic scale measures s “ ’ s of gratitude as an affective trait, which should partially set people ’ s religion as a means for reaching other, nonreligious ends such the use of one typical levels of gratitude in daily mood (Rosenberg, 1998), were as comfort or status (Allport & Ross, 1967). The quest scale measures the indeed correlated with mean levels of gratitude in daily mood, with extent to which an individual conceptualizes religion as an open-ended search. s These scales essentially measure the functional nature of an individual ’ effect size correlations of r  .37. religion; therefore, it is necessary to include only those participants that The findings of Study 1 led to four other questions. First, categorize themselves as religious to begin with. For this reason, we conducted because Study 1 involved adults with chronic physical health analyses involving these scales only with the 85 participants that indicated at problems, we wondered whether the results from Study 1 would “ least a moderate interest in religion (marked 3 or higher on the item How emerge also in a sample of basically healthy university students.  to 9 Not at all  on a scale from 1 ” interested are you in religion? Second, the correlations of the Big Five with gratitude in daily ). The intrinsic, extrinsic, and quest scales had reliability estimates of Extremely mood in Study 1 differed somewhat from what we have found in  .89, .69, and .83, respectively, in this sample.  earlier work in which we assessed gratitude as an affective trait In addition, participants completed a 15-item version of the Self- (McCullough et al., 2002). We wondered if this divergence from Transcendence subscale of the Character and Temperament Inventory previous results was due to differences in the assessment of grat- (Cloninger, Svrakic, & Przybeck, 1993). This measure assesses three aspects of spirituality: Self-forgetful versus self-conscious experience, itude or to the nature of the sample used in Study 1. Third, we transpersonal identification versus self-isolation, and spiritual acceptance wished to explore the association of gratitude in daily mood with versus rational materialism. Kirk, Eaves, and Martin (1999) developed a a wider variety of spiritual and religious constructs. 15-item form of the full-length 33-item subscale. Cloninger et al. (1993) Fourth, we wanted to explore how gratitude in daily mood might reported alpha in the low .70s for the 33-item version. In McCullough et be shaped not only by personality and affective traits but also by  s (2002) study, internal consistency of the 15-item version was ’ al. .86.  the gratitude-inducing events that people experience and their Items were endorsed on a 4-point Likert-type scale (1 agree   ;4 emotional reactions to them. In this context, we also wished to test disagree ). (Analyses using the Self-Transcendence subscale included all the conductance and resistance hypotheses. We investigated these participants regardless of their religious interest.) questions in Study 2. The tendency to experience empathy for others Dispositional empathy. was measured with the widely used Empathic Concern and Perspective- Taking subscales of the Interpersonal Reactivity Index (Davis & Oathout, Study 2  1987). These subscales have adequate internal consistency (  s .73 and .71, respectively; Davis & Oathout, 1987). Envy. We measured envy with R. H. Smith, Parrott, Diener, Hoyle, and Method Frankly, “ s (1999) eight-item Dispositional Envy Scale. Items (e.g., Kim ’ ” the success of my neighbors makes me resent them ) were rated on a Participants strongly agree 5-point Likert-type scale (1  strongly disagree ;5  ). R. H. Smith et al. reported internal consistency reliability estimates in the Participants were 112 students (84 women, 27 men, 1 unrecorded) in an r .80.  – range of   .83 – .86 and 2-week test retest stability of undergraduate psychology course at Southern Methodist University (age 18  range  20.54). Most participants were European 44 years, – M Measures From Daily Diaries American (85). Smaller numbers of individuals classified themselves as Gratitude-relevant daily events and discrete emotional reactions to Latino/a American (13), African American (7), and Asian American (4), with 3 individuals declining to state their ethnicity. Participants received them. Participants used daily diaries to report information regarding course credit for participating. discrete gratitude-relevant episodes each day. Participants described (typ-

8 CULLOUGH, TSANG, AND EMMONS C M 302 scores on the three time-dependent or within-subjects covariates, each of ically with a single phrase) up to eight situations in which they felt grateful which was centered on each person ’ s mean (Nezlek, 2001). These analyses each day. Next to the description of each gratitude-eliciting situation, enabled us to determine whether the associations between the three time- participants listed the specific people to whom they felt grateful in the dependent covariates and gratitude in daily mood varied as a function of corresponding situation. Then, they rated the intensity of the gratitude ’ scores on the GQ-6, which is a measure of gratitude as an participants Somewhat elicited by each situation on a 3-point Likert-type scale (1  affective trait. ). From these data, we were able to derive  Extremely grateful ;3 grateful three measurements of discrete gratitude-inducing events in participants ’ daily lives and their emotional reactions to these events: (a) gratitude Results and Discussion frequency (the number of discrete daily situations that elicited gratitude), (the number of people to whom participants felt (b) gratitude density Means, standard deviations, and reliabilities for major study mean episodic gratitude intensity grateful each day), and (c) (the mean variables appear in Table 3. gratitude intensity rating for all of the discrete situations that elicited gratitude each day). Longitudinal Trajectories of Gratitude in Daily Mood Gratitude in daily mood. In the next section of each daily diary entry, Ratings participants were instructed to rate 38 emotion words including three and appreciative ) to indi- gratitude-related adjectives ( grateful, thankful, First, as in Study 1, we decomposed people s 14 daily ratings of ’ the extent to which you felt each of these emotions today. ” As in cate “ gratitude in daily mood into initial status and linear change com- Study 1, scores on these three adjectives were averaged and used as a ponents. The mean initial status was above the midpoint on the 1 5 – measure of gratitude in their daily moods. Across the 14 days of the study,  3.75, 0.07). In addition, there was a scale (coefficient  SE the mean internal consistency reliability of this three-item composite was small mean reduction in daily gratitude as the study progressed .90.   (111) .001.  p 5.91,  t 0.01),  SE 0.04,  (coefficient These general trends notwithstanding, there were substantial indi- Procedure vidual differences among participants in both the initial status Participants received a questionnaire packet and daily diaries at the 0.002, (variance  0.48, p   .001) and linear change (variance beginning of class during an undergraduate course in the spring semester of .001) parameters. These two parameters accounted for 48% of p  2001. Participants were instructed to complete the questionnaire packet as ’ the variance in the amount of gratitude in people s daily mood soon as possible and to complete one diary entry each night for 14 days. reports. We proceeded to look for personality variables that might Materials were collected each class session (four times total) during the explain individual differences in these initial status parameters. 2-week period. There was no gender difference for the initial status parameter or s  linear change parameter ( p .30). Analyses Correlations of Dispositional Measures With Typical We conducted multilevel models as in Study 1 (see Equations 1 and 2) Levels of Gratitude in Daily Mood ’ to study the personality and affective trait correlates of participants mean levels of gratitude in daily mood across the 14 days. In addition, we In a series of Level 2 models, we estimated the correlations of examined the within-subjects correlations of three time-varying covariates several personality and affective traits with individual differences (gratitude frequency, gratitude density, and mean episodic gratitude inten- in the mean amounts of gratitude in participants ’ daily moods sity) with the three-item measure of gratitude in daily mood. These within- across the 14 days. subjects correlations allowed us to estimate the extent to which people who had more gratitude in their daily moods on a given day than was typical for Correlations with the measures of gratitude as an affective trait. them also had (a) more discrete events that caused them to feel grateful on ’ Table 4 shows that participants scores on the GQ-6 were corre- that day (gratitude frequency), (b) more people to whom they felt grateful lated with individual differences in mean levels of gratitude in the on that day (gratitude density), and (c) more intense emotional responses to 0.10), (110) daily mood ratings (coefficient  0.57, SE   t 5.90, those gratitude-eliciting episodes they experienced that day (mean episodic r .001 (  p .49), as was the three-item adjective measure of  gratitude intensity). These models took the form 0.57,  gratitude as an affective trait (coefficient 0.10),  SE p .47). These results are very similar (110)  5.59, t  .001 ( r  GMR       gratitude-inducing events  .    day  r ij 2j 1j 0j ij to those obtained in Study 1. (4) Consistent with the results Life satisfaction and affectivity. from Study 1, people with high levels of life satisfaction, happi- ’ Equation 4 decomposes Person j s GMR on Day i into a mean level across ness, and PA had higher mean levels of gratitude in their daily  the 14 days (centered on the first day of the study, ); a linear effect for 0j time,  moods. Also, as in Study 1, NA was not related to mean levels of (day); an effect due to the number of gratitude-inducing events 1j specified in Person j ’ s daily diary on Day i,  (gratitude-inducing events); gratitude in daily mood (see Table 4). 2j r and a residual . In other words, this equation allows one to determine ij Similar to Study 1, people Religious and spiritual variables. whether Person j experienced more gratitude in daily mood than is per- who reported high levels of spirituality had higher mean levels of sonally typical on days when he or she experienced more gratitude- gratitude in their daily moods, as did people higher in religious inducing events than is personally typical, controlling for typical level of interest, general religiousness, and intrinsic religious orientation. gratitude in daily mood. We conducted similar within-person models for However, extrinsic religious orientation and quest religious orien- each of the three above-mentioned time-varying covariates and then ex- tation were not significantly correlated with mean levels of grati- amined them jointly as simultaneous within-subjects predictors. daily moods. These findings suggest that tude in participants ’ Finally, we tested the conductance and resistance hypotheses by exam- people who are involved in conventional forms of religiousness, s scores on the GQ-6 (a ’ ining cross-level interactions between people between-subjects covariate, centered on the sample mean) and their daily especially people for whom religion is a fundamental organizing

9 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 303 Table 3 Means, Standard Deviations, and Reliability Estimates for Major Variables, Study 2 Variable Range MSD Reliability Daily measures (initial status parameters only) a Gratitude in Daily Mood 1 – 5 3.75 0.69 .76 a 8 .88 1.47 2.27 Gratitude frequency – 0 a Gratitude density 1  3.03 2.16 .83 a Mean episodic gratitude intensity 1 – 3 2.10 0.58 .48 Gratitude as an affective trait – 1 GQ-6 .76 0.65 6.31 7 Gratitude Adjectives Scale – 5 4.28 0.71 .89 1 Satisfaction With Life Scale 1 – 7 5.32 1.11 .84 Empathic Concern 1 – .76 0.53 3.98 5 1 – .74 0.53 3.60 5 Perspective-Taking 1 – 9 6.27 2.29 — Religious interest b General religiousness 8 – 34 24.90 22.18 .88 b 1.59 Intrinsic Religious Orientation 9 – 1 5.88 .89 b Extrinsic Religious Orientation 1 – 9 4.30 0.97 .69 b .83 1 – 9 4.78 1.31 Quest Religious Orientation .90 Self-Transcendence 1 – 4 2.31 0.64 – 1 5 3.84 0.55 .83 Positive Affectivity Negative Affectivity 1 – 5 2.00 0.63 .86 Envy 1 – 5 1.84 1.72 .86 Big Five Openness 3.70 – 1 .76 5 0.57 Conscientiousness – 5 3.63 0.72 .84 1 1 3.60 5 Extraversion 0.76 .85 – Agreeableness – 5 3.91 0.63 .81 1 Neuroticism 1 – 5 2.90 0.81 .85 Note.  Gratitude Questionnaire, six-item form. GQ-6 a Reliability calculated as percentage of true parameter variance per Bryk and Raudenbush (1992). All other b reliabilities estimated with Cronbach ’ s alpha. n  87. All other n s  112. mean ’ measure of gratitude as an affective trait with participants principle (i.e., people who are intrinsically religious) and people levels of gratitude in daily mood remained statistically significant who report high levels of spiritual transcendence experience more .001 (effect size  p 3.53, (103)  .11),  SE .38,  (coefficient t gratitude in their daily moods than do their less religious or r .32). Thus, the association of gratitude as an affective trait with  spiritual counterparts. mean levels of gratitude in daily mood could not be explained Prosocial traits. The Empathic Concern subscale (but not the solely in terms of the Big Five, trait PA, and trait NA. Perspective-Taking subscale) from Davis s Interpersonal Reactiv- ’ ity Index was strongly associated with mean levels of gratitude in people ’ s daily moods. Dispositional envy was not related to mean Predicting the Day-to-Day Fluctuations of Gratitude in levels of gratitude in daily mood. Daily Mood With Gratitude-Relevant Daily Events and The Big Five. Among the Big Five, Agreeableness and Open- Discrete Emotional Reactions to Them ness were positively and significantly correlated with mean levels We proceeded to examine the extent to which the day-to-day s daily moods. These results differ substan- ’ of gratitude in people fluctuations of gratitude in people ’ s daily moods varied as a tially from those of Study 1, in which Extraversion was the major function of scores reflecting day-to-day changes in gratitude fre- ’ Big Five predictor of mean levels of gratitude in people s daily quency, gratitude density, and mean episodic gratitude intensity. moods. We examined these associations by individually adding each of these three time-dependent covariates into the Level 1 model in Associations of Gratitude as an Affective Trait With s daily moods was already decomposed ’ which gratitude in people Gratitude in Daily Mood, Controlling for Higher Order into terms representing initial status, linear change, and residual Personality and Affective Traits variance (as in Equation 4). The resulting associations between day-to-day fluctuations of gratitude in daily mood and the three As in Study 1, we estimated the associations of our two mea- time-dependent covariates are partial associations, with the effects sures of gratitude as an affective trait with mean levels of gratitude for initial status and linear change controlled simultaneously. in daily mood, controlling for their associations with the Big Five As can be seen in Table 5, all three time-dependent covariates and trait PA and NA. When these additional seven variables were were strongly correlated with day-to-day fluctuations in partici- included with the GQ-6, the association of the GQ-6 with mean gratitude mood reports (effect size ’ pants .70, – s ranging from .61 r levels of gratitude in daily mood remained statistically significant  s .001). Thus, people experienced more gratitude in their daily p t  p 2.97,  (103) .003 (effect size .11),  SE .33,  (coefficient moods on days when (a) they experienced many gratitude-eliciting r .27). Similarly, the association of the three-item adjective 

10 CULLOUGH, TSANG, AND EMMONS C M 304 Table 4 Correlations of Personality and Affective Traits With Initial Status Parameters for Gratitude in Daily Mood, Study 2 Effect size Personality measure t r Coefficient SE (110) Affective trait measures of gratitude 5.90*** 0.10 0.57 GQ-6 .49 5.59*** Three-item adjective measure 0.57 0.10 .47 Life satisfaction and affectivity 0.23 Life Satisfaction .30 3.28** 0.07 .30 0.27 0.08 3.29** Happiness .38 Positive Affectivity 0.57 0.13 4.27***   .03  0.34 0.11 Negative Affectivity 0.04 Religious/spiritual variables 0.09 Religious Interest .26 0.03 2.79** 4.09*** .36 0.01 0.05 General Religiousness Intrinsic Religious Orientation 3.25** 0.04 0.13 .30  0.55  0.10 0.06  Extrinsic Religious Orientation .05 .11  0.08 0.07  1.13  Quest Religious Orientation Self-Transcendence 0.49 0.12 3.95*** .35 Prosocial traits Empathic Concern 0.09 .65 8.90*** 0.84 .18 1.95 0.15 0.30 Perspective-Taking Envy 0.07   .07 0.77  0.09 The Big Five .19 2.02* 0.13 0.27 Openness  0.11 0.02  Conscientiousness 0.20 .02  .16 1.74 0.09 0.16 Extraversion 0.13 0.53 Agreeableness .37 4.20*** Neuroticism 0.12 0.09 .01 0.01 GQ-6 Note.  Gratitude Questionnaire, six-item form. .001.  p * .01. ***  .05. ** p  p dict day-to-day fluctuations in participants gratitude mood re- ’ events (high gratitude frequency), (b) many people did things that ports, both gratitude frequency and mean episodic gratitude inten- caused them to feel grateful (high gratitude density), and (c) they r  .39 sity maintained significant unique associations (effect size s experienced a high mean degree of grateful emotion per gratitude- and .65, respectively,  .001), although the number of people p s eliciting event (high mean episodic gratitude intensity). When we r .05,  who caused participants to feel grateful did not (effect size entered all three time-dependent covariates simultaneously to pre- Table 5 Relationship Between Gratitude in Daily Mood and Gratitude Density, Gratitude Frequency, and Mean Episodic Gratitude Intensity, Study 2 Gratitude measure r Coefficient t SE (110) Effect size Individual entry of time-dependent covariates .61 8.17*** 0.01 Gratitude density 0.09 .66 Gratitude frequency 0.20 0.02 9.14*** .70 Mean episodic gratitude intensity 0.33 0.03 10.16*** Simultaneous entry of time-dependent covariates .05 0.52 Gratitude density 0.01 0.01 0.11 0.03 4.49*** .39 Gratitude frequency .65 0.29 0.03 Mean episodic gratitude intensity 9.05*** .001.  p ***

11 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 305 ns ). Thus, the amount of gratitude that people reported in their Additional Construct Validity Data on Measures of daily mood on any given day was uniquely related to the number Gratitude as an Affective Trait of discrete events that caused them to feel grateful that day and Finally, we examined the associations of the GQ-6 and the how grateful they felt, on average, in response to the gratitude- three-item adjective measure of gratitude as an affective trait with eliciting events experienced that day. the daily diary measures of gratitude frequency, gratitude density, and mean episodic gratitude intensity, which we decomposed into initial status and linear change components via HLM 5. As can be Interactions of the Grateful Disposition With Gratitude- seen in Table 6, both measures of gratitude as an affective trait Relevant Daily Events and Discrete Emotional Reactions were correlated with all of these daily measures (initial status to Them: Testing the Conductance and Resistance s r values) in the theoretically expected directions (effect size Hypotheses ranged from .14 to .25). The GQ-6 was significantly correlated typical daily scores on mean episodic gratitude with participants ’ The conductance hypothesis states that for people with a strong intensity and gratitude density (the mean number of people to disposition toward gratitude, the gratitude-relevant events they whom they felt grateful each day). The three-item adjective mea- encounter in daily life and their episodic emotional reactions to sure of gratitude as an affective trait was significantly correlated them are strongly predictive of the amount of gratitude in their only with gratitude density. Neither the GQ-6 nor the three-item daily moods. Conversely, for people with a weak disposition adjective measure of gratitude as an affective trait was correlated toward gratitude, the gratitude-relevant events they encounter in with linear change in these three time-varying measures. daily life and their episodic emotional reactions to them are less strongly predictive of gratitude in daily mood. The resistance Summary hypothesis yields the opposite predictions. We evaluated these hypotheses by examining the interaction Study 2 replicated many findings from Study 1 regarding the s scores on the GQ-6 as a terms resulting from the entry of people ’ personality and affective traits that predict mean levels of gratitude between-subjects predictor of the associations of each of the time- in daily mood. In addition, we found that people high in empathic varying covariates and day-to-day fluctuations in grateful mood. In concern experience more gratitude in their daily moods than do other words, we examined whether the magnitudes of the within- their less empathic counterparts. Unlike Study 1, however, in subjects associations were moderated by gratitude as an affective which Extraversion appeared as the most important Big Five trait. Scores on the GQ-6 were negatively related to the strength of predictor of mean levels of gratitude in daily mood, we found in the associations of gratitude in daily mood with gratitude fre- Study 2 that Agreeableness was the strongest Big Five correlate of 2 .10, quency (coefficient  p 2.86,  (110) t 0.03),  SE  mean levels of gratitude in daily mood.  SE 0.04,  .005, gratitude density (coefficient (110) t  0.01), Study 2 also yielded evidence that gratitude in people s daily ’ moods was strongly related to the number of discrete interpersonal   2.94, .004, and marginally so with mean episodic gratitude p events in their daily lives that elicited gratitude (gratitude fre- 1.76, p   0.08, SE  intensity (coefficient t (110)  .05), quency) as well as the mean degree of gratitude they experienced .08. In other words, gratitude as an affective trait appeared to in response to those discrete events (mean episodic gratitude moderate the associations of gratitude-relevant social events (and intensity). The correlations of these daily measures were strongest s emotional reactions to these events) and the amount of ’ people for people with relatively weak dispositions toward gratitude. gratitude that participants experienced in their daily mood. To interpret these interactions, we compared people who scored half a standard deviation or more above the mean on the GQ-6 1 To ensure that these differences in the correlations for people high (6.63 or higher;  39) with people who scored half a standard n versus low in gratitude as an affective trait were not the artifactual result of n  25). For deviation or more below the mean (5.97 or lower; range restrictions in the relevant measures, we examined their standard people scoring high on the GQ-6, correlations between day-to-day deviations for participants who were half a standard deviation above versus below the mean on the GQ-6. For these two groups of participants, the fluctuations of gratitude in daily mood and gratitude frequency standard deviations for the four variables were as follows: gratitude in daily p r .22,  r .16), as well as gratitude density (  p .21,  ( .17)  mood: 0.74 versus 0.64; gratitude density: 2.66 versus 1.37; gratitude were relatively small and nonsignificant. Only the correlation frequency: 1.53 versus 1.20; mean episodic gratitude intensity: 0.49 versus between day-to-day fluctuations of gratitude in daily mood and 0.50. Because the standard deviations for people scoring high on the GQ-6 p  mean episodic gratitude intensity was significant ( r  .61, were no smaller than they were for people scoring low on the GQ-6, the .001). In contrast, for people scoring low on the GQ-6, day-to-day results could not be explained as an artifact of range restriction. We also fluctuations of gratitude in daily mood were correlated with grat- obtained the same pattern of results when we divided the sample on the basis of scores on the three-item adjective-based measure of gratitude as an  .72, p  itude frequency ( r  r .001), with gratitude density ( affective trait rather than the GQ-6. p  .006), and with mean episodic gratitude intensity ( r .58,  .59, 2 - A significance test for the difference between two independent corre p  .002). In support of the resistance hypothesis, the association lations revealed that the correlation of Extraversion with gratitude in daily of day-to-day fluctuations in the amount of gratitude in people ’ s s  .38 and mood was significantly stronger in Study 1 than in Study 2 ( r mood reports with day-to-day fluctuations in the number of events 96 and 112, respectively; .16, N s  .05). The correlation of Agree- p  for which they felt grateful each day and the number of people to ableness with gratitude in daily mood was marginally significantly stronger whom they felt grateful was relatively weak for people who had in Study 2 than in Study 1 ( r s  .37 and .16, N s  112 and 96, respectively; 1 relatively strong dispositions toward gratitude. p  .06).

12 CULLOUGH, TSANG, AND EMMONS M C 306 emotions (which have a more focused, action-oriented, short-lived Table 6 influence). In the present study, we conceptualized mood in terms Correlations of Affective Trait Measures of Gratitude With s average of two components (a) a component representing people ’ Daily Diary Measures (Initial Status Parameters Only), Study 2 levels of gratitude in their moods over a 2-week or 3-week period Effect s grateful moods around and (b) day-to-day fluctuations of people ’ r size Daily diary measure Coefficient SE t (110) their mean levels. Gratitude Questionnaire, six-item form Affective Traits as Predictors of Mean Levels of Gratitude 0.24 Mean episodic gratitude intensity 0.09 2.71** .25 .21 0.71 Gratitude density 2.24* 0.32 in Daily Mood 0.19 1.53 .14 0.29 Gratitude frequency Adults with neuromuscular disorders and university students who tend to experience high mean levels of gratitude in daily Three-item gratitude adjective scale mood appear to be satisfied with their lives, happy, optimistic, and 1.92 Mean episodic gratitude intensity 0.17 0.09 .18 prone to positive emotionality. Results from both studies were 0.29 0.60 2.09* .20 Gratitude density consistent on this point. In contrast, measures of negative affective 1.77 .17 Gratitude frequency 0.32 0.18 traits (e.g., NA, depression, and envy) were correlated negatively  * .01. .05. **  p p but at rather low magnitudes (e.g., less than  |.25|) with mean r levels of gratitude in participants ’ daily mood reports. These findings corroborate those of McCullough et al. (2002), who found that measures of gratitude as an affective trait were more closely Finally, Study 2 yielded further evidence that measures of grati- linked to positive affective traits than to negative ones. Thus, tude as an affective trait are useful for predicting several dimen- gratitude both as an affective trait and as a mood appears to be sions of gratitude in people ’ s daily interpersonal and emotional characteristic of happy, contented, optimistic people (see also experience. Watkins, 2004). General Discussion Spiritual and Religious Predictors of Mean Levels of The present studies represent the first attempt of which we are Gratitude in Daily Mood aware to examine the relationships among gratitude as a mood, McCullough et al. (2002) reported that measures of gratitude as gratitude as an affective trait, and gratitude as a discrete emotional an affective trait were consistently associated with higher scores response to gratitude-relevant interpersonal events. Specifically, on measures of religious involvement and spirituality. Consistent these results make four major contributions to a scientific under- with those findings, the present results demonstrate that people standing of gratitude. First, they help to identify the personality who score high on traditional measures of religious commitment, and affective traits that are associated with the stable interindi- intrinsic (but not extrinsic or quest) religious motivation, and more vidual differences in the typical degree of gratitude that people nontraditional measures of spirituality experience more gratitude experience in their daily moods. Second, they help to identify key in their typical daily moods than do less religious and spiritual aspects of people ’ s day-to-day social lives (and their emotional people. Again, Studies 1 and 2 were remarkably consistent in this reactions to them) that explain why people experience more grat- regard. As far as we are aware, the present studies yield the first itude in their moods on some days than on others. Third, they evidence that gratitude may be a positive mood characteristic of suggest that gratitude as an affective trait may actually weaken the religiously and spiritually engaged people, just as an absence of s ’ link between gratitude-relevant interpersonal events (and people depressive symptoms is a negative mood characteristic of spiritu- fleeting emotional reactions to those events) and the degree of ally and religiously engaged people (T. B. Smith, McCullough, & gratitude in people ’ s daily moods. Finally, on a methodological Poll, 2003). note, whereas nearly all previous research on experiences of grat- itude has made use of hypothetical gratitude scenarios (e.g., Gra- ham, 1988; Tesser, Gatewood, & Driver, 1968) or retrospective The Big Five as Predictors of Mean Levels of Gratitude accounts of gratitude (Gray et al., 2001), the present set of studies in Daily Mood examined grateful affect as it unfolded in daily life and how it was associated with personality and affective traits as well as events in Studies 1 and 2 provided different pictures of the Big Five ’ people s interpersonal worlds. daily moods. correlates of mean levels of gratitude in participants ’ The present findings are novel in part because they illuminate In Study 1, in which we studied adults with neuromuscular disor- how gratitude operates in people the intermediate — s daily moods ’ ders, mean levels of gratitude in daily mood were related most terrain between affective traits and discrete emotion episodes strongly to Extraversion. In Study 2, in which we studied basically (Rosenberg, 1998) which has been neglected in gratitude re- — healthy university students, mean levels of gratitude in daily mood search to date. Grateful moods are an important level of affect at were related most strongly to Agreeableness. In three previous which to understand gratitude because moods generally have studies, McCullough et al. (2002) found that Agreeableness was greater distributive breadth and pervasiveness in consciousness correlated with gratitude as an affective trait in three samples of than do emotions; as a result, grateful moods probably have basically healthy university students and adults. However, Extra- broader effects on other psychological systems (e.g., cognition, version was also uniquely associated with gratitude as an affective person perception, physiology, coping with stress) than do grateful trait (and Neuroticism negatively so) in several tests. One expla-

13 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 307 high levels of gratitude in daily mood. Although no one knows for nation for these results is that Extraversion and Agreeableness play different roles in facilitating grateful moods depending on whether sure where the boundary between an emotion and a mood really s health status limits exposure to potential benefactors. one ’ lies, it seems plausible from these data that discrete episodes of Basically healthy university students are likely, on average, to grateful emotions diffuse into daily mood, thereby casting a grate- have relatively high access to socially supportive relationships ’ ful affective tone over people s daily mood experiences. (friends, roommates, classmates, other members of organizations to which they belong). Moreover, in their in loco parentis capacity, How Gratitude as an Affective Trait Interacts With Daily universities encourage students to avail themselves of a wide Experience variety of academic, social, and health services (which are, among other things, structured forms of social support) to address nearly In her theoretical analysis of the relationships between affective every possible problem in living. Thus, for university students, the traits, moods, and emotions, Rosenberg (1998) focused primarily key to experiencing grateful moods in daily life may not be simply on the bivariate causal associations among these three levels of exposing oneself to benevolent people. Rather, grateful moods affect. In the present article we have explored how two of these may be a matter of appreciating the wide range of beneficial levels of affect, namely, affective traits and emotions, might op- relationship partners with whom one might interact in a given erate interactively to influence gratitude in daily mood. We found week. Agreeableness is the Big Five dimension most closely that the disposition toward gratitude can be conceptualized as a linked with perceiving others in a charitable and benevolent fash- trait that establishes consistently high amounts of gratitude in ion and may therefore be the trait that determines the degree of ’ s daily moods irrespective of day-to-day fluctuations in people gratitude experienced by people who live in social worlds that gratitude-relevant interpersonal events and people ’ s emotional re- teem with potential benefactors. To be in a grateful mood, perhaps actions to them. Because dispositionally grateful people ’ s prone- university students merely need to recognize the benevolent nature ness to grateful moods is driven so strongly by personality, their of the social worlds they typically inhabit. grateful moods may be less dependent on the ebb and flow of In contrast, for people with chronic illness, whose mobility and gratitude-relevant life events (the number of events for which they social interaction may be limited to some extent by physical are grateful each day), the number of people to whom they feel symptoms, experiencing grateful moods may be more contingent grateful each day, and (marginally) the amount of gratitude they on actually interacting with an adequate number of supportive experience in the typical gratitude-eliciting situation. These find- relationship partners. Extraversion is the Big Five dimension with ings are supportive of the resistance hypothesis of the grateful the most relevance for seeking out and enjoying social interaction disposition. (Ashton, Lee, & Paunonen, 2002), so perhaps the more extroverted This finding seems counter to much existing work. For example, among adults with chronic physical illnesses are more likely to some work (e.g., Gross et al., 1998; Gunthert, Cohen, & Armeli, encounter benevolent interaction partners. For people with chronic 1999; Larsen & Ketelaar, 1991) has suggested that affective traits illnesses that can limit functional abilities, feeling grateful in daily PA moderate the link – NA and Extraversion – such as Neuroticism life may be largely a matter of ensuring that one interacts with between emotionally valenced life events and emotional reactions, other people who can provide one with such benefits. that is, that people with a personality-based proneness to positive In any case, the contribution of the Big Five to predicting mean emotions (e.g., people high in Extraversion) experience more levels of gratitude in daily mood was modest (i.e., they accounted positive emotional reactions to pleasant events, and people with a for approximately 18% – 26% of the variance in the amount of personality-based proneness to negative emotions (e.g., people s daily moods). In this respect, the present gratitude in people ’ high in Neuroticism) experience more negative emotional reac- results mirror the relationship between gratitude as an affective tions to unpleasant events. trait and the Big Five found by McCullough et al. (2002), who However, not all of the existing research supports the conduc- reported that the Big Five accounted for 21% – 33% of the variance tance hypothesis. David, Green, Martin, and Suls (1997) found no in self-ratings of the disposition toward gratitude. Therefore, we evidence that Neuroticism and Extraversion moderated the links can conclude confidently that the Big Five (in some still unclear between desirable or undesirable life events and daily mood. Also, combination that might vary depending on one ’ s life situation) consistent with the resistance hypothesis and the present findings, account for some of the variance in mean levels of gratitude in Affleck et al. (1992) reported that among patients with rheumatoid daily mood, though only a fraction of it. arthritis, Neuroticism appeared to weaken the association between daily pain ratings and negative affect in daily mood. One feature Daily Gratitude-Relevant Events as Predictors of Day-to- that seems to differentiate the present study from many of the Day Fluctuations in Grateful Mood others in this area is that rather than examining the conductance and resistance hypotheses vis-a `-vis global affective traits (e.g., Day-to-day fluctuations in grateful mood are strongly related to PA and Neuroticism NA), we looked at a discrete – – Extraversion the day-to-day fluctuations in gratitude-relevant aspects of peo- affective trait (i.e., the grateful disposition) that predisposes people ’ ple s daily lives. Participants reported the most gratitude in their to experience a particular discrete affect in daily mood (i.e., daily moods on days (a) when they reported many situations that gratitude). Perhaps other investigators using similar methods will caused them to feel grateful, (b) when they listed many people to — for example, hostility — find that discrete affective traits predis- whom they felt grateful, and (c) when they felt high levels of pose people to experience (a) higher levels of discrete affects (e.g., grateful emotion per daily episode. In other words, discrete expe- anger) in daily mood and (b) emotion-inducing (i.e., anger- riences of grateful emotions — along with the interpersonal events appeared to foster — and appraisals that ostensibly occasion them inducing) events more frequently but not a stronger link between

14 CULLOUGH, TSANG, AND EMMONS C M 308 quasi-signal detection analysis of daily interactions between close rela- the number of anger-inducing events experienced and the amount Psychological Science, 14, tionship partners. – 105. 100 of anger experienced in daily mood. Graham, S. (1988). Children ’ s developing understanding of the motiva- Cognitive Development, 3, tional role of affect: An attributional analysis. Summary 71 88. – Gray, S. A., Emmons, R. A., & Morrison, A. (2001, August). Distinguish- The results presented herein stand alongside recent work that ing gratitude from indebtedness in affect and action tendencies. Poster has helped to clarify the nature of gratitude as a discrete emotional presented at the 109th Annual Convention of the American Psycholog- experience (McCullough et al., 2001) and an affective trait (Mc- ical Association, San Francisco, CA. Cullough et al., 2002). On the basis of these findings, we conclude Greenberg, M. S. (1980). A theory of indebtedness. In K. J. Gergen, M. S. that grateful moods are created both through top-down effects (i.e., Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory the effects of personality and affective traits), bottom-up effects – (pp. 3 and research 26). New York: Plenum Press. Gross, J. J., Sutton, S. K., & Ketelaar, T. (1998). Relations between affect (i.e., the effects of discrete interpersonal and emotional episodes), and personality: Support for the affect-level and affective reactivity and the interaction of these effects. More broadly, we have at- views. 288. – 279 Personality and Social Psychology Bulletin, 24, tempted to articulate a view of how the three levels of affect — Gunthert, K. C., Cohen, L. H., & Armeli, S. (1999). The role of Neuroti- relate to one another using — affective traits, moods, and emotions cism in daily stress and coping. Journal of Personality and Social the affect of gratitude as a case in point. As such, the present 1087 – Psychology, 77, 1100. results may be useful not only for better understanding gratitude Methods of meta-analysis: Cor- Hunter, J. E., & Schmidt, F. L. (1990). specifically but as a useful model for articulating how these three Newbury Park, CA: Sage. recting error and bias in research findings. levels of affect relate to one another for other discrete affects as John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five well. Berkeley: University of California, — Versions 4a and 54. Inventory Berkeley, Institute of Personality and Social Research. John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, References measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. Affleck, G., Tennen, H., Urrows, S., & Higgins, P. (1992). Neuroticism – 138). New York: Guilford Press. 102 and the pain – mood relation in rheumatoid arthritis: Insights from a Kirk, K. M., Eaves, L. J., & Martin, N. G. (1999). Self-transcendence as a prospective daily study. Journal of Personality and Social Psychology, measure of spirituality in a sample of older Australian twins. Twin 126. 60, 119 – Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and – Research, 2, 61 87. Journal of Personality and Social Psychology, 5, 432 443. – prejudice. Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to Ashton, M. C., Lee, K., & Paunonen, S. V. (2002). What is the central Journal of Personality and Social positive and negative emotional states. Jour- feature of Extraversion? Social attention versus reward sensitivity. 132 – 140. Psychology, 61, nal of Personality and Social Psychology, 83, 245 – 251. Passion and reason: Making sense Lazarus, R. S., & Lazarus, B. N. (1994). Batson, C. D., & Schoenrade, P. A. (1991). Measuring religion as quest: 1. New York: Oxford University Press. of our emotions. Journal for the Scientific Study of Religion, 30, Validity concerns. McCullough, M. E., Emmons, R. A., Kilpatrick, S. D., & Mooney, C. N. 429. 416 – ” (2003). Narcissists as : The role of narcissism in the perception “ victims Social Beckie, T. M., & Hayduk, L. A. (1997). Measuring quality of life. – of transgressions. 885 Personality and Social Psychology Bulletin, 29, 21 Indicators Research, 42, – 39. 893. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models. McCullough, M. E., Emmons, R. A., & Tsang, J. (2002). The grateful Newbury Park, CA: Sage. disposition: A conceptual and empirical topography. Journal of Person- Campbell, A., Converse, P. E., & Rodgers, W. L. (1976). The quality of 127. – 112 ality and Social Psychology, 82, New York: Russell Sage Foundation. American life. McCullough, M. E., Kilpatrick, S. D., Emmons, R. A., & Larson, D. B. Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobi- (2001). Is gratitude a moral affect? Psychological Bulletin, 127, 249 – Archives of General ological model of temperament and character. 266. – Psychiatry, 50, 975 990. Nezlek, J. B. (2001). Multilevel random coefficient analyses of event- and David, J. P., Green, P. J., Martin, R., & Suls, J. (1997). Differential roles interval-contingent data in social and personality psychology research. of Neuroticism, Extraversion, and event desirability for mood in daily 771 Personality and Social Psychology Bulletin, 27, 785. – life: An integrative model of top-down and bottom-up influences. Jour- Piedmont, R. L. (1999). Does spirituality represent the sixth factor of 159. 149 nal of Personality and Social Psychology, 73, – Journal personality? Spiritual transcendence and the five-factor model. Davis, M. H., & Oathout, H. A. (1987). Maintenance of satisfaction in 1013. – 985 of Personality, 67, romantic relationships: Empathy and relational competence. Journal of Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for – Personality and Social Psychology, 53, 397 410. research in the general population. Applied Psychological Measurement, Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The 385 1, – 401. Journal of Personality Assessment, 49, Satisfaction With Life Scale. Raudenbush, S., Bryk, A., Cheong, Y. F., & Congdon, R. (2000). HLM5: – 71 75. Hierarchical linear and nonlinear modeling [Computer software]. Lin- Emmons, R. A., & Crumpler, C. A. (2000). Gratitude as human strength: colnwood, IL: Scientific Software International. Appraising the evidence. Journal of Social and Clinical Psychology, 19, Rosenberg, E. L. (1998). Levels of analysis and the organization of affect. 69. – 56 270. Review of General Psychology, 2, 247 – Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing burdens: An experimental investigation of gratitude and subjective well- optimism from neuroticism (and trait anxiety, self-mastery, and self- being in daily life. Journal of Personality and Social Psychology, 84, esteem): A reevaluation of the Life Orientation Test. Journal of Person- 389. 377 – 1078. – ality and Social Psychology, 67, 1063 Gable, S. L., Reis, H. T., & Downey, G. (2003). He said, she said: A

15 GRATITUDE IN INTERMEDIATE AFFECTIVE TERRAIN 309 (6th ed.). Oxford, Smith, A. (1976). (pp. The psychology of gratitude Emmons & M. E. McCullough (Eds.), The theory of moral sentiments England: Clarendon Press. (Original work published 1790) 192). New York: Oxford University Press. – 167 Smith, R. H., Parrott, W. G., Diener, E., Hoyle, R. H., & Kim, S. H. (1999). Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali- Personality and Social Psychology Bulletin, 25, Dispositional envy. dation of brief measures of positive and negative affect: The PANAS – 1020. 1007 – Journal of Personality and Social Psychology, 54, scales. 1063 1070. Smith, T. B., McCullough, M. E., & Poll, J. (2003). Religiousness and Weiner, B., & Graham, S. (1989). Understanding the motivational role of depression: Evidence for a main effect and the moderating influence of affect: Life-span research from an attributional perspective. Cognition stressful life events. 636. – 614 Psychological Bulletin, 129, and Emotion, 3, 401 419. – – environment fit and its Suls, J., Martin, R., & David, J. P. (1998). Person limits: Agreeableness, neuroticism, and emotional reactivity to personal – conflict. Personality and Social Psychology Bulletin, 24, 88 98. Received December 3, 2001 Tesser, A., Gatewood, R., & Driver, M. (1968). Some determinants of Revision received May 13, 2003 236. – gratitude. Journal of Personality and Social Psychology, 9, 233 Watkins, P. C. (2004). Gratitude and subjective well-being. In R. A. Accepted May 13, 2003 

Related documents

What is an Honors Student?

What is an Honors Student?

Univ ersit ebraska - L in coln y of N coln gitalC niv ersit y of N ebraska - L in [email protected] omm Di ational Collegiate Honor s Counc il - Journal of the N il s Counc National Collegiate Honor rchive -Onl i...

More info »
Title IX at 45 Advancing Opportunity through Equity in Education

Title IX at 45 Advancing Opportunity through Equity in Education


More info »
Microsoft Word   20190227 Chinas Impact on the US Education System

Microsoft Word 20190227 Chinas Impact on the US Education System

United States Senate PERMANENT SUBCOMMITTEE ON INVESTIGATIONS Committee on Homeland Security and Governmental Affairs Rob Portman, Chairman Tom Carper, Ranking Member CHINA’S IMPACT ON THE U.S. EDUCAT...

More info »
What Matters to Student Success:  A Review of the Literature

What Matters to Student Success: A Review of the Literature

What Matters to Student Success: A Review of the Literature Commissioned Report for the National Symposium on Postsecondary Student Success: Spearheading a Dialog on Student Success George D. Kuh Jill...

More info »
RE Law

RE Law

Real Estate License Law 6 ) ( 201 June New York DEPARTMENT OF STATE Division of Licensing Services Andrew M. Cuomo Rossana Rosado Gove rnor Secr e tary of State

More info »