A Systematic Review with a Meta-Analysis of the Motivational Climate and Hedonic Well-Being Constructs: The Importance of the Athlete Level

Motivational climate is known to relate to individual behaviors, emotions, and thoughts. Hedonic or subjective well-being includes self-assessed positive affect (i.e., pleasant affect, moods, and emotions), negative affect (i.e., unpleasant affect, moods, and emotions), and life or domain-specific satisfaction. The aim of this review was to quantify the relationships between task and ego motivational climate scales and measures representing hedonic well-being with sports participants. Potential moderators of the motivational climate and hedonic well-being were examined. This review followed the PRISMA guidelines (PROSPERO ID CRD42023470462, registered 28 October 2023). From five relevant databases, one relevant review, and hand searching, 82 articles totaling 26,378 participants (46.3% female) met the inclusion criteria. The articles spanned publication dates from 1993 to 2023, representing 18 countries, various team and individual sports, and athletes competing in elite (e.g., Olympic) to grassroot (e.g., club sport) competitions. To meta-analyze the motivational climate and hedonic well-being relationships, the random-effects model was used. For the moderation analyses, the mixed-effects model was used. The task or mastery climate relationships were medium in magnitude with positive affect and satisfaction and small with negative affect. The ego or performance climate relationships were small in magnitude for positive affect, negative affect, and satisfaction. Evidence of bias existed in the motivational climate and hedonic well-being relationships. For moderation analyses, athlete level (i.e., elite vs. non-elite) moderated (p < 0.05) the task (elite, r = 0.23; non-elite, r = 0.34) and ego motivational climate (elite, r = −0.02; non-elite, r = −0.13) and positive affect and satisfaction combined relationships. In conclusion, the motivational climate and hedonic well-being relationships were stronger for the task climate than for the ego climate. The finding that elite athlete correlations appeared dampened is important for future research. Even with the damped relationships, practitioners, from the Olympics to local clubs, should ensure the promotion of the task climate to maximize positive affect and satisfactions in and around the sport experience.


Introduction
Achievement Goal Theory (AGT) research began in the late 1970s, leading to influential publications in the 1980s [1][2][3][4][5].In the sport and physical activity domains, research has flourished, resulting in quantitative reviews [6][7][8].Intertwined with the flourishing of AGT research was great interest in the athletes' motivations, cognitions, continued participation, and well-being as influenced by their coaches, and parents/peers [9,10].To best study these relationships, sport psychology researchers began creating and validating motivational climate measures with two appearing in 1992, the Perceived Motivational Climate in Sport Questionnaire [11] and the Parent-initiated Motivational Climate Questionnaire [12].Of the correlates often studied with motivational climate measures [9], no review has with intention examined hedonic well-being or subjective well-being [13] in the sport context or explored potential moderators.Hedonic or subjective well-being appeared in the literature in the 1980s along with AGT and is a meaningful psychological construct in the human experience comprised of positive and negative affect, and life-or domain-specific satisfaction such as sport.Well-being, hedonic and eudaimonic, within sport and exercise psychology is becoming a popular research topic [14].Hence, this systematic review aimed to quantify the task and ego motivational climates and measures consistent with the three components of subjective or hedonic well-being relationships.

AGT and Motivational Climate History
AGT is one of the most researched motivation theories across education, psychology, and business.AGT, including both the individual's predisposition and situational influences (e.g., teachers), originated from independent and collaborative research teams in education [1][2][3][4][5].Via Professor Glyn C. Roberts being part of the early discussions at the University of Illinois, sport researchers began studying and publishing on AGT [15,16].Since the 1980s, books [17], meta-analyses [6][7][8]18], and influential articles [19,20] multiplied and thus provide all the relevant information and background of AGT to interested readers.In addition, pertinent to this review, Ntoumanis and Biddle [10] published a review in 1999.More than 20 years later, Lacerda and colleagues provided an extensive review of motivational climate measures in sport [21].Across sport and physical activity, Harwood et al.'s quantitative review provides a comprehensive listing of motivational climate measures [9].Thus, we wrote a brief review of Nicholls' AGT and motivational climate measures.
Nicholls [5] built his framework upon the following assumptions: individuals operate in a rational manner and the predominant achievement goal guides the individual's decisions and behaviors in achievement contexts.The demonstration of competence is the goal of action in AGT frameworks.Thus, individual ability perceptions are central to AGT.How individuals reference ability perceptions refer to conceptions of ability.Nicholls theorized ability in two concepts, differentiated and undifferentiated.These ability conceptions define the task or mastery and ego or performance achievement goals, both of which are assumed to be orthogonal and implicit.These two implicit orientations are theorized to determine the vast array of beliefs, emotions, cognitions, and behaviors within achievement settings.Also, both goal orientations reflect ways in which individuals or athletes, the focus of this quantitative review, define success and failure and ways competence is inferred.The task orientation is adopted when personal mastery, achievement of higher ability, and improvement are the prime reasons for motivation.When a task orientation is the focus, athletes define success and failure by self-referenced perceptions of their performance.In contrast to a task orientation, an ego orientation is characterized when an individual's motivation for action is to demonstrate competence, defined by demonstrating superior ability or beating an opponent.Hence, self-comparisons define a task orientation, and other comparisons define an ego orientation.
An athlete's task or mastery or ego or performance involvement is determined by their proneness for each goal state and the current perceived situation [1,2].As found in Table 1, the PMCSQ, PMCSQ-2, and MCSYS were developed and incorporated into research agendas [9].The PMCSQ includes two subscales, whereas the second generation of the PMCSQ includes three subscales for each achievement goal orientation.Within their long history of youth sport research at the University of Washington, Smith and Smoll [22] developed a 12-item mastery and ego motivational climate measure.Most recent in the line of motivational climate measures, Appleton and colleagues developed the coach-created Empowering and Disempowering Motivational Climate Questionnaire [23].This scale includes task-and ego-involving subscales in addition to autonomy-supportive, socially supportive, and controlling coach subscales.For the purpose of this quantitative review, we included all subscales measuring AGT task or mastery and ego or performance subscales.The two main aims of this quantitative review were to update and extend knowledge of the relationships between the dichotomous motivational climates, task and ego, and hedonic well-being constructs researched in a sport.To our knowledge, only two motivational climate meta-analyses exist in the physical activity domains [9,25] as opposed to motivational climate as one of many correlates (e.g., [7]).Braithwaite and colleagues meta-analyzed motivational climate interventions in PE settings.They reported that task climate interventions improved student self-rated enjoyment (g = 0.15) and decreased anxiety (g = −0.25)and boredom (g = −0.27)along with other outcome variables.Over their 17 categories of correlates, Harwood and colleagues reported upon two main components of hedonic well-being, positive and negative affect.The reported effect size values were consistent with AGT as the task climate was positively related to positive affect (r = 0.47) and negatively related to negative affect (r = −0.17).In contrast, the ego climate was negatively related to positive affect (r = −0.11)and positively related to negative affect (r = 0.25).To expand upon Harwood et al.'s review, we searched a broad range of potential positive and negative affect constructs as well as searching for satisfaction.Our secondary aim was to explore potential moderators such as sample makeup (i.e., percent females), athlete level (i.e., elite vs. non-elite), and the sport type (i.e., individual vs. team) of the quantified motivational climates and hedonic well-being relationships.
Based on longstanding AGT proposed and verified relationships, we hypothesized the task climate to correlate positively with positive affect and satisfaction and negatively with negative affect.Conversely, we hypothesized the ego goal climate to be negatively related to positive affect and satisfaction and positively related with negative affect.Regarding our proposed moderators, the motivational climate literature is absent from moderator testing of relationships.Differences between male and female participants within dichotomous AGT studies stem from Duda [19] hypothesizing females to endorse the task orientation more than males and males to endorse the ego orientation more than females.Lochbaum and his colleague [1] found support, with studies using one of the main AGT measures, for males endorsing the ego orientation more than females.Whether differences in relationships exist, with males potentially being more sensitive to an ego motivational climate, is unknown; yet it is testable.From a series of Norwegian elite athlete research studies, the authors suggested that elite athletes might be more sensitive to their motivational climates [6] and the climate is more influential [26].Whether this sensitivity or ability to be influenced changes the relationships among the two motivational climates and our hedonic-based correlates is unknown, but worthy of investigation.Concerning sport type, some evidence exists that individual sport athletes are more ego-oriented than team sport athletes [6].Again, as with our other potential moderators, whether individual athletes are more sensitive to an ego/performance climate, and this sensitivity's impact on the meta-analyzed relationships, is unknown.By using the updated CMA program, our quantitative statistics are more comprehensive than the previous motivational climate quantitative review [9] and thus there is potential for some insights not yet found in the literature such as in addition to our moderator tests.

Materials and Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [27] guided all aspects of this manuscript.The PRISMA checklist corresponding to this manuscript is found in Supplemental Table S1.For our computations and result interpretations, we utilized Borenstein, Hedges, Higgins, and Rothstein's Comprehensive Meta-Analyses (CMA) Version 4 program and materials [28][29][30].The registration information is as follows: PROSPERO ID CRD42023470462, registered 28 October 2023.To avoid self-plagiarism, our methodology and such aspects' subheadings, table titles, and figure captions come from both authors' recent meta-analyses [31][32][33].

Eligibility Criteria and Selection Process
The inclusion criteria were as follows: (a) a task/mastery or ego/performance motivation climate measure, (b) a hedonic well-being measure, (c) participants engaged in a sport, (d) sufficient data provided for effect size calculation between at least one motivational climate and one hedonic well-being measure, and (e) original data published in a peerreviewed academic journal.The main exclusion criteria for studies reporting a climate and hedonic well-being measure were as follows: participants sampled in a non-sport setting (e.g., physical education class or leisure non-competitive settings such as exercising at a fitness club) or insufficient data for effect size calculation.
Concerning our search terms, we searched terms within the well-being domain such as flourishing, resilience, burnout, positive affect, negative affect, mental health, depression, satisfaction with life, and satisfaction, with the goal of capturing all relevant studies.Variables such as perceived competence, self-efficacy, confidence, and physical well-being (e.g., injuries) were excluded from the search.Inquiry about missing data or need for clarifications of any kind did not occur.For articles in a language other than English, we used Google Translate, https://translate.google.com/(accessed on on 1 December 2023), to help find the required data and coding information.

Information Sources, Search Strategy, and Search Protocol
As detailed here and in Figure 1, information sources included references from Harwood et al., databases found within EBSCOhost (search ended 1 November 2023), and hand searching (search ended 1 November 2023).Within EBSCOhost, we selected the following databases: APA PsycArticles, ERIC, Psychology and Behavioral Sciences Collection, PsychINFO, and SPORTDiscus.All search details are outlined below with more details located in Supplemental Table S2.[9] (n = 1 non-duplicate).

Data Collection and Items Retrieved
The developed data collection worksheet followed past co-authored systematic reviews and meta-analyses [32,33] with the following data retrieved: climate measurement name, context (sport, PE, or leisure), participant description (e.g., athletes, PE students), correlate data found (yes, no), climate referenced agent other than coach/team (e.g., peers, mother, father), % female participants, participant athletic level description (e.g., Olympic, regional, grassroots, club, university), mean age or age range, sport, apparent country of most participants, well-being measure name, and citation.We used Lochbaum, Cooper, and Limps' [33] classification system (Table 2) to organize the article-published participant descriptions as best as possible.3 contains the quality questions from Kmet et al. [35].Both authors rated the studies together with discrepancies discussed until agreement.Based on the question and rating explanations, we eliminated questions 5-7, 9, and 12. Scoring for each question followed Kmet's system of 2, 1, or 0. For the risk of bias across studies often referred to as publication bias, we used the following statistics: Orwin's fail-safe n [36], the classic fail-safe n [37], the funnel plot [38], and Duval and Tweedie's trim and fill [39].Orwin's fail-safe n represents the potential missed studies that would move the correlation past a predetermined threshold.We chose zero as our missed study value and 0.10 or −0.10 as this is the threshold for a small in meaningfulness interpretation.Thus, the greater the value for both fail-safe n calculations, the greater the confidence that the result is safe from bias.The classic fail-safe n statistic represents the number of null samples required to change a significant value into a nonsignificant value.We specified the one-tailed test when we conducted the classic fail-safe n analysis.Funnel plots were examined to determine whether the entered studies dispersed in a comparable manner on either side of the overall effect.Symmetry indicates that the retrieved studies captured the essence of all studies.For our last risk of bias across studies metric, we examined Duval and Tweedie's trim and fill analysis.The trim and fill analyses are used to adjust for potential missing studies.Data points filled to the right increase the effect size value, whereas those filled to the left lower the effect size value.

Summary Statistics, Planned Analyses, and Certainty Assessment
The correlation coefficient (r) was the summary statistic.The coefficient was based on the random effects model.The random effects model is the logical model given the gathered studies are best thought of as a random sampling of studies published in the literature [30].Cohen's [40] guidelines of 0.10-0.29 as small, 0.30-0.49as medium, and 0.50 as large defined meaningfulness.For the most parsimonious and least interrelated summary statistics, we reported only one summary statistic for our six relationships per study.Hence, if a study reported more than one negative affect or mood or the subscales of one measure, those were combined to one effect size.For each overall relationship (e.g., task climate and satisfaction), the number of samples, summary statistic, 95% confidence and prediction intervals, Tau-squared (τ 2 ) and I-squared (I 2 ), and publication bias statistics were reported.To examine the proposed categorical moderators, a mixed-effects model was used for the calculations.For Orwin's n, only the fixed-effects analysis is provided in the CMA program.To assess the potential impact of sample sex makeup, we used a random effects meta-regression model.
To further assess robustness, we conducted the remove-one study and cumulative analyses provided in the CMA program in addition to the classic fail-safe n and Orwin's fail-safe n, both of which provide statistics indicating robustness.The remove-one study remove-one analysis gauges each study's impact.The remove-one analysis runs the data with all studies except the first, and then all studies except the second, and so on with the resulting data and forest plot depicting the impact of each study.We ran the CMA cumulative analysis program by study publication year.The cumulative analysis run by year allowed us to determine the consistency and thus robustness of the examined relationship over time.Lastly, and in line with the PRISMA guidelines, we examined our results (e.g., 95% confidence and prediction intervals, risk of bias assessments, and differences between moderator groups) to judge certainty related to our hypothesized motivational climate and hedonic well-being relationships.

Study Selection, Characteristics, and Quality
The 82 included studies are found in Table 4.The 82 studies resulted in 457 extracted correlations entered into the CMA program (see the Supplemental File for all entered correlations).The study publication years ranged from 1993 to 2023, with studies from the following decades: 1990s (n = 3), 2000s (n = 19), 2010s (n = 46), and 2020s (n = 14).The studies included 26,378 (M = 321.68,SD = 273.12,range 27 to 1430) participants from Europe (Croatia, Finland, Germany, Greece, Ireland, Italy, Norway, Poland, Portugal, Serbia, Spain, Sweden, Turkey, UK), Asia (China, Japan), and North America (Canada, Mexico, USA).Participants included children, adolescents, and adults, M age = 25.20 (SD = 3.65).Of samples reporting male and female composition, 37% of the samples were greater than 50% female participants (M = 46.20%females).Studies reported on both individual sports athletes (e.g., tennis, swimming, gymnastics) and team sports athletes (e.g., handball, soccer, and volleyball).Concerning the quality score (see Figure 2 for details), the mean summary score was 0.92 (SD = 0.05) for the rated samples.Though cross-sectional studies are of low quality compared to experimental or quasi-experimental designs, for our purpose of metaanalyzing correlate relationships, the studies were of sufficient quality.The most neglected category was #10 as few studies reported correcting for alphas.

Task Climate Individual Study Data, Synthesis of Results, and Risk of Bias across Studies
Table 5 contains all the summary data for the task climate analyses.The individual study data with corresponding forest plots and the trim and fill plots for the task climate analyses are located in Figures 3-8.For both positive affect and satisfaction, the random effect sizes were medium in magnitude.The 95% confidence intervals remained in the same effect size interpretation range.Of note, the task climate positive affect true prediction interval did not cross zero.Heterogeneity was present though the bias statistics suggested the relationships to be free or mostly free from bias (see funnel plots in Figure 4 for positive affect and Figure 6 for satisfaction).

Task Climate Individual Study Data, Synthesis of Results, and Risk of Bias across Studies
Table 5 contains all the summary data for the task climate analyses.The individual study data with corresponding forest plots and the trim and fill plots for the task climate analyses are located in Figures 3-8.For both positive affect and satisfaction, the random effect sizes were medium in magnitude.The 95% confidence intervals remained in the same effect size interpretation range.Of note, the task climate positive affect true prediction interval did not cross zero.Heterogeneity was present though the bias statistics suggested the relationships to be free or mostly free from bias (see funnel plots in Figure 4 for positive affect and Figure 6 for satisfaction).Abbreviations: PA = positive affect constructs, NA = negative affect constructs, k = number of samples, CI = confidence interval, PI = prediction interval, Q = Q total between statistics, τ 2 = tau-squared, I 2 = ratio of excess dispersion to total dispersion, FS = fail-safe number.Superscript: A = Abrahamsen and Kristiansen [87] data point removed.The task climate and negative affect relationship unlike the positive affect/mood and satisfaction relationships was small in magnitude with the 95% confidence intervals crossing 0. As with the positive affect/mood and satisfaction analyses, heterogeneity was present.The trim and fill analysis suggested that bias was present.As seen in the individual study data (see Figure 7) and corresponding funnel plot (see Figure 8), the Abrahamsen and Kristiansen [87] data point appears as an obvious deviation from the other studies.Thus, we examined the task climate and negative affect relationship without Abrahamsen and Kristiansen.However, these analyses resulted in little to no change in the effect size statistics (refer back to Table 5).The task climate and negative affect relationship unlike the positive affect/mood and satisfaction relationships was small in magnitude with the 95% confidence intervals crossing 0. As with the positive affect/mood and satisfaction analyses, heterogeneity was present.The trim and fill analysis suggested that bias was present.As seen in the individual study data (see Figure 7) and corresponding funnel plot (see Figure 8), the Abrahamsen and Kristiansen [87] data point appears as an obvious deviation from the other studies.Thus, we examined the task climate and negative affect relationship without Abrahamsen and Kristiansen.However, these analyses resulted in little to no change in the effect size statistics (refer back to Table 5).Figure references [24,26,41,42,44,46,[48][49][50]56,57,60,61,63,64,[69][70][71]78,80,81,84,86,87,93,94,96,98,[101][102][103]107,109,110,112,116,117,120].

Ego Climate Individual Study Data, Synthesis of Results, and Risk of Bias across Studies
Individual study data with corresponding forest plots for the ego climate analyses are located in Figures 9-14.Table 6 contains all the summary data for the ego climate analyses.For both positive affect and satisfaction, the random effect sizes were small in magnitude.For both sets of measures, the 95% confidence intervals remained just inside 0. However, the true prediction intervals crossed zero.Heterogeneity was present for both sets of measures.For positive affect, Orwin's n was 0 as this analysis utilizes the fixedeffect r (−0.08).The trim and fill analysis suggested that the ego climate and positive affect relationship needed correction, but the overall relationship changed only from −0.11 to −0.09 (see Figure 10).The ego climate and satisfaction relationship appeared to be influenced by Bekiari and Syrmpas [88] (see individual study data in Figures 11 and 12 for the funnel plot) in that the publication bias statistic adjusted from −0.18 to −0.30.Removal of Bekiari and Syrmpas resulted in no trim and fill adjustment and a resultant random effects correlation of −0.11.Abbreviations: PA = positive affect constructs, NA = negative affect constructs, k = number of samples, CI = confidence interval, PI = prediction interval, Q = Q total between statistics, τ 2 = tau-squared, I 2 = ratio of excess dispersion to total dispersion, FS = fail-safe number.Superscript: A = Bekiari and Syrmpas [88] data point removed.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Ego Climate Individual Study Data, Synthesis of Results, and Risk of Bias across Studies
Individual study data with corresponding forest plots for the ego climate analyses are located in Figures 9-14.Table 6 contains all the summary data for the ego climate analyses.For both positive affect and satisfaction, the random effect sizes were small in magnitude.For both sets of measures, the 95% confidence intervals remained just inside 0. However, the true prediction intervals crossed zero.Heterogeneity was present for both sets of measures.For positive affect, Orwin's n was 0 as this analysis utilizes the fixed-effect r (−0.08).The trim and fill analysis suggested that the ego climate and positive affect relationship needed correction, but the overall relationship changed only from −0.11 to −0.09 (see Figure 10).The ego climate and satisfaction relationship appeared to be influenced by Bekiari and Syrmpas [88] (see individual study data in Figures 11 and 12 for the funnel plot) in that the publication bias statistic adjusted from −0.18 to −0.30.Removal of Bekiari and Syrmpas resulted in no trim and fill adjustment and a resultant random effects correlation of −0.11.Abbreviations: PA = positive affect constructs, NA = negative affect constructs, k = number of samples, CI = confidence interval, PI = prediction interval, Q = Q total between statistics, τ 2 = tau-squared, I 2 = ratio of excess dispersion to total dispersion, FS = fail-safe number.Superscript: A = Bekiari and Syrmpas [88] data point removed.The ego climate and negative affect relationship like the positive affect and satisfaction relationships was small in magnitude.Unlike the other relationships, the ego climate and negative affect 95% confidence interval did not cross zero and remained small in magnitude while the true prediction interval crossed zero.As with the positive affect and satisfaction analyses, heterogeneity was present.The trim and fill analysis suggested eight missing samples though the effect size changed only to 0.15 from 0.19.Last, the bias statistics suggested that this relationship requires many studies for the relationship to change, confirming a fairly robust relationship.The ego climate and negative affect relationship like the positive affect and satisfaction relationships was small in magnitude.Unlike the other relationships, the ego climate and negative affect 95% confidence interval did not cross zero and remained small in magnitude while the true prediction interval crossed zero.As with the positive affect and satisfaction analyses, heterogeneity was present.The trim and fill analysis suggested eight missing samples though the effect size changed only to 0.15 from 0.19.Last, the bias statistics suggested that this relationship requires many studies for the relationship to change, confirming a fairly robust relationship.Figure references [24,41,42,44,46,48,49,51,53,56,57,60,61,63,64,[69][70][71]78,80,81,84,86,93,94,96,98,[101][102][103]107,109,112,116,117,120].

Study name
Statistics for each study Correlation and 95% CI

Additional Sensitivity Analyses
The remove-one study analysis forest plots are located in the supplemental file.The remove-one study analysis gauges the impact of each included study.For the task climate remove-one study analyses, the individual point estimates for each correlate category appeared to be consistent as the range of point estimates varied little even with the Abrahamsen and Kristiansen [87] data.At best, the Bekiari and Syrmpas [88] data slightly im-

Discussion
Researchers continue to study motivation from the AGT from the original dichotomous perspective in the sport literature.The present study was a systematic review with a meta-analysis of the published literature of the task or mastery and ego or performance motivational climate and three constructs within subjective or hedonic well-being.With minimal study overlap with the Harwood and colleagues' meta-analysis (20 of 82), a focus within sport, and the examination of potential moderators, we believe that this review advances the AGT-based motivational climate literature.

Summary of Findings
Concerning the task climate results, our findings place a high degree of certainty that this climate is positively related to positive affect and satisfaction measures and negatively related to negative affect.For positive affect and satisfaction, both relationships resulted in medium meaningfulness correlations, whereas the negative affect correlation was small in effect size interpretation.Of interest is the task climate and positive affect effect size in this review being less than that of Harwood and colleagues [9].In fact, the 95% confidence intervals do not overlap.Though with no sport participants, Braithwaite et al. [25] quantified task climate interventions within physical education classes.The resultant effect size for enjoyment was small.With the present data, the CMA (version 4) program provides a true effect prediction interval, which is interpreted as the range of plausible values that can include the true effect.The true predicted interval ranged from a minimal effect to a large effect.With all the information and past meta-analyses, the task climate as hypothesized since inception has no downside with positive affect and mood measures.The same conclusion can be drawn with self-rated satisfaction measures and task climate perceptions.Our satisfaction data seem to be unique to the literature and thus of great importance.Why athlete level moderated the task climate and positive affect/satisfaction relationship is unknown and open to speculation.Further down in our discussion, we propose more elite athlete research as a future direction.
As with Harwood et al. [9], the task climate and negative affect relationship was small.The ego climate relationships were all small in meaningfulness interpretation.The ego climate relationships provided a confirmation of the small relationships with positive and negative affect that Harwood and colleagues [9] reported.With negative affect, the true prediction interval provides certainty that the effect size falls between no relationship to a medium relationship.The two other quantified relationships, positive affect and satisfaction, had wider true prediction interval ranges from positive to negative values, thus casting doubt on the true effect size.As with the task climate and positive affect/satisfaction measures, the athlete level moderated the ego climate and positive affect/satisfaction relationship.This moderation, even with the one outlier removed, resulted in the elite athlete category, comprised of elite and advanced/elite athletes, resulting in a negligible correlation.The relationship with the sub-elite samples was small and negative, but even a small negative relationship to desired states lessens the potential joy of sport participation at any level.

Strengths, Limitations, Future Directions, and Applications
The strengths of our meta-analysis were the inclusion of 62 articles beyond the Harwood et al. [9] quantitative review with an extensive search strategy, following the PRISMA guidelines, the inclusion of satisfaction as a correlate, reporting the true prediction interval statistic provided by CMA version 4, and the examination of longstanding proposed AGT moderators.Our search resulted in a number of positive affect, negative affect, and satisfaction measures.This is also a strength.Example measures included the PANAS [121], the vigor and enthusiasm subscales from the Athlete Engagement Questionnaire [122], the pleasant and unpleasant subscales and related moods from emotional state questionnaires [123], the Sport Anxiety Scales [51], the Sport Satisfaction Scale from Duda and Nicholls joint education and sport publication [20], and the enjoyment subscale from Scan-lan's original and updated Sport Commitment Questionnaire [124,125].We limited each sample to only one effect size per task and ego climate analysis for each of the hedonic categories, which is another strength.For example, within the CMA program, the reporting of the three subscales by Smith and colleagues [51] of the Sport Anxiety Scale-2 was merged as were studies reporting multiple positive affect or negative affect correlations.Despite our strengths while following the structured PRISMA approach [27] to formulate and conduct a systematic review with a meta-analysis, limitations existed, stemming from the process and information provided in the included articles.
The first limitation is the number of missed studies, as we used only English in our search.The number of motivational climate studies in non-English languages (e.g., journals) is lacking in this review.Larger research teams from different countries or at least a research team member with multiple-language expertise is required to remedy this limitation.For instance, Lochbaum and colleagues' [31] meta-analysis on the 3 × 2 achievement goal framework included a search in the Turkish language in an attempt to minimize the language bias [126].Biddle and colleagues' [127] systematic review on martial arts, combat sports, and mental health is an example as they searched in six different languages.The non-English studies included in our review were retrieved as the title, abstract, or keywords were written in English and supplied with the published manuscript.To extract the relevant details of methodology and results, we used Google Translate.For a few study quality ratings, we were unable to reach confidence in the provided translation.Though providing a unique finding, our coding of athlete level is another potential limitation.We applied the coding system found in the Lochbaum et al. [33] athletic identity meta-analysis based on Kyllo and Landers [128] and Swann and colleagues' [129] coding systems.The limitation stems from within sample level study participant sections, as coding depends upon the author-provided descriptions.Research with sport samples following Swann and colleagues' system will move sports science research forward.A last limitation stems again from the studies themselves as little if any random sampling or any such sampling other than convenience.However, the standard, convenience sampling could impact the data in ways unknown, as there is not enough of such studies for a comparison.
In terms of future directions, a further examination of elite sport and hedonic wellbeing is important, even though access to elite sport is limited.For instance, in the dichotomous AGT research, though not a firm estimate of all the AGT literature, Lochbaum et al. [6] reported that nearly half of the 260 included studies were from youth sport and approximately 19% with elite sport participants.Whether elite sport participants are less influenced by or interested in the motivational climate is unknown.Our results only indicated that the relationships were dampened.The study of potential moderators or mediators is needed to best understand the dampened relationships.For example, Ntoumanis and Biddle [42] reported self-confidence mediated the ego climate to ego orientation to state anxiety relationship.In addition to variables such as self-confidence, the athlete's relationship with their coach is a variable needing more attention as a mediator, or moderators such as the athlete's standing within the team, playing status, or length of time with the team.Another future direction concerns athletes competing at the Masters level of sport.A surprise is that not one of the mean ages in any of our studies exceeded 26, let alone approaching the age to enter for Masters athletics of 35.Hence, research with older athletes is an undeveloped area for future motivational climate and hedonic well-being research.Last, researching hedonic or eudaimonic well-being with intentionality is a future research direction.Eudaimonic well-being unlike hedonic well-being is more disputed in terms of the key concepts needed.Readers should consider Trainor and Bundon's [14] well-being commentary to gain an understanding of eudaimonic well-being frameworks.

Conclusions
In conclusion, though limitations exist with correlational data, the knowledge gained in this review is of value.First, it is evident that AGT from the original dichotomous perspective is still popular and a stronghold in the sport environment.From the list of The levels of competition included elite (n = 6), advanced/elite (n = 8), advanced (n = 15), intermediate (n = 20), intermediate/advanced (n = 5), mixed (n = 13), youth/intermediate (n = 4), and youth (n = 11) samples.Researchers utilized a variety of motivational climate scales with the most frequently used scales being the PMCSQ-2 (n = 42), the PMCSQ (n = 21), and the MCSYS (n = 12).It is notable that of the included 82 studies, only 20 overlapped with the Harwood et al. [9] quantitative review.

Figure 4 .
Figure 4. Task climate and positive affect random effects plot trimmed and filled.The open circles are the data points.The clear rhombus is the mean effect size.

Figure 6 .
Figure 6.Task climate and satisfaction random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 6 .
Figure 6.Task climate and satisfaction random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 8 .
Figure 8. Task climate and negative affect random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 8 .
Figure 8. Task climate and negative affect random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 10 .
Figure 10.Ego climate and positive affect random effects plot trimmed and filled.The open circles are the data points.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 12 .
Figure 12.Ego climate and satisfaction random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 12 .
Figure 12.Ego climate and satisfaction random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 14 .
Figure 14.Ego climate and negative affect random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Figure 14 .
Figure 14.Ego climate and negative affect random effects plot trimmed and filled.The open circles are the data points, and the filled circles are the result of the trim and fill analysis.The clear rhombus is the mean effect size, and the filled rhombus is the trim and filled mean effect size.

Table 1 .
The dominant sport motivational climate measures.

Table 2 .
Athlete-level categories and specifics used for classification.

Table 3 .
Individual study bias questions and rating explanations summed to a study quality score.

Table 5 .
Task/mastery climate and hedonic well-being results.

Table 5 .
Task/mastery climate and hedonic well-being results.

Table 6 .
Ego/performance climate and hedonic well-being results.

Table 6 .
Ego/performance climate and hedonic well-being results.