C
Psychology in the Schools, Vol. 53(3), 2016 View this article online at wileyonlinelibrary.com/journal/pits
2016 Wiley Periodicals, Inc. DOI: 10.1002/pits.21903
WHAT GOOD IS GRATITUDE IN YOUTH AND SCHOOLS? A SYSTEMATIC REVIEW AND META-ANALYSIS OF CORRELATES AND INTERVENTION OUTCOMES TYLER L. RENSHAW AND RACHEL M. OLINGER STEEVES
Louisiana State University
The development of gratitude in youth has received increasing attention during the past several years, and gratitude-based interventions have often been recommended for use in schools. Yet, the empirical status of the correlates of gratitude and the effects of gratitude-based interventions on youths’ outcomes remains unclear. The present study addressed this ambiguity by systematically reviewing and meta-analyzing original empirical journal articles investigating gratitude in youth through September 2014 (N = 20). Findings from the meta-analysis of correlates indicate small-to-moderate convergent and discriminant evidence for gratitude as a subjective well-being indicator in youth, yet other results indicate that gratitude measures have relatively poor test–retest reliability and/or predictive validity and that they have questionable concurrent validity with other gratitude measures. Moreover, findings from the meta-analysis of intervention outcomes indicate that gratitude-based interventions are, as a whole, generally ineffective and that much more intervention research is warranted. Implications of these findings for theory, future research, and the C 2016 Wiley Periodicals, Inc. practice of school psychology are discussed.
A positive youth psychology has been developing under various names, including resilience studies (e.g., Masten, 2001), strength-based approaches (e.g., Jimerson, Sharkey, Nyborg, & Furlong, 2004), and positive youth development (e.g., Larson, 2002). Scholars and practitioners contributing to this movement hail from a variety of disciplines and training backgrounds (i.e., developmental psychology, social psychology, school psychology, child-clinical psychology, social work, and education), yet they share a common interest in cultivating a subfield of applied scientific inquiry wherein youth’s strengths, assets, and well-being are understood and promoted via methodologically rigorous and socially valid means. A broad look at the literature emerging from the subfield of positive youth psychology suggests that scientific headway is indeed being made, as recent books (e.g., Furlong, Gilman, & Huebner, 2014), book chapters (e.g., Kirschman, Johnson, Bender, & Roberts, 2009), intervention studies (e.g., Froh, Sefick, & Emmons, 2008), measurement studies (e.g., Froh, Fan, et al., 2011), other construct-validation studies (e.g., Froh, Bono, & Evans, 2010), and review articles (e.g., Froh, Miller, & Snyder, 2007) indicate a burgeoning base of scholarship seeking to describe, predict, and influence the positive functioning of youth across developmental periods (i.e., early childhood, adolescence, and young adulthood) and life contexts (i.e., home, school, and community settings). Commenting on the current state-of-the-art of a positive psychology for youth, Huebner and Hills (2011) concluded that although the available empirical research “provides a reasonable scientific foundation for enhanced professional knowledge and practice with children and youth” (p. 89), such research has, to date, made little headway into applied psychological practice and, therefore, has likely had little practical influence on the well-being of youth in schools. Although there may be myriad reasons for this research-to-practice gap—such as lacking pre-service and in-service training opportunities in positive-psychological assessment and intervention strategies—this disconnect might also be fueled by one major-yet-fixable problem within the contemporary scholarship itself. Specifically, we suggest that the lack of systematic reviews and meta-analyses of the empirical status of correlates and intervention outcomes for particular positive-psychological constructs in Correspondence to: Tyler Renshaw, Department of Psychology, Louisiana State University, 236 Audubon Hall, Baton Rouge, LA 70803. E-mail:
[email protected]
286
Gratitude Meta-Analysis
287
youth is likely to facilitate, albeit unintentionally, ambiguity and confusion among researchers and practitioners, and that this ambiguity, in turn, impedes the progress of evidence-based practice. At worst, such lack of systematic and meta-analytic research may result in scholars and trainers misunderstanding or miscommunicating the research regarding positive youth psychology to practitioners and, as a result, practitioners may misunderstand and misapply relevant scholarly work in the schools. Moreover, in the absence of systematic and meta-analytic research, scholars and practitioners are likely to selectively attend to and privilege positive research findings over discomforting results, as well as to overemphasize the statistical significance of individual studies at the expense of evaluating overarching trends and cumulative effect sizes within the extant research base (Schmidt, 1992). To illustrate the need for systematic and meta-analytic research within the subfield of positive youth psychology in particular, we turn next to a discussion of one of the most well-known positive psychological constructs, which has recently found its way into schools: gratitude. G RATITUDE
IN
YOUTH
AND
S CHOOLS
As one of the 24 original “character strengths” outlined in Peterson and Seligman’s (2004, p. 3) so-called “Manual of Sanities,” gratitude is arguably one of the most popular of positive psychology’s constructs. Although work in this area originated with adults, within the past 5 years, gratitude in youth has become the subject of much scholarly press (e.g., Bono, Froh, & Forrett, 2014), semi-scholarly press (e.g., Froh & Bono, 2012), and even popular press (e.g., Williams, 2012). Yet, despite receiving so much attention, the construct of gratitude in youth has been inconsistently defined. Emmons and McCullough (2003) noted that gratitude “defies easy classification” and can been conceptualized variously as “an emotion, an attitude, a moral virtue, a habit, a personality trait, or a coping response” (p. 377). This loose operationalization is apparent in the empirical work conducted with youth as well. For example, gratitude in youth has been defined cognitively as “an estimate of gain coupled with the judgment that someone else is responsible for that gain” (Chen & Kee, 2008, p. 319), emotionally as “a sense of thankfulness and joy in response to receiving a gift, whether the gift be a tangible benefit from a specific other or a moment of peaceful bliss evoked by natural beauty” (Froh, Yurkewicz, & Kashdan, 2009, p. 554), and behaviorally as “a positive response to receiving a benefit” (Froh, Emmons, Card, Bono, & Wilson, 2011, p. 291). For the purposes of this work, which aims to meta-analyze empirical findings across studies investigating gratitude in youth using varying operationalizations, we employed the latter and broader behavioral definition, positing that “positive responses to receiving a benefit” adequately accounts for the cognitive and emotional facets of gratitude referenced in other common definitions. Despite its operational ambiguity, the gratitude construct has received significant professional attention within the field of school psychology in the last decade, especially from the National Association of School Psychologists (NASP). Beginning in 2007, NASP published a review article by Froh et al. (2007) in the School Psychology Forum, which recommended gratitude-based interventions for practitioners to use with students (as well as with themselves) and concluded that gratitude “has clear and important implications for school psychology” (p. 10). Next, in 2009a, NASP teamed with the Fishful Thinking initiative and released a resource called “Fostering an Attitude of Gratitude: Tips for Parents,” which offered simple strategies for caregivers to use for the purposes of cultivating gratitude in youth. Shortly thereafter, NASP (2009b) developed and disseminated a gratitude-based intervention program, called “Gratitude Works,” which was promoted as the keystone initiative of School Psychology Awareness Week in both 2009 and 2013.
Psychology in the Schools
DOI: 10.1002/pits
288
Renshaw and Olinger
In the interim, NASP also published two narrative reviews of gratitude-based interventions in the “Research-Based Practice” section of the Communiqu´e (Froh & Bono, 2011; Reivich, 2009), the former of which concluded that “gratitude can lead to many positive outcomes of central importance for children and adolescents” (Froh & Bono, 2011, p. 28). In addition, a third NASP article was published in the Communiqu´e, reporting on practitioners’ experiences implementing gratitude-based interventions in schools (Molony & Henwood, 2010). That said, the visibility of and interest in gratitude within the profession of school psychology, and by NASP in particular, does not necessarily imply that its correlates are well understood or that the effects of gratitude-based interventions for children and adolescents are well established. Indeed, a cursory review of the scholarly literature suggests otherwise. A MBIGUITIES
IN
G RATITUDE R ESEARCH
WITH
YOUTH
According to Wood (2014), who is one of the pioneers of empirical research regarding gratitude in adults, the scientific and semi-scientific literature on gratitude is often plagued by too much enthusiasm and too little skepticism, resulting in overly optimistic interpretations of research findings that fail to highlight empirical ambiguities. For example, despite being lauded and advocated in popular-press books as a well-established, science-based practice (e.g., Froh & Bono, 2014), the efficacy of gratitude-based interventions for youth seems unclear in the peer-reviewed literature. Although there is some evidence indicating that gratitude-based interventions have helped to enhance at least one aspect of youths’ subjective well-being (i.e., school satisfaction; Froh et al., 2008), there is other evidence indicating that gratitude-based interventions have had null effects on several other aspects of subjective well-being (i.e., positive affect, negative affect, and general life satisfaction; Froh, Kashdan et al., 2009). Furthermore, given that narrative reviews of gratitudebased interventions tend to cite only one or two popular intervention studies (i.e., Froh et al., 2008; Froh, Kashdan et al., 2009), it is unknown just how many other protocols for gratitude-based intervention have been developed, tested, and reported, as well as which specific outcomes have been targeted via such interventions, and, most importantly, what effect sizes result from such interventions. Studies investigating gratitude-based interventions have also used a variety of comparison conditions to gauge the efficacy of the intervention, ranging from passive controls (i.e., “educationas-usual”) to active-positive controls (i.e., intentionally engaging in another constructive behavior) to active-negative controls (i.e., intentionally engaging in an aversive behavior) to alternative interventions (i.e., previously validated treatments). Considering that a thoroughgoing understanding of intervention effects requires consideration of the nature of comparison groups, meta-analysis is also warranted to see whether the effects of gratitude-based interventions vary as a function of control conditions. Although scholars promoting the use of gratitude-based interventions have admitted in some instances that the relevant research “is still in its early stages” and that “the implementation of school-based programs to enhance gratitude is limited in scope” (Bono et al., 2014, p. 68), they have also, in other instances, promoted the use of gratitude-based interventions in a more sweeping manner for improving youths’ well-being (Bono et al., 2014; Froh & Bono, 2011), which seems to ignore the contextual conditions in which these interventions were tested and, in turn, the reasonable interpretations and generalizations that can be made from available data. Furthermore, although there is certainly evidence to suggest that youths’ gratitude is a bona fide psychological construct that is measurable via psychometrically sound assessment instruments (e.g., Froh, Fan et al., 2011), and several high-profile public presentations have emphasized the strong predictive validity of gratitude in youth (e.g., Bono, 2014, September 29), the scope, consistency, and effect sizes of the correlates of these measures—as well as the strength of concurrent associations among different measures of youths’ gratitude itself—have yet to be systematically investigated. Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
289
Thus, although it has been demonstrated that gratitude has at least small-to-moderate, positive associations with several subjective well-being outcomes and similarly negative associations with subjective distress outcomes (e.g., van Eeden et al., 2008), it is unclear how these correlations vary in magnitude across assessment instruments (e.g., the Gratitude Questionnaire–6 [GQ–6] vs. the Gratitude Adjective Checklist [GAC]), as well as whether different measures of the construct have practically meaningful associations with other types of measures, such as informant-rated or performance-based assessments of student well-being that are highly valued within school contexts (e.g., teachers’ perceptions of student behavior or direct measures of academic skill proficiency). Considering the importance of such well-being measures within school-based service delivery frameworks that are guided by the problem-solving model and multitiered systems of student support (Ervin, Gimpel Peacock, & Merrell, 2010; Hawkins, Barnett, Morrison, & Musti-Rao, 2010), a systematic understanding of the correlates of gratitude in children and adolescents could help inform best practices relative to promoting positive youth psychology in schools. P URPOSE
OF THE
P RESENT S TUDY
Given the previously discussed context, the purpose of the present study was to investigate the scientific goodness—or, more technically, the empirical utility—of gratitude in youth and schools by systematically reviewing and, where appropriate, meta-analyzing its correlates and intervention outcomes for children and adolescents. Specifically, given school psychology and NASP’s apparent interest in the gratitude construct, as well as some scholars’ recommendations for gratitude-based interventions to “be encouraged as early in life as possible” due to their supposed multifaceted benefits for students, teachers, and school climates (Froh, Bono, Fan et al., 2014, p. 148), we aimed to systematically examine and interpret the extant research to determine the empirical grounding underlying the present enthusiasm for gratitude. M ETHOD Study Search and Identification Following the recommendations of Beretvas (2010) and Field (2009), we conducted the systematic review and meta-analysis using a multi-stage protocol that comprised three phases: study search and identification, study coding, and data analyses. In the first phase, we developed inclusion criteria for relevant studies, deciding that, given our aims, an identified study must be (a) an original empirical investigation that (b) explicitly targeted the correlates or intervention outcomes of gratitude in youth and (c) was published or scheduled to be published (i.e., advance online publication) in a peer-reviewed academic journal (for the purposes of ensuring quality control) that (d) was reported in English (as the researchers were only fluent in this language). Following, we put this inclusion criteria to use in identifying potentially relevant studies for meta-analysis by conducting a computerized search of abstracts in the PsycINFO, ERIC, MEDLINE, Anthropology Plus, SocINDEX, EconLit, and Psychology and Behavioral Sciences Collection databases using the topical keyword “gratitude,” paired with iterations of the population keywords “children,” “adolescents,” “youth,” and “students,” while limiting results to peer-reviewed articles published in academic journals. Findings from this initial multi-database search, which spanned all publication years up to the date of the present search, September 2014, yielded 44 unique study abstracts that appeared to meet inclusion criteria and thus warranted further review. In the third step of the sampling procedure, both authors independently re-reviewed the complete abstracts of each of the 44 studies and screened the “Method” section of studies with ambiguous abstracts to systematically identify all articles meeting the inclusion criteria. Findings from both authors’ independent reviews yielded 100% agreement regarding 17 original empirical investigations Psychology in the Schools
DOI: 10.1002/pits
290
Renshaw and Olinger
that met all inclusion criteria. Of the remaining studies that were not identified for inclusion in the systematic review and meta-analysis, three were qualitative empirical investigations with youth (i.e., Baumgarten-Tramer, 1938; Freitas, Pieta, & Tudge, 2011; Gordon, Musher-Eizenman, Holub, & Dalrymple, 2004); five were quantitative investigations that included measures of youths’ gratitude as indicators of higher-order latent variables but did not examine direct correlations at the lowerorder level of the gratitude construct itself (i.e., Furlong, You, Renshaw, O’Malley, & Rebelez, 2013; Furlong, You, Renshaw, Smith, & O’Malley, 2014; Gillham et al., 2011; Weber, Ruch, Littman-Ovadia, Lavy, & Gai, 2013; You et al., 2013); four were empirical studies conducted with young-adult college students; three were only published in a language other than English (i.e., Chinese or Japanese); and 12 were nonempirical articles of some sort, such as conceptual papers and narrative reviews. After identifying the articles that met all inclusion criteria, both authors conducted an archival review of the “References” section within each of the identified articles, searching for studies that might meet inclusion criteria yet did not turn up in the computerized database searches. Results of this archival review of references yielded 100% agreement regarding three additional studies that met all inclusion criteria (as well as three articles that met most inclusion criteria but were published in Portuguese), resulting in a grand total of 20 identified empirical studies that were appropriate for coding. Study Coding As the initial step in the study coding phase, the 20 identified studies were sorted according to the type of data they reported, resulting in 14 studies reporting only correlates of gratitude measures in youth, five studies reporting only intervention outcomes, and one study reporting both correlates and intervention outcomes. Next, each of the 15 studies reporting on correlates of gratitude in youth was coded by the first author using a basic coding scheme that accounted for (a) general study characteristics (i.e., authors, year of publication, and journal), (b) sample characteristics (e.g., age/grade-level, gender composition, number of sample participants), (c) measurement characteristics (e.g., name of gratitude measures and type of constructs represented by concurrent and/or predictive validity measures), and (d) Pearson r effect size coefficients, which indicated the magnitude of the association between each gratitude measure and the concurrent and/or predictive validity measures reported in each study. Each of the six studies reporting intervention outcomes was coded by the first author using a similar basic coding scheme, which accounted for (a) general study characteristics, (b) sample characteristics, and (c) measurement characteristics, while including additional codes accounting for (d) research design (e.g., types of conditions and timing of data collection), (e) intervention characteristics (e.g., name and description of the intervention protocol, intervention dosage, and treatment fidelity), and (f) descriptive outcome data (e.g., group sizes, means, and standard deviations) that could be used to calculate Hedges’ g effect size estimates, which would indicate the magnitude of the standardized mean differences among gratitude-based intervention and comparison conditions reported in each study. Data Analysis Given that only 20 studies were identified for inclusion in the systematic review and metaanalysis, we determined that a minimalistic or “bare bones” approach to data analysis would be employed for both correlates and intervention effects, calculating only the overall weighted mean effect size for each correlate or outcome of interest, followed by the calculation of summary effect sizes for specific subgroups of outcomes (i.e., positive vs. negative correlates, post-test vs. followup effects, and intervention comparison conditions). Thus, no potential moderating variables (e.g., Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
291
geographic location, type of gratitude measure, or treatment fidelity) were selected for further investigation and no artifact correction procedures (e.g., for measure reliability or dichotomization of variables) were deemed appropriate or necessary for our purposes (see Beretvas, 2010). Furthermore, given the nature of psychological research and considering the variance in participant age and geographic location of the samples represented in the identified studies, as well as the small number of total studies (N = 20) eligible for the meta-analysis, we conceptualized our analyses from a random-effects perspective, which assumes that the identified studies represented populations with naturally varying average effect sizes across the constructs of interest and helps enhance statistical power (see Field, 2009; Valentine, Pigott, & Rothstein, 2010). After deciding on the analytic approach, the first author entered the relevant coded data (i.e., study name, sample and subsample group sizes, measurement characteristics, and effect size information) into two separate electronic files, one for correlates and the other for intervention effects, within the Comprehensive Meta-Analysis, Version 2 software (CMA; Borenstein, Hedges, Higgins, & Rothstein, 2006–2014). Information representing a total of 89 correlate effect sizes were entered into the first CMA file and information representing a total of 93 intervention outcome effect sizes (48 post-test comparisons and 45 follow-up comparisons) were entered in the second file. The CMA program was used to meta-analyze the weighted mean effect sizes, using the “random effects” and “cumulative analysis” functions for all coded correlates (resulting in an overall Pearson r), as well as summary effect sizes for total positive and negative correlates. Using similar procedures, all coded intervention effects were then systematically reviewed and meta-analyzed (resulting in an overall Hedges’ g) and summary effect sizes were obtained for total intervention effects, as well as post-test, follow-up, passive control, positive-active control, negative-active control, and alternative intervention comparison conditions. Additionally, for all correlates and intervention effects derived from at least two samples, Cochran’s Q and Higgins and Thompson’s I2 were calculated to determine the relative heterogeneity of the collective individual effect sizes represented by each weighted mean effect size. For all correlates and intervention effects derived from at least three samples, Orwin’s fail-safe N, which is a more robust version of the classic fail-safe N, was calculated to estimate potential publication, or “file-drawer,” bias. R ESULTS Systematic Review of General Study Characteristics Table 1 presents a systematic review of the descriptive characteristics of the 20 original empirical studies included in the meta-analysis. Descriptive information indicated that the first study investigating the correlates of gratitude in youth was published in 2006 (Park & Peterson), whereas the first evaluation of a gratitude-based intervention with youth was published in 2008 (Froh et al., 2008). Although no gratitude studies were apparently published in 2007, at least two empirical studies were published per year from 2008 to 2013, with four published to date (September) in 2014. The majority of studies were conducted with youth in adolescence (e.g., Weber & Ruch, 2012) or middle-to-late childhood (e.g., Hoy, Suldo, & Mendez, 2013), whereas only one study targeted youth in early childhood (Owens & Patterson, 2013). To date, 11 empirical studies targeting gratitude in youth were conducted in the United States, five were carried out in Asia, three in Europe, and one in Africa. The largest investigation (N = 1,691) was a survey study of youths’ gratitude that was embedded within a multidimensional measure of character strengths (van Eeden, Wissing, Dreyer, Park, & Peterson, 2008), and the smallest study (N = 69) was a gratitude-based intervention with elementary-age children (Owens & Patterson, 2013). For the correlation studies, as well as the one intervention study that used a gratitude measure (Froh, Kashdan, Ozimkowski, & Miller, 2009), the GQ–6 (McCullough, Emmons & Tsang, 2002) was the Psychology in the Schools
DOI: 10.1002/pits
Psychology in the Schools
DOI: 10.1002/pits
2006 2008 2008 2008 2009 2009 2010 2010 2011 2011 2012 2012 2013 2013 2014 2014 2014 2014 2014 2014
Park & Peterson Chen & Kee Froh et al. van Eeden et al. Froh, Kashdan et al. Froh, Yurkewicz et al. Akhtar & Boniwell Froh et al. Froh, Emmons et al. Froh, Fan et al. Li, Zhang, Li, Li, & Ye Weber & Ruch Hoy et al. Owens & Patterson Chen & Chang Datu Ruch, Weber, Park, & Peterson Tian et al. Froh et al. Froh et al.
Correlates Correlates Outcomes Correlates Outcomes, Correlates Correlates Outcomes Correlates Correlates Correlates Correlates Correlates Correlates Outcomes Correlates Correlates Correlates Correlates Outcomes Outcomes
Data 1 1 1 1 1 1 1 1 1 6 1 2 2 1 1 1 1 2 1 1
k 250 169 221 1,691 89 154 20 700 1,035 1,405 1,252 174 148 62 293 210 247 760 122 98
N — (10–17) 16.43 (15–18) 12.17 (n/a) — (13–17) 12.74 (8–19) 12.14 (11–13) 17.50 (14–20) 11.74 (10–14) 15.67 (14–19) 14.47 (10–19) 15.00 (12–19) 16.45 (13–19) 10.22 (9–11) 7.35 (5–11) 16.14 (n/a) 17.76 (15–21) 11.77 (10–14) 11.07 (8–14) 9.03 (8–10) 9.50 (8–11)
M Age (Range) United States Taiwan United States Africa United States United States England United States United States United States China Switzerland United States United States Taiwan Philippines Switzerland China United States United States
Location VIA–Y GQ–6 — VIA–Y GAC GAC — GAC GQ–6, GAC, GRAT–S GQ–6, GAC, GRAT–S GQ–6 VIA–Y GQ–6 — GQ–6 GQ–6 VIA-Y GQ–6 GAC GAC
Gratitude Measure
— — Counting Blessings — Letter Writing — Positive Psychology Curriculum — — — — — — Picture Drawing — — — — Benefit Appraisal Curriculum Benefit Appraisal Curriculum
Gratitude Intervention
Note. k = number of independent samples per study; GQ–6 = Gratitude Questionnaire–6; GAC = Gratitude Adjective Checklist; GRAT–S = Gratitude Resentment Appreciation Test–Short Form; VIA–Y = Values in Action Inventory of Strengths for Youth.
Year
Authors
Table 1 Descriptive Characteristics of Studies Included in the Meta-Analysis
292 Renshaw and Olinger
Gratitude Meta-Analysis
293
most commonly used measure (eight studies), followed by the GAC (McCullough et al., 2002; seven studies), the Values in Action Inventory of Strengths for Youth (VIA–Y; Park & Peterson, 2006; four studies), and, lastly, the Gratitude Resentment Appreciation Test—Short Form (GRAT–S; Thomas & Watkins, 2003; two studies).
Systematic Review of Intervention Protocol Characteristics A systematic review of the descriptive characteristics of the five gratitude-based intervention studies indicated that each used a different experimental protocol. The first published intervention study, conducted by Froh et al. (2008), adapted a “counting blessings” exercise previously validated with adults (see Emmons & McCullough, 2003) for use with older children and early adolescents in a school setting. Eleven classrooms were randomly assigned to either the gratitude-based intervention or one of two control conditions. Participants in the intervention condition (n = 76) engaged in the “counting blessings” exercise, in which they wrote a list of up to five things they were grateful for since yesterday, once during each school day, every other day, and over the course of 2 weeks. Outcomes for this condition were compared with those of participants within an active-negative control condition (n = 80), who counted hassles instead of blessings, and a passive control or “education-as-usual” condition (n = 65). The second intervention study, conducted by Froh, Kashdan et al. (2009), adapted a gratitude letter-writing exercise also previously validated with adults (see Watkins, Woodward, Stone, & Kolts, 2003) for use with students in a primary–secondary parochial school. Participants were matched by grade and then randomly assigned to the gratitude-based intervention or control condition. Those in the intervention condition (n = 44) wrote letters to persons who had been kind to them but whom they had not properly thanked, for 10 to 15 minutes during every school day over the course of 2 weeks. Outcomes for this condition were compared with those of participants in an active-positive control condition (n = 45), which required students to journal about activities they had engaged in the previous day, as well as how they felt about those activities. The third published gratitude-based intervention study, conducted by Akhtar and Boniwell (2010), differed substantively from the previous two intervention studies with regard to both protocol and sample characteristics. Unlike the previous two studies, which tested stand-alone gratitude writing exercises with developmentally typical youth in schools, this study included a combination of gratitude-based psychoeducation and exercises as a component of a larger positive psychology treatment package for adolescents enrolled in an alcohol and drug rehabilitation center. Thus, the study used a quasiexperimental research design that tested the effects of the intervention on preexisting groups of youth attending the center. The positive psychology curriculum consisted of eight weekly sessions, which targeted themes associated with subjective and objective well-being (e.g., gratitude, optimism, goal-setting, resilience), using group work, experiential exercises, and discussion. Outcomes for participants in the intervention condition (n = 10) were compared with those of participants within a passive control or treatment-as-usual condition (n = 10). The fourth published study, conducted by Owens and Patterson (2013), also differed substantively from the first two studies, although, similar to those studies, it tested a stand-alone gratitudebased exercise. In this study, participants were children enrolled in after-school care programs. Those in the intervention condition (n = 22) drew pictures of things they were thankful for once a week for 4 to 6 weeks. Outcomes for this condition were compared with those of participants within an alternative intervention condition (n = 23), who drew pictures of themselves in situations wherein they were “at their best,” as well as with those of participants in a positive-active control condition (n = 17) that involved drawing pictures of an activity they engaged in that day. Psychology in the Schools
DOI: 10.1002/pits
294
Renshaw and Olinger
The fifth gratitude-based intervention study consisted of two substudies, which were presented in a single article and conducted by Froh, Bono, Fan, and colleagues (2014). Although these studies also substantively differed from all previous intervention studies with youth, they both investigated a new gratitude-based intervention curriculum with young schoolchildren. Participating classrooms were randomly assigned to either the gratitude-based intervention or control conditions. In the first substudy, participants in the intervention condition (n = 62) received the Benefit Appraisal Curriculum, which comprised instruction targeting the social–cognitive perceptions that elicit gratitude, during 30 minutes each school day over the course of 1 week. Outcomes for this condition were compared with those of participants in a positive-active control condition (n = 60), who received similarly structured lesson plans that focused on emotionally neutral topics. The second substudy was similar to the first, using the Benefit Appraisal Curriculum for the intervention condition (n = 44) as well as the same procedures for the control condition (n = 40). However, unlike the previous study, the intervention and control conditions in this study were carried out once per week during the school day over the course of 5 weeks, instead of once per day during the course of 1 week. Taken together, then, these intervention studies represent five distinct approaches to gratitude-based intervention with youth, only one of which, to date, has been replicated (i.e., the Benefit Appraisal Curriculum). Meta-Analysis of the Correlates of Gratitude Coding of the 15 identified studies that reported empirical correlates of gratitude in youth yielded 29 distinct correlates, 14 of which were investigated in at least two samples and the remainder of which were investigated in only a single sample. Meta-analysis of the correlates derived from at least two samples revealed two correlates with large positive effect sizes (social integration, r = .63; future gratitude, r = .50), seven with medium positive effect sizes (other gratitude measures, r = .47; life satisfaction, r = .45; positive affect, r = .40; positive outlook, r = .40; positive self-appraisal, r = .37; happiness, r = .37; informant [i.e., teacher or parent]-rated gratitude, r = .31), two with small positive effect sizes (prosocial behavior, r = .26; parent well-being, r = .15), two with small negative effect sizes (negative affect, r = –.17; depression, r = –.28), and one with a medium negative effect size (athletic burnout, r = –.34). Table 2 provides a full presentation of the overall weighted mean r [95% CI] for each of these correlates, as well as summary effect sizes for both total positive and total negative correlates derived from these samples. Distributions of the weighted effect sizes across samples are provided for the two most commonly researched positive correlates (i.e., life satisfaction [18 samples] and positive affect [11 samples]) and negative correlates (i.e., negative affect [10 samples] and depression [seven samples]) via forest plots in Figures 1 and 2. Moreover, the systematic review of correlates derived from only a single sample yielded two medium positive effect sizes (i.e., agreeableness, r = .36; absorption, r = .34), seven small positive effect sizes (grade point average, r = .28; assortative relationship preferences, r = .24; openness, r = .23; neuroticism, r = .22; extraversion, r = .20; social support, r = .19; conscientiousness, r = .13), one negligible effect size (stressful life events, r = –.09), three small negative effect sizes (suicidality, r = –.10; somatic symptoms, r = –.16; psychological distress, r = –.21), and two medium negative effect sizes (materialism, r = –.34; envy, r = –.35). Table 3 provides a full presentation of these single-sample correlates, as well as summary effect sizes for both total positive and total negative correlates derived from these samples. Additionally, combining and weighting all samples represented in Tables 2 and 3, the meta-summary effect for total positive correlates was medium (r [95% CI] = .41 [.37, .45]; K = 15, k = 23, p < .001), whereas the meta-summary effect for total negative correlates was small (r [95% CI] = –.22 [–.28, –.15]; K = 8, k = 14, p < .001). Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
295
Table 2 Gratitude Correlates Derived from Multiple Samples Correlate
K
k
N
r [95% CI]
Q
I2
NOfs
Social Integration Future Gratitude Other Gratitude Life Satisfaction Positive Affect Positive Outlook Positive Self-Appraisal Happiness Informant Gratitude Prosocial Behavior Parent Wellbeing Negative Affect Depression Athletic Burnout Total Positive Total Negative
2 3 1 11 5 2 3 2 2 3 2 4 2 2 14 7
2 3 6 18 11 2 3 2 2 4 2 10 7 2 23 11
2,005 697 1,401 6,722 3,766 299 2,855 364 467 1,614 246 3,677 2,706 462 8,743 5,444
.63 [.24, .85]b .50 [.31, .65]a .47 [.40, .54]a .45 [.37, .52]a .40 [.34, .47]a .40 [.01, .68]c .37 [.27, .45]a .37 [.26, .47]a .31 [.22, .39]a .26 [.20, .32]a .15 [.02, .27]c –.17 [–.23, –.11]a –.28 [–.38, –.17]a –.34 [–.44, –.23]a .44 [.36, .51]a –.23 [–.30, –.16]a
119a 18a 15c 243a 49a 13a 13b 1ns 1.
intervention). Moreover, the distribution of weighted effect sizes across samples for the two most commonly researched intervention outcomes (i.e., positive affect [five studies] and negative affect [five studies]) are displayed via forest plots in Figures 3 and 4. Regarding the four intervention outcomes derived from multiple samples, the resulting Q and I2 values indicated that for positive affect, the variability in sample effect sizes could again be substantively accounted for by population heterogeneity or between-study variance (I2 > 25; see Table 2), whereas the outcome effect sizes appeared to be derived from more homogenous populations with less between-study variance (I2 < 25; see Heudo-Medina, Sanchez-Meca, Marin-Martinez, & Botella, 2006). These findings partially support the appropriateness of our a priori random-effects approach to meta-analysis, suggesting that, at least for positive affect, conclusions regarding the findings can be generalized to other populations of youth who were not directly represented in samples included in the present analyses (see Cook et al., 2010). However, the null findings for the other three multi-sample outcomes, as well as the lack of multiple samples for the majority of outcomes, suggest that more gratitude-based intervention research is warranted using more diverse samples prior to generalizing these findings to youth in schools more broadly. Additionally, for the four intervention outcomes derived from at least three samples, resulting Orwin’s fail-safe N values indicated that, for negative affect, affective gratitude, and life satisfaction, zero unpublished studies were needed—and, for positive affect, only one unpublished study was needed—indicating a null effect size (g = .00) to lower the overall weighted mean effect size for the outcome to a negligible level (g < .20).
FIGURE 4. Negative affect intervention outcomes: Forest plot of Hedge’s g effect sizes with 95% confidence intervals for the five samples included in the meta-analysis. Square = individual effect size estimate; diamond = overall effect size estimate; line = confidence interval; arrow = upper bound of confidence interval > 1. Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
299
D ISCUSSION Interpretation of Findings The purpose of the present study was to investigate the scientific goodness of gratitude in youth and schools by systematically reviewing and meta-analyzing its correlates and intervention outcomes in children and adolescents. We expected that findings from this study would help clarify the empirical status of gratitude within the developing subfield of positive youth psychology, but did not intend for such findings to resolve the larger conceptual issues related to gratitude in youth and schools, such as how best to define the gratitude construct or how to value the promotion of gratitude amid the many competing demands for other potential interventions. Indeed, such conceptual issues can be considered extra-scientific or pre-scientific, as they function as the impetus for the empirical work and must be evaluated on their own terms. That said, given school psychology and NASP’s apparent interest in the gratitude construct and, specifically, in gratitude-based interventions (see Froh & Bono, 2011; Froh et al., 2007; Molony & Henwood, 2010; NASP, 2009a, 2009b), we estimated that enough interested parties within the field of school psychology value the gratitude construct to make the present study useful for guiding practice and future research related to youth and schools. To begin interpreting findings from the present study, we suggest that the systematic review of descriptive characteristics obtained from the 20 identified studies included in the meta-analysis, which represented 28 independent samples, indicate that (a) empirical interest in gratitude in youth is relatively new (as the first study was published in 2006), (b) interest in the correlates of gratitude in youth has so far outweighed interest in the effects of gratitude-based interventions on youths’ outcomes (i.e., correlate study–outcome study ratio = 2.7:1), (c) most studies have targeted youth in late childhood or adolescence instead of early childhood, (d) some gratitude measures have been used much more frequently than have others (i.e., GQ–6 > GAC > VIA–Y > GRAT), and (e) there is little agreement on preferred intervention approaches for youth (as, to date, only one replication study has been conducted; i.e., Froh et al., 2014). Considered together, these descriptive findings suggest that the scientific study of gratitude in youth and schools is in a nascent stage, as it has yet to establish a well-rounded empirical basis that places equal emphasis on experimental designs, spans developmental levels of youth, cuts across settings, clarifies measurement issues, and refines intervention protocols. Given this context, we suggest that the meta-analytic findings discussed in this section are likely to be especially fruitful for guiding future research and practice relative to gratitude in youth. Furthermore, for readers who may be concerned about the small number of studies included in the present systematic review and meta-analysis (N = 20), especially those investigating intervention outcomes (n = 5), we wish to reiterate Valentine et al.’s (2010) response to the question, “How many studies do you need to do a meta-analysis?” In brief, Valentine et al. argue that a meta-analysis of two or more studies should be considered a more transparent and more valid indicator of the empirical status of a domain of inquiry than consideration of a single study or a narrative review of selected studies. Considered in this light, then, what follows is intended to be a careful and conservative interpretation of meta-analysis findings, followed by a discussion of key limitations. Results from the meta-analysis of correlates indicate that gratitude in youth is positively associated with several indicators of youths’ subjective well-being (e.g., positive affect, positive outlook, and positive self-appraisal), as well as negatively associated with several indicators of youths’ subjective distress (e.g., negative affect, depression, and somatic symptoms; see Tables 2 and 3). Such correlations provide convergent and discriminant validity evidence suggesting that gratitude has small to moderate associations with other indicator of subjective well-being in youth, at least for older children and adolescents. However, a closer look at some particular correlates suggests potenPsychology in the Schools
DOI: 10.1002/pits
300
Renshaw and Olinger
tial reason for concern regarding the empirical status of gratitude measures in youth, as well as the utility of gratitude measures in schools. First, the three samples investigating future gratitude (or gratitude at Time 2) yielded an overall r [95% CI] of .50 [.31, .65], suggesting relatively poor test–retest reliability and/or predictive validity, with one measure sharing only 25% of the variance with the other. It is noteworthy that such reliability differed substantially between measures, with the GQ–6 showing much higher correlations (r = .60; Chen & Chang, 2014) than the VIA–Y (r = .32; Park & Peterson, 2006), and that test–retest reliability and/or predictive validity has yet to be evaluated for the GAC and GRAT when used with youth samples. Likewise, it is potentially concerning that the concurrent cross-correlations of different gratitude measures (e.g., GQ–GAC, GAC–GRAT) yielded a similar overall effect size as the test–retest and/or predictive validity correlations, r [95% CI] = .47 [.40, .54], indicating that the measures share only about 22% of the total variance with each other and, thus, that they are measuring imprecise or, potentially, related-yet-distinct constructs. This latter claim—that different gratitude measures may be measuring different constructs—is further supported by the overall correlation between gratitude and life satisfaction (r [95% CI] = .45 [.37, .52]), which approximates the overall cross-correlation among gratitude measures and the observed concurrent relationship between gratitude and social integration (r [95% CI] = .63 [.24, .85]), which far exceeds the effect size of all gratitude–gratitude correlations. Second, findings from the meta-analysis of correlates also highlight some potential concerns regarding the empirical utility of gratitude for youth in school contexts. Although, as mentioned previously, the results clearly show that gratitude is substantively associated with measures of subjective well-being and distress, which are important dimensions of youths’ mental health, there is little evidence to suggest that gratitude in youth has practically meaningful associations with performance-based or informant-rated variables that educators and school psychologists tend to value. Specifically, although a medium effect size was observed for the association between gratitude and grade point average (r [95% CI] = .34 [.29, .39]; r2 = .12), this finding was derived from a single study and used self-reported (as opposed to school-reported or teacher-reported) grades (Froh, Emmons et al., 2011). Although a significant effect size was also noted in the relation between gratitude and prosocial behavior, this effect was notably small (r [95% CI] = .26 [.20, .32]; r2 = .07) and was based entirely on self-reported prosocial behaviors (Froh et al., 2010; Froh, Yurkewicz et al., 2009; Tian, Du, & Huebner, 2014), which may differ substantially from directly observed or informant-rated prosocial behaviors (see McCullough & Tabak, 2010). That said, there appears to be no available correlational evidence indicating associations between youths’ gratitude and directly observed or teacher-reported school functioning. Moreover, several of the correlates that have been investigated—including agreeableness, assortative relationship preferences, materialism, envy, openness, and neuroticism—seem to lack face validity within school contexts, as they measure constructs that are highly unlikely to be targeted via assessment and intervention practices for promoting student success. Thus, to make gratitude research more relevant for applied work with youth in schools, we recommend that much more correlational research be performed to investigate associations with what might be called the “bread-and-butter” variables of student success, such as academic skill proficiency, prosocial and academic behaviors, social– emotional skills (see Furlong et al., 2014), school climate indicators (see O’Malley, Voight, Renshaw, & Eklund, 2014), and school-specific subjective well-being (see Renshaw, Long, & Cook, 2015). In addition to more correlational research, findings from the present study also suggest that increased attention to and testing of the effects of gratitude-based interventions with youth and in schools is sorely needed. Results from the meta-analysis of intervention effects indicate that, on the whole, gratitude exercises for promoting youths’ subjective well-being and decreasing subjective distress are generally ineffective, although isolated studies have deemed them effective for select Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
301
outcomes. Specifically, the summary effect size for all outcomes across all samples yielded a negligible total intervention effect (g [95% CI] = .18 [–.02, .39]), with similarly negligible findings for the summary effects across post-test outcomes (g [95% CI] = .19 [–.01, .39]) and follow-up outcomes (g [95% CI] = .11 [–.11, .33]). Again, similarly negligible outcomes were observed for the summary effects when both positive-active control conditions (g [95% CI] = .18 [–.04, .40]) and negative-active control conditions (g [95% CI] = .14 [–.18, .45]) were used as comparison groups, yet a small positive effect size was observed when passive control conditions were used as the comparison group (g [95% CI] = .49 [–.50, 1.47]). Moreover, a small negative effect size (g [95% CI] = –.33 [–1.18, .53]) was observed in the one study comparing a gratitude-based intervention with a structurally similar alternative intervention (Owens & Patterson, 2013), indicating that the alternative intervention was actually more effective. Taken together, then, the upshot of these summary effect sizes is that gratitude-based interventions are, on the whole, likely more effective than doing nothing different (i.e., passive control or education-as-usual), but are not likely practically different from the effects yielded when doing something different (i.e., active control or alternative intervention). That said, although the majority of summary findings yielded null overall intervention effects, the meta-analysis findings for individual outcomes suggest several positive effects that warrant further research. Regarding these positive outcomes, the only one derived from several samples was positive affect, which yielded a small effect size (g [95% CI] = .23 [–.15, .61]; see Figure 4). The three largest positive outcomes, for optimism (g [95% CI] = 1.71 [.72, 2.70]), happiness (g [95% CI = 1.44 [.49, 2.39]), and alcohol dependence (g [95% CI] = .43 [–.42, 1.28]), were all derived from the same single-sample study (Akhtar & Boniwell, 2010), which was exceptional compared with the other gratitude-based intervention studies in several respects: (a) it was the only one to target participants with identified preexisting psychosocial impairment (i.e., attending an alcohol and drug treatment rehabilitation center), (b) it had by far the smallest sample size (N = 20), and (c) it used the most robust intervention (a positive psychology treatment package including a core gratitude component) that was (d) carried out over the longest duration (8 weeks). Therefore, prior to deeming this gratitude-based intervention effective for promoting these aspects of subjective well-being in youth, much more research is needed to test the effects of this positive psychology treatment package—and, specifically, its gratitude component—with more diverse and typical samples of youth, especially in schools. Finally, one additional intervention, the Benefits Appraisal Curriculum (Froh et al., 2014), yielded a small positive effect size across two studies for the benefit appraisals outcome (g [95% CI] = .43 [.15, .70]) and a small effect in one study for the behavioral gratitude outcome (i.e., writing gratitude notes; g [95% CI] = .40 [.03, .76]), suggesting that future research is also warranted to investigate the efficacy of this particular protocol. However, prior to further empirical validation, we recommend that discussion is warranted regarding the utility of these particular outcomes, which are entailed by the nature of the curriculum itself, as intervention targets within school psychological service delivery. It is to this issue that we will turn in closing, after discussing some key limitations. Limitations of the Present Study Given that the present study only used published original empirical studies of gratitude in youth for the systematic review and meta-analysis, the primary limitation of this work is likely the ubiquitous potential for publication, or file-drawer, bias. In an attempt to account for this potential bias, we calculated Orwin’s fail-safe N (NOfs ), as opposed to the classic fail-safe N, for all correlates (see Table 2) and intervention outcomes (see Table 4) derived from at least three samples. NOfs values from the meta-analysis of correlates derived from these samples indicated that many unpublished studies (range: 9–82) would be needed to void the substantive effect sizes observed in the published Psychology in the Schools
DOI: 10.1002/pits
302
Renshaw and Olinger
samples, suggesting that publication bias for at least these eight correlates is unlikely. Yet, there were six additional correlates derived from two samples (see Table 2) and 14 correlates derived from single samples (see Table 3) that also yielded practically meaningful effect sizes, but, because of their few-sample status, they were all more amenable to publication bias (although NOfs values could not technically be calculated) and therefore warrant further research. Moreover, given the generally null and unfavorable summary effect sizes observed for the gratitude-based intervention effects, excepting passive control comparisons (see Table 4), potential publication bias is not necessarily a concern, as the effects are already practically negligible on the whole. Yet, given that the majority of intervention effects were only investigated in one or two studies, representing outcomes at the individual level (as opposed to the summary or meta-analytic level) is more prone to publication bias and, thus, requires cautious interpretation and careful, if any, generalization. Implications for Theory and Practice Findings from the present meta-analysis of the correlates and intervention outcomes of gratitude in youth have a few key implications for gratitude theory and gratitude-based practices in schools. To begin, as mentioned earlier (see the Introduction section), much has been written and otherwise publicly presented, and often with much enthusiasm, regarding the empirical utility of gratitude in youth (see Bono, 2014, September 29; Froh & Bono, 2011, 2012, 2014; Froh et al., 2007). Furthermore, many in the field of school psychology, including NASP, have advocated for the use of gratitude-based interventions in schools (e.g., NASP, 2009a, 2009b, 2013). Perhaps the strongest articulation of the empirical utility of gratitude in youth was offered by Bono et al. (2014), who claimed that (a) “gratitude [in youth] is particularly suited to building and strengthening supportive ties” (p. 69), (b) gratitude-based interventions could be an integral practice for “turning schools into environments that do not just address deficit and pathology but enhance strengths and psychological and social development” (p. 70), (c) gratitude-based practices “should help youth fulfill their potential and find greater meaning, success, and well-being—that is, to flourish” (p. 73), and (d) “gratitude felt and expressed by students and the improved behaviors that could ensue would likely spread to teachers and staff, encouraging them to work harder on students’ behalf and helping to prevent burnout” (p. 76). Despite such enthusiasm, Bono et al. (2014) also noted that gratitude research “is still in its early stages,” that “the implementation of school-based programs to enhance gratitude is limited in scope” (p. 68), and that “the development of gratitude or factors that promote or inhibit its emergence remain empirically uncharted” (p. 69). Considering the findings from the present meta-analysis, we suggest that the current empirical status of gratitude in youth and schools lends much support to these latter caveats and very little support to the aforementioned enthusiasm. Regarding theory, although there is correlational evidence indicating that gratitude in youth is likely an indicator of subjective well-being and that it is positively associated with social integration, prosocial behavior, and social support (see Tables 2 and 3), there is currently no direct evidence indicating that gratitude-based interventions build or strengthen relationships among youth or between youth and caregivers (see Table 4). There is also no evidence that such interventions improve academic performance or behaviors of any kind. There is, however, some evidence suggesting that gratitude-based interventions may be useful for enhancing other areas of youth development—but, considered in context, generalizations of this evidence to typical youth in schools is ill-advised and warrants further investigation. Regarding school-based practice, then, our interpretation of the empirical evidence presented herein is that gratitude-based interventions are not currently advised as an approach for improving prosocial behavior, relationship quality, or subjective well-being, which are the purposes for which Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
303
they have been primarily recommended, and that there is currently no empirical warrant for implementing them “as early in life as possible” to help boost students’ interest and success in school, as has also been recommended (see Froh, Bono, Fan et al., 2014, p. 148). Indeed, there is not even evidence to suggest that gratitude-based interventions are likely to improve youths’ affective gratitude; yet, this null effect may be confounded by the poor construct validity of self-report gratitude measures for youth, as discussed earlier. Considering these findings and implications, we wish to close by drawing attention to the practicality and utility of the correlates and intervention outcomes that have been targeted thus far via gratitude-based research in schools. Although evidence indicates, for instance, that gratitude has small positive correlations with common personality traits, such as agreeableness and extraversion, what good are these findings for informing the practice of psychology in the schools? In addition, if the correlational evidence suggests suboptimal measurement of self-reported gratitude to begin with, how much stock should scholars and practitioners place on this construct in the first place? Moreover, although evidence suggests, for example, that a particular gratitude-based curriculum may enhance youths’ ability to appraise benefits and increase their likelihood of writing gratitude letters, what is the value of these outcomes compared with other common school outcomes (e.g., improving academic performance and social behavior)? Also, what are the trade-offs of focusing on improving gratitude-related skills instead of targeting other potentially useful behaviors for intervention? Our purpose in raising these questions is not to provide answers, but rather, to spark further discussion about the goodness of gratitude in youth and, especially, the goodness-of-fit between the values of researchers interested in gratitude in youth and the values of educators and school psychologists practicing in schools. Given that there is always limited time and resources when working in schools, we believe that clarifying these conceptual issues is of utmost importance for shaping future empirical work targeting gratitude in youth and schools.
R EFERENCES References marked with an asterisk indicate studies included in the meta-analysis. *Akhtar, M., & Boniwell, I. (2010). Applying positive psychology to alcohol misusing adolescent: A group intervention. Groupwork, 20(3), 6–31. Baumgarten-Tramer, F. (1938). “Gratefulness” in children and young people. The Journal of Genetic Psychology, 53, 53–66. Beretvas, S. N. (2010). Meta-analysis. In G. R. Hancock & R. O. Mueller (Eds.), The reviewer’s guide to quantitative methods in the social sciences (pp. 255–263). New York, NY: Routledge. Bono, G. (2014, September 29). How can we cultivate gratitude in schools? Berkeley, CA: Greater Good Science Center. Retrieved from http://goo.gl/0ZaFai Bono, G., Froh, J. J., & Forrett, R. (2014). Gratitude in school: Benefits to students and schools. In M. J. Furlong, R. Gilman, & E. S. Huebner (Eds.), Handbook of positive psychology in schools (2nd ed., pp. 67–81). New York, NY: Routledge. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2006–2014). Comprehensive Meta-Analysis, Version 2. Englewood, NJ: Biostat. Retrieved from www.meta-analysis.com *Chen, L. H., & Chang, Y-P. (2014). Cross-lagged association between gratitude and adolescent athlete burnout. Current Psychology. Advance online publication. *Chen, L. H., & Kee, Y. H. (2008). Gratitude and adolescent athletes’ well-being. Social Indicators Research, 89, 361–373. Cook, C. R., Williams, K. R., Guerra, N. G., Kim, T. E., & Sadek, S. (2010). Predictors of bullying and victimization in childhood and adolescence: A meta-analytic investigation. School Psychology Quarterly, 25, 65–83. *Datu, J. A. D. (2014). Forgiveness, gratitude and subjective wellbeing among Filipino adolescents. International Journal for the Advancement of Counselling, 36, 262–273. Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: An experimental investigation of gratitude and subjective well-being in daily life. Journal of Personality and Social Psychology, 84, 377–389. Ervin, R. A., Gimpel Peacock, G., & Merrell, K. W. (2010). The school psychologist as a problem solver in the 21st century: Rationale and role definition. In G. Gimpel Peacock, R. A. Ervin, E. J. Daly, & K. W. Merrell (Eds.), Practical handbook of school psychology: Effective practices for the 21st century (pp. 3–12). New York, NY: Guilford. Psychology in the Schools
DOI: 10.1002/pits
304
Renshaw and Olinger
Field, A. P. (2009). Meta-analysis. In R. E. Millsap & A. Maydeu-Olivares (Eds.), The SAGE handbook of quantitative methods in psychology (pp. 404–422). Thousand Oaks, CA. Freitas, L. B. L., Pieta, M. A. H., & Tudge, J. R. H. (2011). Beyond politeness: The expression of gratitude in children and adolescents. Psicologia: Reflex˜ao e Cr´ıtica, 24, 757–764. Retrieved from www.scielo.br/prc Froh, J. J., & Bono, G. (2011). Gratitude in youth: A review of the gratitude interventions and some ideas for applications. Communiqu´e, 39, 1, 26–28. Froh, J. J., & Bono, G. (2012, November 19). How to foster gratitude in schools. Berkeley, CA: Greater Good Science Center. Retrieved from www.greatergood.berkeley.edu Froh, J. J., & Bono, G. (2014). Making grateful kids: The science of building character. West Conshohocken, PA: Templeton. *Froh, J. J., Bono, G., & Emmons, R. (2010). Being grateful is beyond good manners: Gratitude and motivation to contribute to society among early adolescents. Motivation and Emotion, 34, 144–157. *Froh, J. J., Bono, G., Fan, J., Emmons, R. A., Henderson, K., Cheray, H. Leggio, H., & Wood, A. M. (2014). Nice thinking! An educational intervention that teaches children to think gratefully. School Psychology Review, 43, 132–152. *Froh, J. J., Fan, J., Emmons, R. A., Bono, G., Huebner, E. S., & Watkins, P. (2011). Measuring gratitude in youth: Assessing the psychometric properties of adult gratitude scales in children and adolescents. Psychological Assessment, 23, 311–324. *Froh, J. J., Emmons, R. A., Card, N. A., Bono, G., & Wilson, J. A. (2011). Gratitude and the reduced costs of materialism in adolescents. Journal of Happiness Studies, 12, 289–302. *Froh, J. J., Kashdan, T. B., Ozimkowski, K. M., & Miller, N. (2009). Who benefits most from a gratitude intervention in children and adolescents? Examining positive affect as a moderator. The Journal of Positive Psychology, 4, 408–422. Froh, J. J., Miller, D. N., & Snyder, S. F. (2007). Gratitude in children and adolescents: Development, assessment, and school-based intervention. School Psychology Forum: Research in Practice, 2, 1–13. *Froh, J. J., Sefick, W. J., & Emmons, R. A. (2008). Counting blessings in early adolescents: An experimental study of gratitude and subjective well-being. Journal of School Psychology, 46, 213–233. *Froh, J. J., Yurkewicz, C., & Kashdan, T. B. (2009). Gratitude and subjective well-being in early adolescence: Examining gender differences. Journal of Adolescence, 32, 633–650. Furlong, M. J., Gilman, R., & Huebner, E. S. (2014).(Eds.). Handbook of positive psychology in schools (2nd ed.). New York, NY: Routledge. Furlong, M. J., You, S., Renshaw, T. L., O’Malley, M. D., & Rebelez, J. (2013). Preliminary development of the positive experiences at school scale. Child Indicators Research, 6, 753–775. Furlong, M. J., You, S., Renshaw, T. L., Smith, D. C., & O’Malley, M. D. (2014). Preliminary development and validation of the Social and Emotional Health Survey for secondary students. Social Indicators Research, 117, 1011–1032. Gillham, J., Adams-Deutsch, A., Werner, J., Reivich, K., Coulter-Heindl, V., Linkins, M., . . . Seligman, M. E. P. (2011). Character strengths predict subjective well-being during adolescence. The Journal of Positive Psychology, 6, 31–44. Gordon, A. K., Musher-Eizenman, D. R., Holub, S. C., & Dalrymple, J. (2004). What are children thankful for? An archival analysis of gratitude before and after the attacks of September 11. Applied Developmental Psychology, 25, 541– 553. Hawkins, R. O., Barnett, D. W., Morrison, J. Q., & Musti-Rao, S. (2010). Choosing targets for assessment and intervention. In G. Gimpel Peacock, R. A. Ervin, E. J. Daly, & K. W. Merrell (Eds.), Practical handbook of school psychology: Effective practices for the 21st century. New York, NY: Guilford. *Hoy, B. D., Suldo, S. M., & Mendez, L. R. (2013). Links between parent’s and children’s levels of gratitude, life satisfaction, and hope. Journal of Happiness Studies, 14, 1343–1361. Heudo-Medina, T., Sanchez-Meca, J., Marin-Martinez, F., & Botella, J. (2006). Assessing heterogeneity in meta-analysis: Q statistic or I2 index? CHIP Documents, Paper 19. Retrieved from http://digitalcommons.uconn.edu/chip_docs/19 Huebner, E. S., & Hills, K. J. (2011). Does the positive psychology movement have legs for children in schools? The Journal of Positive Psychology, 6, 88–94. Jimerson, S. R., Sharkey, J. D., Nyborg, V., & Furlong, M. J. (2004). Strength-based assessment and school psychology: A summary and synthesis. The California School Psychologist, 9, 9–19. Kirschman, K. J. B., Johnson, R. J., Bender, J. A., & Roberts, M. C. (2009). Positive psychology for children and adolescents: Development, prevention, and promotion. In S. J. Lopez & C. R. Snyder (Eds.), The Oxford handbook of positive psychology (2nd ed., pp. 133–148). New York, NY: Oxford. Larson, R. W. (2000). Toward a psychology of positive youth development. American Psychologist, 55, 170–183. *Li, D., Zhang, W., Li, X., Li., N., & Ye, B. (2012). Gratitude and suicidal ideation and suicide attempts among Chinese adolescents: Direct, mediated, and moderated effects. Journal of Adolescence, 35, 55–66. Masten, A. S. (2001). Ordinary magic: Resilience process in development. American Psychologist, 56, 227–238. McCullough, M. E., Emmons, R. A., & Tsang, J.-A. (2002). The grateful disposition: A conceptual and empirical topography. Journal of Personality and Social Psychology, 82, 112–127.
Psychology in the Schools
DOI: 10.1002/pits
Gratitude Meta-Analysis
305
McCullough, M. E., & Tabak, B. A. (2010). Prosocial behavior. In R. F. Baumeister & E. J. Finkel (Eds.), Advanced social psychology: The state of the science (pp. 263–302). New York, NY: Oxford. Molony, T., & Henwood, M. (2010). Gratitude works in schools. Communiqu´e, 38, 6. Retrieved from www.nasponline.org National Association of School Psychologists. (2009a). Fostering an attitude of gratitude: Tips for parents. Bethesda, MD: Author. Retrieved from www.nasponline.org National Association of School Psychologists. (2009b). Gratitude Works program guidelines. Bethesda, MD: Author. Retrieved from www.nasponline.org National Association of School Psychologists. (2013). School Psychology Awareness Week 2013. Bethesda, MD: Author. Retrieved from www.nasponline.org O’Malley, M. D., Voight, A., Renshaw, T. L., & Eklund, K. (2014). School climate, family structure, and academic achievement: A study of moderation effects. School Psychology Quarterly. Advance online publication. *Owens, R. L., & Patterson, M. M. (2013). Positive psychological interventions for children: A comparison of gratitude and best possible selves approaches. The Journal of Genetic Psychology, 17, 403–428. *Park, N., & Peterson, C. (2006). Moral competence and character strengths among adolescents: The development and validation of the Values in Action Inventory of Strengths for Youth. Journal of Adolescence, 29, 891–909. Peterson, C., & Seligman, M. E. P. (2004). Character strengths and virtues: A handbook and classificaiton. Washington, DC: American Psychological Association. Reivich, K. (2009). Fishful Thinking: Cultivating gratitude in youth. Communiqu´e, 38, 1, 20–22. Renshaw, T. L., Long, A. C. J., & Cook, C. R. (2015). Assessing adolescents’ positive psychological functioning at school: Development and validation of the Student Subjective Wellbeing Questionnaire. School Psychology Quarterly, 30, 534–552. doi:10.1037/spq0000088. *Ruch, W., Weber, M., Park, N., & Peterson, C. (2014). Character strengths in children and adolescents: Reliability and initial validity of the German Values in Action Inventory of Strengths for Youth (German VIA-Youth). European Journal of Psychological Assessment, 30, 57–64. Schmidt, F. L. (1992). What do data really mean? Research findings, meta-analysis, and cumulative knowledge in psychology. American Psychologist, 47, 1173–1181. 3 Thomas, M., & Watkins, P. (2003, April). Measuring the grateful trait: Development of the revised GRAT. Poster presented at the Annual Convention of the Western Psychological Association, Vancouver, British Columbia, Canada. *Tian, L., Du, M., & Huebner, E. S. (2014). The effect of gratitude on elementary school students’ subjective well-being in schools: The mediating role of prosocial behavior. Social Indicators Research. Advance online publication. Valentine, T. D., Pigott, J. C., & Rothstein, H. R. (2010). How many studies do you need? A primer on statistical power for meta-analysis. Journal of Behavioral and Educational Statistics, 35, 215–247. *van Eeden, C., Wissing, P., M., Dreyer, J., Park, N., & Peterson, C. (2008). Validation of the Values in Action Inventory of Strengths for Youth (VIA–Youth) among South African learners. Journal of Psychology in Africa, 18, 143–154. Watkins, P. C., Woodward, K., Stone, T., & Kolts, R. D. (2003). Gratitude and happiness: The development of a measure of gratitude and its relationship with subjective well-being. Social Behavior and Personality, 31, 431–452. *Weber, M., & Ruch, W. (2012). The role of character strengths in adolescent romantic relationships: An initial study on partner selection and mate’s life satisfaction. Journal of Adolescence, 35, 1537–2546. Weber, M., Ruch, W., Littman-Ovadia, H., Lavy, S., & Gai, O. (2013). Relationships among higher-order strengths factors, subjective well-being, and general self-efficacy: The case of Israeli adolescents. Personality and Individual Differences, 55, 322–327. Williams, M.-J. (2012, December 19). This holiday season, give your kids the gift of gratitude. The Washington Post. Retrieved from www.washingtonpost.com Wood, A. (2014, September 22). The trouble with the science of gratitude. Greater Good Science Center. Retrieved from http://goo.gl/YzIFUK You, S., Furlong, M. J., Dowdy, E., Renshaw, T. L., Smith, D. C., & O’Malley, M. D. (2013). Further validation of the Social and Emotional Health Survey for high school students. Applied Research in Quality of Life. Advance online publication.
Psychology in the Schools
DOI: 10.1002/pits