Skip to main content
Download PDF

The structural validity and latent profile characteristics of the Abbreviated Profile of Mood States among Chinese athletes

BMC Psychiatry volume 24, Article number: 636 (2024) Cite this article

Abstract

Background

This study examines the structural validity of the Chinese version of the Abbreviated Profile of Mood States (POMS) among Chinese athletes and analyzes potential profiles to provide evidence for its effective use and recommendations for its application.

Methods

A total of 340 Chinese athletes completed the Chinese version of the Abbreviated POMS. Initially, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were conducted to identify and verify the extractable dimensions of the Abbreviated POMS. Subsequently, the fit of the six-factor and seven-factor models of POMS was tested directly based on their theoretical structures. Finally, latent profile analysis was used to examine profiles based on the four-factor model derived from the factor analysis, six-factor model, and seven-factor model.

Results

The Abbreviated POMS was refined to a four-factor model, consisting of 27 items across four factors: positive mood, anger, fatigue, and confusion. The hypothesized six-factor and seven-factor models did not demonstrate satisfactory fit, suggesting that the seven dimensions function better as independent subscales. Iceberg and inverse iceberg profiles were observed across the four-factor model, six-factor model, and seven-factor model.

Conclusion

The Abbreviated POMS does not support its initial hypothesized structure among Chinese athletes. Caution is advised when using the Abbreviated POMS with athletes; it is recommended to use the four-factor model or evaluate each emotion as an independent subscale. The iceberg and inverse iceberg profiles can be used to categorize athletes’ emotional characteristics.

Peer Review reports

Background

The Profile of Mood States (POMS), introduced in 1971, serves as a tool for measuring immediate or transient emotional states and has primarily been utilized to quantify progress in counseling and psychotherapy [1]. Since its introduction to the field of sports in 1975, the POMS has been extensively employed in numerous studies, becoming a prominent instrument within sports psychology [2,3,4]. It has even been selected as an indicator for evaluating overtraining syndrome [5]. William Morgan’s early work on the “iceberg profile,” derived from the pre-competition mood states of Olympic athletes in rowing, endurance running, and wrestling, became a criterion for identifying elite athletes and assessing athletes’ psychological health over time [4, 6]. Responding to the research needs in sports, Grove and colleagues developed the Abbreviated Profile of Mood States (Abbreviated POMS), consisting of 40 items organized into seven dimensions [7]. Shortly after its introduction, this scale was adopted domestically, becoming one of the primary tools for measuring athletes’ moods in China [8].

The introduction of the Profile of Mood States (POMS) has provided an effective quantitative tool for predicting athletes’ mental health and competitive performance, as well as for implementing proactive interventions. However, researchers have gradually raised concerns about the effectiveness of POMS in evaluating athletic performance. For instance, some studies have found that POMS fails to distinguish changes in certain training states [9]. Others argue that the predictive power of POMS for athletic achievement is relatively low and requires further verification [10]. Additionally, a meta-analysis suggested that while POMS can predict outcomes, it is not suitable for predicting performance [11]. These issues may partly stem from the complex relationship between athletic performance and outcomes [11], but the internal issues within POMS itself should not be overlooked. From early explorations and validations to the increasing number of studies revealing varying structures, the structural validity of POMS has consistently been both confirmed and challenged over time [12, 13]. Although emotions indeed influence athletic performance [14], if the structural validity of POMS is poor or does not support the original dimensions, effective prediction becomes difficult.

Since its development, POMS has sparked extensive discussion regarding its dimensions [1, 15]. It is important to note that POMS was initially developed as a tool for clinical interventions, with its dimensions proposed a priori based on the developers’ clinical needs rather than being empirically validated through data [1]. This subjectively driven dimension setting may align with practical needs but does not necessarily reflect the actual structure of the scale. This issue is particularly evident in the development of the Abbreviated POMS. The “Esteem” dimension in the Abbreviated POMS was directly added to the original POMS, with some items modified, yet its revised structure was never empirically tested [7]. Moreover, due to the large number of items in the original POMS, it has undergone several simplifications and revisions to meet practical demands, with further modifications made based on varying needs [12, 13, 16]. It is crucial to recognize that each modification of the items may affect the structure of POMS [1]. Therefore, the possibility of structural changes in POMS, especially in specific versions or populations, must be carefully considered.

This possibility exists and has been confirmed by some studies. Early research did indeed support the structural validity of both the original and abbreviated versions of the POMS based on its six dimensions [16,17,18,19]. However, more recent studies have increasingly found inconsistencies in the structure following simplifications and revisions for specific populations. For instance, in the clinical domain, a study of patients with chronic pain symptoms found that the original scale’s structure did not fit well, with negative emotions forming a distinctive structure among patients [13]. Similarly, a study conducted among pregnant women derived four dimensions from a reduced 27-item POMS, with two negative emotion dimensions forming a second-order factor [20]. In the general population, researchers in Portugal revised the POMS-36 among students, resulting in a version with three dimensions and 12 items [21]. Revision among Australian students revealed three additional state dimensions beyond the six primary dimensions [15]. In a version revised among the general population in Germany, 35 items formed four dimensions, while another version comprised 16 items [22]. In similar East Asian cultures, a revision among elderly Koreans yielded three dimensions, while among elderly individuals in Taiwan, only one dimension was found [23, 24]. Furthermore, a study involving athletes and university students, based on the simplification of the original POMS to the 30-item EPOMS, extracted only five dimensions, retaining a six-dimensional structure after introducing one weaker dimension [12].

The Abbreviated POMS, which is widely used in China, similarly may face structural validity issues. During the revision process of both the Abbreviated POMS and the more prevalent POMS-SF into Chinese, their structures were not rigorously examined [8, 25]. Consequently, our understanding of their structural validity in the Chinese athletic population is limited. Therefore, despite the widespread application of the Abbreviated POMS in China, it is essential to conduct an in-depth revalidation of its structure. This step is crucial for the effective use of the Abbreviated POMS among Chinese athletes.

Additionally, the profile chart is a classic method for evaluating athletes’ states using POMS. The application of POMS in athletes largely depends on the interpretation of these profiles. However, few studies have analyzed the characteristics of profiles obtainable from the Abbreviated POMS based on its scale structure. Furthermore, an increasing number of studies have identified various profiles beyond the classic iceberg and inverted iceberg profiles in the six dimensions of POMS [6, 26]. Can the structure of the Abbreviated POMS similarly produce these classic profiles? Addressing this question, based on the exploration of scale structure, can further provide effective recommendations for using the Abbreviated POMS.

Specifically, this study will first focus on structural validity by exploring and examining the dimensions of the Abbreviated POMS through factor analysis from a data-driven perspective, thereby establishing a model of the Abbreviated POMS for Chinese athletes. Subsequently, based on a strong theoretical framework from previous research [16, 27], the study will directly test the fit of the seven-factor model of the Abbreviated POMS and the six-factor model of the POMS with the Esteem dimension removed. By analyzing these two approaches, the study aims to provide a comprehensive description of the structural validity of the Abbreviated POMS among Chinese athletes. Subsequently, latent profile analysis will be used to examine profile characteristics based on the structure obtained in this study (four-factor model), the classic six dimensions of POMS (six-factor model), and the seven dimensions of the Abbreviated POMS (seven-factor model). This analysis will test whether the profile-based interpretations are supported by the data and provide more reliable recommendations for the application of the Abbreviated POMS with athletes.

Methods

Participants

A total of 340 Chinese professional athletes participated in this study, comprising 127 males and 213 females, with a mean age of 18.84 years (SD = 4.01). Participants in this study, representing approximately one-third of Shanghai’s professional athletes, were selected on a team basis with permission from coaches and administrators. These athletes were engaged in various sports disciplines such as fencing, modern pentathlon, badminton, and others, undergoing centralized training at training centers. They were all eligible to represent Shanghai in national-level competitions or had been selected for the Chinese national team, frequently utilizing the POMS in training monitoring. Ethical approval for the study was obtained from the Ethics Committee of Shanghai Research Institute of Sports Science (Shanghai Anti-doping Agency) (Ethics Approval Number LLSC 20220005), and all procedures adhere to the ethical principles outlined in the latest version of the Declaration of Helsinki. Written consent was obtained from all participants before the commencement of the test.

For exploratory factor analysis (EFA), 200 athletes were randomly selected, stratified by gender and sport, with 68 males and 132 females, with a mean age of 18.81 years (SD = 4.03). The remaining 140 athletes were used for CFA to examine the structure obtained from EFA, with 59 males and 81 females, with a mean age of 18.87 years (SD = 4.00). Data from all athletes were used in other confirmatory factor analyses and LPA.

Scale

The Chinese version of the Abbreviated POMS was employed, consisting of 40 emotion-related items organized into seven dimensions (tension, anger, fatigue, depression, vigor, confusion, and esteem), widely utilized in research and practical applications among athletes. Participants rated themselves on a scale of 0–4 based on their current emotional state, with higher scores indicating a more prominent expression of the respective emotion. After its introduction to China in the 1990s, the original version underwent Chinese revision, with norms established for Chinese students [8].

Consistent with previous research and customary practices, a one-week timeframe was employed in this study [1]. The POMS offers various timeframes, including the past week, today, the past three minutes, and right now [16]. However, the past week timeframe is generally considered better for capturing typical or sustained emotional responses, yet short enough to demonstrate acute intervention effects [22]. Preliminary evidence also suggests the structural similarity between the one-week and immediate timeframes [28].

Procedures

Participants completed the Abbreviated POMS via paper-based or online questionnaires as part of their routine psychological assessments and pre-competition psychological evaluations over approximately one year to avoid interference from specific periods or tasks. Participants’ responses were used not only for this study but also as feedback data for training monitoring and pre-competition psychological assessments provided to coaches.

Statistical analysis

EFA was conducted using SPSS 27.0, employing principal component analysis and maximum variance rotation, with factors extracted based on eigenvalues greater than 1. Items were retained if their loadings exceeded 0.30. Items showing loadings exceeding 0.30 and differences between loadings across dimensions less than 0.15 were indicative of significant cross-loadings and were excluded from further analysis [29]. CFA of the model obtained from EFA was then conducted using lavaan.

Subsequently, both the seven-factor model of the Abbreviated POMS (using all 40 items) and the six-factor model of the POMS (excluding the Esteem dimension, leaving 35 items) were examined through oblique and orthogonal models. Model fit was evaluated using χ²/df, comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). Following this, the seven dimensions of the Abbreviated POMS were treated as separate scales for CFA to assess their individual fit. Due to the limited number of items per dimension, degrees of freedom were low, and χ² statistics might have been influenced by the number of observed variables in the model. RMSEA was found to be less effective in this context, often indicating poor fit, while CFI and SRMR were more reliable [30,31,32]. Therefore, while reporting the same fit indices as the complete scale, the model fit was primarily judged based on CFI and SRMR results.

After standardizing the data, latent profile analysis was conducted using Mplus 7.4 for the four-factor model (derived from EFA), the six-factor model (excluding Esteem), and the full seven-factor model of the Abbreviated POMS. The Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), adjusted BIC (aBIC), entropy, Lo-Mendell-Rubin likelihood ratio test (LMR-LRT), and Bootstrap Likelihood Ratio Test (BLRT) were employed to compare models with different profile numbers [33]. Lower values of AIC, BIC, and aBIC were considered better indicators of a balance between complexity and parsimony. Entropy values greater than 0.80 indicated classification accuracy exceeding 90%. The LRT and BLRT were used to assess whether models with k profiles were more optimal than those with k-1 profiles; a significant p-value indicated that k profiles were better. Mean scores on the 0–4 scale were computed for each dimension and used in LPA. Profile plots were created using standardized scores of each dimension for comparison and naming purposes [34,35,36,37]. When naming the profiles, features of similar profiles identified in previous studies were primarily referenced [6, 26].

It is important to note that LPA was conducted for both the six-factor and seven-factor models to facilitate comparisons with profiles identified in previous research and applications. This approach aimed to provide new evidence regarding the characteristics of the profiles of Abbreviated POMS among Chinese athletes, rather than proposing new models based on the data, which was done with the four-factor model. Therefore, this study followed the method used in prior research, which incorporated the scores from the POMS subscales directly into the latent profile analysis [6, 38]. The scores for the six or seven dimensions of the POMS, each representing a distinct emotion, were included in the analysis in their entirety, without prior selection based on CFA results. This approach was chosen to prevent any alterations in the dimensions included in the profiles, which could affect comparisons with profiles identified in earlier studies.

Results

Factor analysis of the abbreviated POMS

EFA was conducted on all 40 items. The Kaiser-Meyer-Olkin (KMO) measure was 0.938, and Bartlett’s test of sphericity was significant (χ2 = 7260.724, p < .001), confirming the dataset’s suitability for EFA. Initially, five factors with eigenvalues exceeding 1 were extracted, explaining 70.627% of the total variance. Following the criteria for cross-loadings, 11 items were excluded iteratively. This process resulted in retaining 27 items, achieving a KMO measure of 0.913 and a significant Bartlett’s test (χ2 = 4374.854, p < .001). Subsequently, four factors with eigenvalues greater than 1 were extracted, accounting for 71.146% of the variance. The factor loadings are presented in Table 1.

Table 1 Factor loadings of the abbreviated POMS items obtained from EFA
Full size table

Items were categorized based on their loadings (bold in Table 1) and compared with their original dimensions. Factor 1 predominantly encompassed positive emotions such as vigor and esteem, labeled as Positive Emotion. Factor 2 primarily reflected anger, labeled as Anger. Factor 3 included items related to fatigue, termed Fatigue. Factor 4 comprised items from the tension and confusion dimensions, termed Confusion. None of the items originally classified under depression exhibited sufficiently high loadings and were therefore excluded from the refined model. This refined version is referred to as four-factor model, distinguishing it from other versions of the POMS. The allocation of items from the original seven dimensions to dimension of four-factor model can be found in Table 2.

Table 2 The reassignment of items from their original dimensions to the new dimensions
Full size table

CFA was performed on the four dimensions obtained from EFA. When allowing for correlations among these four dimensions, the model fit indices were as follows: χ2/df = 2.046, CFI = 0.875, TLI = 0.863, RMSEA = 0.086, SRMR = 0.085. These indices indicate a satisfactory fit of the model. However, the correlation between the positive emotion dimension and the three negative emotion dimensions did not significantly differ from zero (p1 − 2 = 0.289, p1 − 3 = 0.472, p1 − 4 = 0.066). Subsequently, the correlation between the positive emotion and the three negative emotion dimensions was constrained to zero in a revised CFA model, resulting in a slightly improved fit: χ2/df = 2.040, CFI = 0.875, TLI = 0.863, RMSEA = 0.086, SRMR = 0.096.

CFA of the theoretical structure of the Abbreviated POMS

When setting the seven dimensions as correlated, the model could not be identified. Re-specifying the seven dimensions as orthogonal yielded the following fit indices: χ2/df = 6.330, CFI = 0.648, TLI = 0.629, RMSEA = 0.125, SRMR = 0.358. The model exhibited poor fit under this condition. After removing the Esteem dimension, which was added to the original six dimensions of the POMS, the six-factor model yielded fit indices of χ²/df = 3.575, CFI = 0.854, TLI = 0.840, RMSEA = 0.087, and SRMR = 0.069. Further setting the six dimensions as orthogonal resulted in fit indices of χ²/df = 6.358, CFI = 0.688, TLI = 0.668, RMSEA = 0.126, and SRMR = 0.374. Subsequently, separate models were constructed for each of the seven dimensions, and the fit indices are presented in Table 3.

Table 3 Fit indices for each dimension of abbreviated POMS
Full size table

LPA of the abbreviated POMS

For the four-factor model, model comparisons across different profile numbers indicated that the two-profile model was optimal (Table 4). While AIC, BIC, and aBIC values continued to decrease, the rate of decrease slowed after the three-profile model, and the entropy at this point was 0.879, higher than for other profile numbers. Results from LMR-LRT suggested that the three-profile model did not significantly outperform the two-profile model. Therefore, for the four-factor model comprising four dimensions, the two-profile model is preferable, with the mean scores for each dimension in the two profiles presented in Table 5.

Table 4 Model comparisons across different profile numbers for four-factor model
Full size table

As illustrated in Fig. 1, the two-profile model revealed distinct patterns. Profile 1 displayed lower scores in positive emotion dimensions, while scores in negative emotion dimensions were higher than the mean. This profile resembles the “Inverted Iceberg” pattern described in previous studies, characterized by high negative and low positive emotion scores. Therefore, it is referred to as the “Inverted Iceberg” profile (74.41%). Conversely, Profile 2 exhibited scores close to the mean for positive emotion dimensions and below the mean for negative dimensions. This profile mirrors the “Iceberg” pattern identified in earlier research, where negative emotion scores are low and positive emotion scores are high. Thus, it is named the “Iceberg” profile (25.59%).

Table 5 Estimated mean scores for the four dimensions of four-factor model across two profiles (M ± SE)
Full size table
Fig. 1
figure 1

Standardized scores of the three profiles of the four-factor model

Full size image

For the six-factor model, model comparison results indicated that the three-profile model was optimal (Table 6). Although the AIC, BIC, and aBIC values continued to decrease, the rate of decrease slowed after the three-profile model. Additionally, the entropy value for the three-profile model was 0.930, higher than that of other models. According to the results of LMR-LRT, the four-profile model was not significantly better than the three-profile model. Therefore, for the original six dimensions of the POMS, the three-profile model was primarily considered optimal. The mean scores for each dimension across the three profiles are shown in Table 7.

The five-profile model appears to outperform the four-profile model. Some researchers have pointed out that although it is possible for the LMR-LRT to yield a significant result after a nonsignificant one, the first nonsignificant result should be taken as a signal to stop further analysis [33]. Therefore, the three-profile model is primarily recommended.

Table 6 Model comparisons for different profile numbers of the six-factor model
Full size table

As shown in Fig. 2, the three-profile model reveals distinct patterns. Profile 1 exhibits relatively uniform scores across all dimensions, resembling the “Surface” profile described in previous research. Therefore, it is designated as the “Surface” profile (33.53%). Profile 2 is characterized by low vigor and high levels of negative emotions, aligning with the “Inverted Iceberg” profile, and is thus labeled as such (19.41%). Profile 3, which displays high vigor and low levels of negative emotions, corresponds to the “Iceberg” profile, and is identified as the “Iceberg” profile (47.06%).

Table 7 Estimated mean scores of the three profiles in the six-factor model (M ± SE)
Full size table
Fig. 2
figure 2

Standardized mean scores of the three profiles of the six-factor model

Full size image

For the seven-factor model, model comparisons across different profile numbers indicated that the three-profile model was optimal (Table 8). While the values of AIC, BIC, and aBIC continued to decrease, the rate of decrease diminished after the three-profile model. Additionally, the entropy at this point was 0.930, higher than for other profile numbers, and the results from LMR-LRT indicated that the four-profile model did not significantly outperform the three-profile model. Therefore, the three-profile model is preferable for the seven dimensions of the POMS. The mean scores for each dimension in the three profiles are presented in Table 9.

Table 8 Model comparisons for different profile numbers of the seven-factor model
Full size table

According to Fig. 3, the seven-factor model identifies three distinct profiles. Profile 1, characterized by low scores in tension, depression, anger, fatigue, and confusion, with high scores in vigor and esteem, aligns with the “Iceberg” profile (47.35%). Profile 2, displaying markedly high scores in tension, depression, anger, fatigue, and confusion, with relatively low vigor and esteem scores, corresponds to the “Inverted Iceberg” profile (19.41%). Profile 3, with relatively uniform scores across all dimensions, is consistent with the “Surface” profile (33.24%).

Table 9 Estimated means of the three profiles in the seven-factor model (M ± SE)
Full size table
Fig. 3
figure 3

Standardized scores of the three profiles of the seven-factor model

Full size image

Discussion

The factor analysis of the Abbreviated POMS extracted one positive emotion dimension and three negative emotion dimensions, retaining only 27 items (Abbreviated four-factor model). The original scale’s seven-factor model did not fit well overall but could be individually fitted as separate subscales. This result may contribute new evidence to the controversy regarding the construct validity of POMS. Most studies supporting the six-dimensional structure of various POMS versions employed CFA for validation [16, 18, 27]. In contrast, EFA often fails to yield the theoretically established structure [12, 13, 15, 20,21,22,23,24].

The primary reason for this inconsistency may be the lack of a solid theoretical basis for the POMS structure. The POMS dimensions were developed based on clinical psychological intervention needs. Even the esteem dimension in the Abbreviated POMS was directly incorporated [1, 7]. Thus, the structure reflects practical experience or diagnostic needs rather than specific emotion theory, which might lead to varied structures when examined in specific populations due to differing relationships and characteristics among emotional categories.

Notably, for the Abbreviated POMS, deficiencies in its structure may be evident only in specific dimensions. In this study, the Abbreviated POMS yielded a four-dimensional structure with 27 items, consisting of anger, positive emotion (vigor and esteem), fatigue, and confusion (tension and confusion). The dimensions of anger, positive emotion, and fatigue mostly retained their original items, while the confusion dimension comprised items primarily from the tension and confusion dimensions. This structure is similar to the anger, vigor, fatigue, and dejection dimensions found in a large-sample study of German students and general populations [22]. However, the depression dimension could be extracted in samples of pregnant women and the elderly [20, 24]. It can be inferred that the four dimensions identified here might represent a general structure of emotional terms included in the POMS for non-clinical populations. In contrast, items related to depression, which are not clearly classified in this study, might be more associated with clinical symptoms or populations with health concerns. The Abbreviated four-factor model derived in this study possibly retains the structure after removing components specific to clinical needs from the POMS.

In the Abbreviated four-factor model, negative emotion dimensions (anger, fatigue, and confusion) were interrelated but not associated with the positive emotion dimension. This result is consistent with previous studies on the POMS among adolescent athletes and the Portuguese short version of the POMS, where negative emotion dimensions were shown to form a higher-order factor [16, 20, 39]. Since mental health encompasses both positive experiences and pathological symptoms as independent aspects [40], considering the POMS’s initial purpose of psychological intervention evaluation, the included positive and negative dimensions may correspond to the two aspects of mental health, reflecting their interrelation.

The dimensions proposed in the theoretical construction of the POMS are not easily extractable from data, and attempts to fit the POMS model directly yield unsatisfactory results. Unlike studies that validated the original POMS dimensional structure through CFA [16, 18, 27], the seven-factor model do not support a scale with interrelated dimensions. Instead, it is more appropriate to view it as a collection of subscales for individual emotions. This finding appears to support the use of individual dimensions within the POMS for measurement in research and practice, as a viable and effective approach [41], which is particularly beneficial for monitoring the training of elite athletes. However, from another perspective, this result may challenge the practice of directly calculating the Total Mood Disturbance (TMD) score by summing negative and positive dimension scores.

It is important to note, however, that the quality of these individual emotional scales varies. Among the seven dimensions, the fitness for the “Confusion” dimension was the least satisfactory. This may be related to the inherent complexity of this emotion. Some researchers suggest that confusion seems to function more as a cognitive feeling, resembling an affective state rather than a true emotion [42]. This could also explain why, in the four-factor model, half of the items originally categorized under the Confusion dimension were reassigned to other emotions, a phenomenon observed in previous studies as well [12]. Furthermore, after removing the Esteem dimension, which was directly added to the Abbreviated POMS, the overall model fit improved, although it still fell short of ideal standards. Given that one of the original intentions behind including the Esteem dimension was to increase the weight of positive emotions [7], this result might indicate that this addition introduced more complications than benefits.

In the profiles derived from the Abbreviated POMS, whether using the seven-factor model, the original six dimensions, or the four-factor model, the classic iceberg (low negative emotion scores and high positive emotion scores) and inverse iceberg (high negative emotion scores and low positive emotion scores) profiles identified in previous research are apparent. However, six-factor and seven-factor model only distinguish an additional surface profile (average scores for all emotion dimensions) beyond these two classic profiles, while the four-factor model yielded only these two classic profiles. The number of effective profiles is much fewer than those identified in previous studies [6, 26, 38, 43].

The “Everest” profile (high levels of vigor and other positive emotions, with lower negative emotions, showing a more pronounced difference compared to the Iceberg profile) and the “Inverted Everest " profile (low levels of vigor and other positive emotions, with high levels of negative emotions, showing a more pronounced difference compared to the Inverted Iceberg profile) did not emerge in this study. Although elite athletes are considered a high-risk group for mental health issues [44, 45], the Inverted “Everest” profile, which is closely associated with mental health risks, was not identified. This suggests that the precision of POMS in screening for mental health risks might be insufficient, and the use of additional tools may be necessary for more accurate assessments [4, 6]. Similarly, the “Submerged” profile (below-average scores across all emotions) was not observed. A large-scale study conducted in Singapore found similar results [46], indicating that individuals engaged in recreational or competitive sports are under-represented in this profile. This phenomenon may be related to the positive effects of physical activity on emotional well-being. Long-term physical activity could provide a physiological basis for positive emotions, reducing the likelihood of the “Submerged” profile. Additionally, the “Shark Fin” profile (high fatigue scores and low scores in other emotions) did not appear in this study. This suggests that the Chinese version of the Abbreviated POMS might not effectively identify states where fatigue is the predominant symptom. Consequently, caution is warranted regarding the use of POMS for assessing fatigue levels and evaluating overtraining syndrome, as recommended in previous studies [5, 47].

It is also noteworthy that in different profiles, negative emotions significantly differ from the mean, while positive emotions remain close to the mean. This is apparent from the profile charts, showing a marked difference from previous studies [6]. This difference might be due to cultural variations. Research indicates that the motivation to reduce unpleasant emotions is lower in East Asian cultures, with less frequent use of emotion regulation strategies [48]. The acceptance and experience of negative emotions or a relatively moderate tendency to maintain positive emotions might contribute to these differences in emotional scores. Although the specific reasons require further analysis, this pattern might indicate that negative emotion levels could be key for POMS to differentiate Chinese athletes.

Overall, the structure of the Abbreviated POMS is not ideal and does not fully support its hypothesized model. However, adjusting the dimensions of the Abbreviated POMS yields a more effective scale structure. Alternatively, if utilizing all seven dimensions, each can be employed as an independent scale. Additionally, both the original and adjusted versions of the scale can be applied to analyze iceberg and inverse iceberg profiles. Based on the results of this study, the following recommendations are suggested for using the Abbreviated POMS with athletes: (1) Use the four-dimensional structure of the Abbreviated four-factor model to measure and separately calculate scores for positive and negative emotion dimensions; (2) when using the 40-item version, select specific dimensions for testing and interpretation according to need, but exercise caution with the Confusion and Esteem dimensions; (3) Interpret results using the iceberg and inverse iceberg profiles as a reference.

However, this study offers only a preliminary analysis of the Abbreviated POMS structure and presents several notable limitations. First, the study employed a one-week time frame, excluding immediate or single-day measures. While the consistency of the structure has been demonstrated for both weekly and immediate time frames in athletes [28], the latest POMS 2 also recommends a one-week measurement frame [49]. However, immediate and single-day timeframes hold significant practical value in competitive sports. Evaluating these timeframes could offer more targeted insights for using or refining the Abbreviated POMS. Second, the study did not differentiate athletes by skill level or role, potentially undermining the interpretive robustness of the structure across different populations and overlooking population-specific characteristics. Additionally, the study did not conduct in-depth analysis or exploration of dimensions such as Confusion, which displayed some challenges in its structure. Lastly, the study did not explore the substantive nature of emotions measured by the Abbreviated POMS. It remains unclear whether POMS results reflect emotional traits or predictable emotional states [49,50,51]. This ambiguity affects the efficacy of profile analysis, a topic that has long been debated [2, 4]. These limitations impede a deeper exploration of the relationship between POMS measurement outcomes, competitive performance, and athlete characteristics.

Future research should undertake more comprehensive examinations of the Abbreviated POMS structure across diverse time frames and among athletes with varying roles and skill levels. Additionally, further investigation into the predictive validity of its measurement structure is warranted. Furthermore, as POMS has evolved to its second edition [49], integrating POMS 2 or incorporating its enhancements into the current Abbreviated POMS within the sports context could offer a more effective and reliable tool for assessing athletes’ emotions.

Conclusions

The structure of the Abbreviated Profile of Mood States (POMS) among Chinese athletes is suboptimal and does not support its initial hypothesized model. Simplifying the Abbreviated POMS to the four-factor model, comprising 27 items across four dimensions: positive mood, anger, fatigue, and confusion, appears more viable. Both the four-dimensional model and evaluating each dimension independently for specific emotional states fit relatively well, offering flexibility in application according to varying needs. The presence of iceberg and inverse iceberg profiles persists across different structural approaches and can effectively categorize athletes’ emotional states.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Abbreviations

POMS:

Profile of Mood States

EFA:

Exploratory factor analysis

CFA:

Confirmatory factor analysis

LPA:

Latent profile analysis

AIC:

Akaike Information Criterion

BIC:

Bayesian Information Criterion

Abic:

Adjusted BIC

LMR-LRT:

Lo-Mendell-Rubin likelihood ratio test

BLR:

Bootstrap Likelihood Ratio Test

KMO:

Kaiser-Meyer-Olkin

References

  1. Spielberger CD. Review of Profile of Mood States [Review of the book Profile of mood states, by D. M. McNair, M. Lorr & L. F. Droppleman]. Prof Psychol. 1972;3:387–8. https://doi.org/10.1037/h0020742.

  2. Leunes A, Burger J. Profile of mood states research in sport and exercise psychology: past, present, and future. J APPL SPORT PSYCHOL. 2000;12(1):5–15.

  3. Leunes A, Burger J. Bibliography on the Profile of Mood States in sport and exercise psychology research, 1971–1998. J Sport Behav. 1998;21(1):53.

    Google Scholar 

  4. Renger R. A review of the Profile of Mood States (POMS) in the prediction of athletic success. J APPL SPORT PSYCHOL. 1993;5(1):78–84.

    Article  Google Scholar 

  5. Carrard J, Rigort A, Appenzeller-Herzog C, Colledge F, Königstein K, Hinrichs T, et al. Diagnosing overtraining syndrome: a scoping review. Sports Health. 2022;14(5):665–73.

    Article  PubMed  Google Scholar 

  6. Heikura EE, Aunola K, Tolvanen A, Ryba TV, Selänne H. Student-athletes’ mood state profiles: the role of sports, sex, and performance level in sports and in school. J Appl Sport Psychol. 2023;35(5):836–53.

    Article  Google Scholar 

  7. Grove JR, Prapavessis H. Preliminary evidence for the reliability and validity of an abbreviated Profile of Mood States. Int J Sport Psychol. 1992;23:93–109.

    Google Scholar 

  8. Zhu B. Brief introduction of POMS scale and its model for China. J Tianjin Inst Phys Educ. 1995;10(1):35–7.

    Google Scholar 

  9. Martin DT, Andersen MB & Gates W. Using Profile of Mood States (POMS) to Monitor High-Intensity Training in Cyclists: Group versus Case Studies. The Sport Psychol. 2000;14(2):138–56. https://doi.org/10.1123/tsp.14.2.138.

  10. Rowley AJ, Landers DM, Kyllo LB, Etnier JL. Does the iceberg profile discriminate between successful and less successful athletes? A meta-analysis. J Sport Exerc Psychol. 1995;17(2):185–99.

    Article  Google Scholar 

  11. Beedie CJ, Terry PC & Lane AM. The profile of mood states and athletic performance: two meta-analyses. J Appl Sport Psychol. 2000;12(1):49–68. https://doi.org/10.1080/10413200008404213.

  12. Bourgeois A, LeUnes A, Meyers M. Full-scale and short-form of the Profile of Mood States: a factor analytic comparison. J Sport Behav. 2010;33(4):355–76.

    Google Scholar 

  13. López-Jiménez CM & Cano-García FJ, Sanduvete-Chaves S, Chacón-Moscoso S. Profile of Mood States Factor Structure Does Not Accurately Account for Patients with Chronic Pain. Pain Med. 2021;22(11):26049–14. https://doi.org/10.1093/pm/pnab127.

  14. Janelle CM, Fawver BJ, Beatty GF. Emotion and sport performance. Handbook of sport psychology. 2020:254 – 98.

  15. Boyle GJ. A cross-validation of the factor structure of the Profile of Mood States: were the factors correctly identified in the First Instance? PSYCHOL REP. 1987;60(2):343–54. https://doi.org/10.2466/pr0.1987.60.2.343.

  16. Terry PC, Lane AM, Fogarty GJ. Construct validity of the Profile of Mood States—adolescents for use with adults. Psychol Sport Exerc. 2003;4(2):125–39.

    Article  Google Scholar 

  17. Curran SL, Andrykowski MA, Studts JL. Short form of the profile of mood states (POMS-SF): psychometric information. PSYCHOL Assess. 1995;7(1):80.

    Article  Google Scholar 

  18. Reddon JR, Marceau R & Holden RR. A confirmatory evaluation of the profile of mood states: convergent and discriminant item validity. J Psychopathol Behav. 1985;7(3):243–59. https://doi.org/10.1007/BF00960756.

  19. Shacham S. A Shortened Version of the Profile of Mood States. J Pers Assess. 1983;47(3):305–6. https://doi.org/10.1207/s15327752jpa4703_14.

  20. Aguiar M, Barros M, Macedo A, Pereira AT. Profile of mood states-27: a valid and reliable measure of negative and positive affect for Brazilian pregnant women. Eur Psychiat. 2021;64(S1):S386–7. https://doi.org/10.1192/j.eurpsy.2021.1036.

    Article  Google Scholar 

  21. Pereira AT, Araújo AI, Cabaços C, Brito MJ, Fernandes M, Rodrigues A, et al. Profile of mood states-12: same validity, more usability. Eur Psychiat. 2023;66(S1):S553–4. https://doi.org/10.1192/j.eurpsy.2023.1168.

    Article  Google Scholar 

  22. Petrowski K, Albani C, Zenger M, Brähler E & Schmalbach B. Revised short screening version of the Profile of Mood States (POMS) from the German general population. Front Psychol. 2021;12:631668. https://doi.org/10.3389/fpsyg.2021.631668

    Article  PubMed  PubMed Central  Google Scholar 

  23. Chen K, Snyder M, Krichbaum K. Translation and equivalence: the profile of mood states short form in English and Chinese. Int J Nurs Stud. 2002;39(6):619–24.

    Article  PubMed  Google Scholar 

  24. Shin YH. A study on verification of the profile of mood states (POMS) for Korean elders. J Nurses Acad Soc. 1996;26(4):743–58.

    Article  Google Scholar 

  25. Wang J, Lin W, Chen Z. POMS for use in China. Acta Psychol Sin. 2000;32(01):110–4.

    CAS  Google Scholar 

  26. Terry P, Parsons-Smith R. Identification and incidence of mood profile clusters among sport participants. J Sci Med Sport. 2019;22:S100.

    Article  Google Scholar 

  27. Zhang C, Si G, Chung P, Du M & Terry PC. Psychometric properties of the Brunel Mood Scale in Chinese adolescents and adults. J Sport Sci. 2014;32(15):1465-76. https://doi.org/10.1080/02640414.2014.898184.

  28. Andrade E, Rodríguez D. Factor structure of mood over time frames and circumstances of measurement: two studies on the Profile of Mood States questionnaire. PLoS ONE. 2018;13(10):e0205892.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Worthington RL, Whittaker TA. Scale development research: a content analysis and recommendations for best practices. Couns Psychol. 2006;34(6):806–38.

    Article  Google Scholar 

  30. Kenny DA, Kaniskan B & McCoach DB. The performance of RMSEA in Models with Small degrees of Freedom. Sociol Method Res. 2014;44(3):486–507. https://doi.org/10.1177/0049124114543236.

  31. Shi D, DiStefano C, Maydeu-Olivares A, Lee T. Evaluating SEM model fit with small degrees of freedom. Multivar Behav Res. 2022;57(2–3):179–207.

    Article  Google Scholar 

  32. Shi D, Lee T & Maydeu-Olivares A. Understanding the model size effect on SEM fit indices. Educ Psychol Meas. 2018;79(2):310–34. https://doi.org/10.1177/0013164418783530.

  33. Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equation Modeling: Multidisciplinary J. 2007;14(4):535–69.

    Article  Google Scholar 

  34. Meyer JP, Morin AJ. A person-centered approach to commitment research: theory, research, and methodology. J Organ Behav. 2016;37(4):584–612.

    Article  Google Scholar 

  35. Morin AJ, Meyer JP, Creusier J, Biétry F. Multiple-group analysis of similarity in latent profile solutions. Organ Res Methods. 2016;19(2):231–54.

    Article  Google Scholar 

  36. O’Neil L, Amorose AJ, Pierce S. Student-athletes’ dual commitment to school and sport: compatible or conflicting? Psychol Sport Exerc. 2021;52:101799. https://doi.org/10.1016/j.psychsport.2020.101799.

    Article  Google Scholar 

  37. Shannon S, Prentice G, Breslin G. Athletes’ psychological needs and coaches’ interpersonal behaviors: a within-Person Latent Profile Analysis. J Sport Exerc PSY. 2021;43(1):71–82. https://doi.org/10.1123/jsep.2019-0295.

    Article  Google Scholar 

  38. Parsons-Smith RL, Terry PC, Machin MA. Identification and description of novel mood profile clusters. Front Psychol. 2017;8:302382.

    Article  Google Scholar 

  39. Terry PC, Lane AM. Normative values for the profile of mood states for use with athletic samples. J Appl Sport Psychol. 2000;12(1):93–109.

    Article  Google Scholar 

  40. DiLeo LL, Suldo SM, Ferron JM & Shaunessy-Dedrick E. Three-Wave Longitudinal Study of a Dual-Factor Model: Mental Health Status and Academic Outcomes for High School Students in Academically Accelerated Curricula. Sch Ment Health. 2022;14(3):514 – 30. https://doi.org/10.1007/s12310-021-09497-9.

  41. Hori M, Yoshikawa E, Hayama D, Sakamoto S, Okada T, Sakai Y, et al. Sense of coherence as a Mediator in the Association between Empathy and moods in Healthcare professionals: the moderating effect of age. Front Psychol. 2022;13:847381.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Hess U. Now you see it, now you don’t–the confusing case of confusion as an emotion: Commentary on Rozin and Cohen (2003). 2003.

  43. Terry PC, Parsons-Smith RL, King R, Terry VR. Influence of sex, age, and education on mood profile clusters. PLoS One. 2021;16(2):e0245341.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Gouttebarge V, Castaldelli-Maia JM, Gorczynski P, Hainline B, Hitchcock ME, Kerkhoffs GM, et al. Occurrence of mental health symptoms and disorders in current and former elite athletes: a systematic review and meta-analysis. Brit J Sport Med. 2019;53(11):700–6.

    Article  Google Scholar 

  45. Peluso MAM, Andrade LHSG. Physical activity and mental health: the association between exercise and mood. Clinics. 2005;60(1):61–70.

    Article  PubMed  Google Scholar 

  46. Han CSY, Parsons-Smith RL, Terry PC. Mood Profiling in Singapore: cross-cultural validation and potential applications of Mood Profile clusters. Front Psychol. 2020;11.

  47. Grant CC, Janse Van Rensburg DC, Collins R, Wood PS, Du Toit PJ. The Profile of Mood State (POMS) questionnaire as an indicator of Overtraining Syndrome (OTS) in endurance athletes. Afr J Phys Health Educ Recreation Dance. 2012;18(sup–1):23–32.

    Google Scholar 

  48. Kaspi L, Hu D, Vishkin A, Chentsova-Dutton Y, Miyamoto Y, Cieciuch J et al. Motivated Emot Regul Cultures Emot. 2024:(In-Press).

  49. Lin S, Hsiao Y, Wang M. Review of the Profile of Mood States 2nd Edition (POMS 2). J Psychoeduc Assess. 2014;32:273–7. https://doi.org/10.1177/0734282913505995.

    Article  Google Scholar 

  50. Lochbaum M, Zanatta T, Kirschling D, May E. The Profile of Moods States and athletic performance: a meta-analysis of published studies. Eur J Invest Health Psychol Educ. 2021;11(1).

  51. Nagle FJ, Morgan WP, Hellickson RO, Serfass RC, Alexander JF. Spotting success traits in olympic contenders. Physician Sportsmed. 1975;3(12):31–4.

    Article  Google Scholar 

Download references

Acknowledgements

The authors would like to thank all the athletes for participating in our study.

Funding

This work was supported by the Science and Technology Commission of Shanghai Municipality (grant number 22dz1204601).

Author information

Authors and Affiliations

  1. Shanghai Research Institute of Sports Science (Shanghai Anti-doping Agency), Shanghai, China

    Chenhao Tan, Jun Yin, Yan An, Jinhao Wang & Jun Qiu

  2. Key Laboratory of General Administration of Sport for Exercise Performance Evaluation, Shanghai, China

    Chenhao Tan, Jun Yin, Yan An, Jinhao Wang & Jun Qiu

Authors
  1. Chenhao Tan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jun Yin
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yan An
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Jinhao Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jun Qiu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

CT and JQ designed and conceptualized the study. CT performed the data analysis and wrote the manuscript. JQ provided critical comments. CT, JY, YA and JW collected the data and provided critical comments on the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Jun Qiu.

Ethics declarations

Ethics approval and consent to participate

Ethical approval for the study was obtained from the Ethics Committee of Shanghai Research Institute of Sports Science (Shanghai Anti-doping Agency) (Ethics Approval Number LLSC 20220005), and all procedures adhere to the ethical principles outlined in the latest version of the Declaration of Helsinki. Written consent was obtained from all participants before the commencement of the test.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tan, C., Yin, J., An, Y. et al. The structural validity and latent profile characteristics of the Abbreviated Profile of Mood States among Chinese athletes. BMC Psychiatry 24, 636 (2024). https://doi.org/10.1186/s12888-024-06092-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12888-024-06092-5

Keywords

BMC Psychiatry

ISSN: 1471-244X

Contact us