Abstract
earchers have extensively examined and reviewed the relationship of the pro le of mood states (POMS) with sport performance since the 1970s. Two decades have passed since the last POMS quantitative review. Our overall objective was to quantify the POMS research with competitive athletes with a prospective measured POMS and a sport performance outcome in the published literature. Additionally, we tested potential moderators of the mental health model (i.e., sport duration, type, and skill) with meta-analytic techniques while considering potential risk bias across study sources. Based on a systematic review, the articles were found using EBSCO and comparing these articles with extensive past POMS in sport and exercise bibliographies. Search terms included pro le of mood states (POMS) or iceberg pro le or the mental health model with sport and performance or sports performance. For selection, articles must have reported data on competitive athletes, an athletic performance outcome, and a valid form of the POMS
EBSCO and comparing these articles with extensive past POMS in sport and exercise bibliographies. Search terms included pro le of mood states (POMS) or iceberg pro le or the mental health model with sport and performance or sports performance. For selection, articles must have reported data on competitive athletes, an athletic performance outcome, and a valid form of the POMS measured prospectively. After screening more than 600 articles for inclusion, 25 articles provided suf cient data for effect size calculations. The included articles spanned from 1975 to 2011, with 1497 unique participants. Hedges' g values were generally small for the six POMS scales: tension ( 0.21), depression ( 0.43), anger ( 0.08), vigor (0.38), fatigue ( 0.13), and confusion ( 0.41). However, the total mood disturbance (TMD) score effect size was medium in magnitude at 0.53. When corrected for potential publication bias, the effect size values increased in magnitude for tension ( 0.47), depression ( 0.64), vigor (0.44), fatigue ( 0.34), and TMD ( 0.84). Moderator analyses for Terry's (1995) propositions and for risk of bias across studies, statistically, resulted in few differences based on conventional statistical signi cance (p< 0.05). Measured before performance, most of the POMS scales and TMD are reliable predictors of sport performance in competitive athletes across a wide variety of sports and athletic performance outcomes. Morgan's (1980, 1985) mental health model or iceberg pro le minus anger is still a viable method for understanding and improving athletic performances. Keywords: mental health model; iceberg pro le; vigor; depression; total mood disturbance; quantita- tive review 1. Introduction Given that sport has performance statistics and winners and losers, an understanding of how to affect performance and thus the outcome of sporting events is valued and re- searched across all sport science areas. In sport psychology research, understanding athletic performance predictors is a main goal and there is much literature on the subject. For instance, researchers have quanti ed many sport psychology interventions and constructs (see Lochbaum [1]) relating (to) athletic performance such as goal setting [2], achievement goals [3], mental practice [4], and team cohesion [5]. As with
all sport science areas. In sport psychology research, understanding athletic performance predictors is a main goal and there is much literature on the subject. For instance, researchers have quanti ed many sport psychology interventions and constructs (see Lochbaum [1]) relating (to) athletic performance such as goal setting [2], achievement goals [3], mental practice [4], and team cohesion [5]. As with the above-mentioned sport Eur. J. Investig. Health Psychol. Educ.2021,11, 5070.
Eur. J. Investig. Health Psychol. Educ.2021,11 51 psychology and performance topics, mood states as measured by the pro le of mood states (POMS) questionnaire [6] is a much-researched topic within sport psychology under Mor- gan's [7,8] mental health model or often termed iceberg pro le. Indeed, two meta-analyses have been previously published [9,10] as well as a number of systematic reviews [1114]. Given the importance of predicting performance, the longevity of Morgan's mental health model with the POMS in sport psychology research, and the passing of at least 20 years since the last reviews [1013], we extensively searched the peer-reviewed pub- lished POMS and sport performance research. This review only included POMS studies using McNair and colleagues' [6] measure. We then meta-analyzed data when available, examined potential sources of bias, and used mixed-effects analysis techniques to test a number of hypotheses with the hopes of solidifying and furthering POMS research in the competitive sport environment regarding predicting performance. 1.1. Morgan's Mental Health Model Morgan and his colleagues appeared to begin using the POMS shortly after its publi- cation [6] to predict US Olympic wrestlers' trial success [15,16] and rowing [17]. Within the next decade, Morgan [7], in his Psychology Today popular press article, termed the unique differences found between successful and unsuccessful sport performers based on the POMS subscales, with successful performers lower in tension, depression, anger, fatigue, confusion, and higher in vigor compared to less successful performers, the iceberg pro le. Researchers enthusiastically took to the POMS in competitive sport contexts to explain sport performances in both individual sports such as distance running [18], speed skating [19] and team sports such as football [20] and netball [21]. It was in this time period, in an edited text titled Limits of Human Performance, that Morgan [8] presented his mental health model, stating positive mental health enhances the likelihood of success in sport, whereas psychopathology is associated with a greater incidence of failure [8] (p. 79). By the 1990s, enough research via academic conference presentations and published articles had accumulated for both qualitative [11,12] and quantitative [9] reviews. In the next
of Human Performance, that Morgan [8] presented his mental health model, stating positive mental health enhances the likelihood of success in sport, whereas psychopathology is associated with a greater incidence of failure [8] (p. 79). By the 1990s, enough research via academic conference presentations and published articles had accumulated for both qualitative [11,12] and quantitative [9] reviews. In the next decade, two qualitative [13,14] and one quantitative [10] reviews kicked off the new decade. A number of inconsistencies pervaded the reviews. For instance, since the rst review [11,12], the notion of the POMS accurately or signi cantly discriminating among lev- els of successful athletes or compared to non-athletes has been viewed as unsubstantiated as well as entirely unreasonable [12] (p. 310). Renger [11] pointed his criticism to Morgan's [7] overly simpli ed popular press article as well as using the POMS as a personality measure. Terry [12] stressed, in a more positive light, that research to best understand the mood and performance relationship both between and within athletes is needed. He highlighted a number of areas, such as early problem identi cation, load monitoring, and the like. Further, in 1995, Rowley and colleagues [9] (most notably his mentor Daniel Landers) published the rst meta-analysis on the POMS. Though casting a less than supportive conclusion ... the utility of the POMS in predicting athletic success is questionable [9] (p. 185), the overall effect size, representing the POMS total mood score, though small (ES = 0.15), was statistically different from zero, indicating that more successful athletes' mood pro les conformed to Morgan's iceberg pro le. By the time of the next round of POMS reviews, research with the POMS ourished. For instance, LeUnes and Burger [13] identi ed 258 published articles using the POMS in sport and exercise contexts based on LeUnes' updated POMS bibliographies [2224]. LeUnes and Burger [13] concluded with improved research techniques and POMS mea- surement (e.g., validated measure for youth) that mood pro ling will be important com- ponents of POMS research and application into the millennium [13] (p. 14). In contrast, Prapavessis [14], in his review, pushed
the POMS in sport and exercise contexts based on LeUnes' updated POMS bibliographies [2224]. LeUnes and Burger [13] concluded with improved research techniques and POMS mea- surement (e.g., validated measure for youth) that mood pro ling will be important com- ponents of POMS research and application into the millennium [13] (p. 14). In contrast, Prapavessis [14], in his review, pushed for Hanin's Individual Zone of Optimal Function model (see Ruiz, Raglin, and Hanin [25] for a historical review of Hanin's work) over Morgan's mental health model to best understand sport performance.
Eur. J. Investig. Health Psychol. Educ.2021,11 52 Beedie and colleagues' [10] meta-analysis provided more de nitive answers by sepa- rating level of achievement studies and performance outcome studies. They coded and attempted to analyze their data based on a number of Terry's [12] mood and performance propositions such as sport type and duration, operational de nition of performance success, timing of mood assessment relating (to) performance, and POMS response set. Addition- ally, unlike Rowley and colleagues' [9] work, Beedie et al. [10] reported data for all POMS subscales when available. For the achievement studies, when averaged together appropri- ately (all coded to positive mental health), the overall effect size (0.10) was nearly identical to that of Rowley's. Separated, the vigor effect size (0.20) was the largest value supporting Morgan's iceberg pro le. The change in tone concerning support of Morgan's mental health model or iceberg pro le stemmed from the mood and performance results. Across 17 samples, the effect size values in order of magnitude were 0.47 (vigor), 0.40 (confusion), 0.34 (depression), 0.27 (anger), 0.25 (tension), and 0.13 (fatigue). Citing small samples across their potential moderator variables based on Terry's [12] propositions, Beedie and colleagues [10] reported effect size representing total mood values without statistical testing for differences. Those values, regardless of the proposition category, ranged from 0.27 to 0.39, thus all still being categorized as small [26]. Though subscale effect size values are found in their review (see Table 2, p. 62), the authors did not expound upon them. Even with limited samples and statistical tests, Beedie and colleagues [10] concluded in their abstract based concerning the POMS predicting performance outcomes, the POMS has utility [10] (p. 49) though, in their discussion, they quali ed this utility as moderate at best [10] (p. 63). Certainly, one could argue the subscale effect size values approaching moderate in magnitude are of value concerning bettering an athlete's performance. 1.2. Objectives In summary, improving athletic success in competitive sport is valued. Thus, based on our examination of the POMS reviews and the potential for more advanced statistical testing, our overall objective was to
at best [10] (p. 63). Certainly, one could argue the subscale effect size values approaching moderate in magnitude are of value concerning bettering an athlete's performance. 1.2. Objectives In summary, improving athletic success in competitive sport is valued. Thus, based on our examination of the POMS reviews and the potential for more advanced statistical testing, our overall objective was to update the research knowledge since Beedie and colleagues' [10] meta-analysis concerning Morgan's mental health model [7,8] as measured by the POMS related only to predicting athletic success. To achieve our overall objective, we reviewed all found published studies that specially assessed the POMS before athletic performance in competitive settings with competitive athletes. By doing so, we tested whether the mental health model or iceberg pro le is still characteristic of successful performances. Then we investigated a number of Terry's [12] proposed moderators that received mixed support in the Beedie and colleagues' [10] meta-analysis. We placed Terry's proposition concerning performance references (i.e., objective or self-referenced) as a potential risk of across-study bias moderators. 1.3. Research Questions Speci cally, we tested the following research questions. In Morgan's [7,8] mental health model, especially higher levels of vigor and lower levels of tension, depression, anger, fatigue, and confusion, characterize successful athletic performances and are invariant to potential across study risks of bias. Based on Terry's [12] proposed moderators, we tested whether the POMS subscales explain (i.e., larger in effect size values) successful athletic performance more in (a) short-duration sportsde ned as less than 10 min, when compared to longer-duration sports; (b) closed-skill sportsde ned as mostly self-paced, little to the interaction between or among competitors, and few external performance in uences when compared to open-skill sports, which are de ned as sports with higher unpredictable competitor interactions or potential external in uences; and (c) individual sportsde ned as sports requiring no teammate cooperation, compared to team sports, which are de ned as sports requiring teammate cooperation, whether open- or closed-skill sports. Across all three moderators, the notion is that mood, as measured by the POMS, before the sporting event should be more accurate of
unpredictable competitor interactions or potential external in uences; and (c) individual sportsde ned as sports requiring no teammate cooperation, compared to team sports, which are de ned as sports requiring teammate cooperation, whether open- or closed-skill sports. Across all three moderators, the notion is that mood, as measured by the POMS, before the sporting event should be more accurate of mood within the event because
Eur. J. Investig. Health Psychol. Educ.2021,11 53 the event is shorter in time or has fewer outside in uences or no teammates to impact competition mood and thus potentially performance. 2. Materials and Methods This meta-analysis reported on each PRISMA statement item [27]. Thus, our reporting is transparent with the goal of being perceived as comprehensive. 2.1. Protocol and Registration Researchers have tested Morgan's mental health model or iceberg pro le since the 1970s, and Terry's proposed moderators are of written record. Thus, we did not register our protocol in a database. We speci ed our search strategy, inclusion criteria, data extraction, and data analyses in advance of writing our manuscript. All details of our work, if in question, are available from the lead author. 2.2. Eligibility Criteria Articles retained for extensive examination met the following inclusion criteria: papers with (a) any methodological design such as mean group, between or within differences or correlates with performance; (b) a publication date after that of the POMS up to 1 January 2020; (c) original data published in peer-reviewed journals; (d) competitive athletic par- ticipants; and (e) a valid full- or short-form POMS questionnaire; (f) the POMS assessed before sport performance; and (g) a measure of sport performance. There was no language of publication restriction. To align with our review objectives, we gave much consideration to study participants and performance outcomes. We based the competitive athletic par- ticipants criteria on the speci cs found in the sample descriptions. We did not consider participants described as recreational or intramural or volunteers at a rehabilitation clinic or exercisers. Speci cally, we de ned sport performance as immediate outcomes such as making an Olympic team, winning and losing sport performance statistics such as baseball batting average, and future athletic success such as becoming a professional athlete. We excluded performance outcomes associated with athletics such as shuttle run performance or vertical jump tests as they in and of themselves are not the outcome of a sporting event. Articles included in the meta-analysis portion of this review met the above inclusion re- quirements and provided necessary data for
and future athletic success such as becoming a professional athlete. We excluded performance outcomes associated with athletics such as shuttle run performance or vertical jump tests as they in and of themselves are not the outcome of a sporting event. Articles included in the meta-analysis portion of this review met the above inclusion re- quirements and provided necessary data for effect size computation. The rst and third authors rigorously checked eligibility and nal inclusion assessments. 2.3. Information Sources We systematically identi ed studies by searching electronic databases, references from published POMS bibliographies, and references from two published meta-analyses. All authors conducted their electronic database search in EBSCO with the following individ- ual databases selected: SPORTDiscus, PsycINFO, and ERIC. The main extensive search concluded in July 2019. The second author then extensively examined the search and expanded the search into January 2020. 2.4. Search Protocol All authors used the following search terms: pro le of mood states or POMS or iceberg pro le or mental health model with sport and performance or sports performance. In EBSCO, we used the advanced search option that provided three separate boxes for search terms such as box 1 (pro le of mood states), box 2 (sport), and box 3 (performance). At each search stage, we restricted EBSCO to a one-year period (e.g., 1974). Once a one-year period, each author restarted with the next year (e.g., 1975). Here are the details of our search strategy: 1. 2. 3. 4.
Eur. J. Investig. Health Psychol. Educ.2021,11 54 5. 6. 7. 8. 2.5. Study Selection As detailed in the PRISMA ow chart [28] (Figure) and the details of inclusion criteria, the study selection process was rigorous. Three of the authors ( rst, third, and fourth) engaged independently in the majority of the study selection process. The rst and third author selected studies for possible inclusion while the second author engaged in a complete review of all students pulled. Through the process, we settled disagreements by consensus while examining the study inclusion criteria.Eur. J. Investig. Health Psychol. Educ. 2021, 11, x FOR PEER REVIEW 5 of 22 1. Profile of mood states, sport, performance; 2. POMS, sport, performance; 3. Iceberg profile, sport, performance; 4. Mental health model, sport, performance; 5. Profile of mood states, sport, sports performance; 6. POMS, sport, sports performance; 7. Iceberg profile, sport, sports performance; 8. Mental health model, sport, sports performance. 2.5. Study Selection As detailed in the PRISMA flow chart [28] (Figure 1) and the details of inclusion cri- teria, the study selection process was rigorous. Three of the authors (first, third, and fourth) engaged independently in the majority of the study selection process. The first and third author selected studies for possible inclusion while the second author engaged in a complete review of all students pulled. Through the process, we settled disagreements by consensus while examining the study inclusion criteria. Figure 1. PRISMA flow chart for the identification of the eventual included studies. Figure 1.PRISMA ow chart for the identi cation of the eventual included studies.
Description
This study quantifies the POMS research with competitive athletes and tests moderators of the mental health model.