Abstract
General Description
The Profile of Mood States 2nd Edition (POMS 2) was published in 2012 by Multi-Health Systems (MHS) to assess transient feelings and mood among individuals aged 13 years and above. Evolving from the original POMS (McNair, Lorr, & Droppleman, 1971, 1992), the POMS 2 was designed for youth (13-17 years old) and adults (18 years old and above) to measure their affective traits, moods, and emotions. The POMS 2 contains four self-report versions: the Profile of Mood States 2nd Edition–Adult (POMS 2-A), the Profile of Mood States 2nd Edition–Youth (POMS 2-Y), the Profile of Mood States 2nd Edition–Adult Short (POMS 2-A Short), and the Profile of Mood States 2nd Edition–Youth Short (POMS 2-Y Short). The POMS 2 is applicable and useful to multiple settings. For clinical practitioners, the POMS 2 instruments can monitor change in mood disturbance during or following intervention. In medical settings, the POMS 2 helps to assess the impact of diagnosis and treatment of physical disease on psychological functioning. In the athletic domain, the test can be used to study the relationship between mood and exercise or physical activity.
Specific Details
The POMS 2 can be completed in either online or paper format within 8 to 10 min for the full version or 3 to 5 min for the short version. The POMS 2 contains six original subscales from the POMS, including Anger-Hostility (AH), Confusion-Bewilderment (CB), Depression-Dejection (DD), Fatigue-Inertia (FI), Tension-Anxiety (TA), and Vigor-Activity (VA). Unlike the POMS, the POMS 2 contains a summary scale called Total Mood Disturbance (TMD) and a Friendliness (F) subscale for testing positive mood complementally (Table 1). Also, the POMS 2 authors modified the POMS items for the extended youth version. Hence, there are 65 items in the POMS 2-A, 60 items in the POMS 2-Y, and 35 items in both short versions.
Descriptions of the POMS 2 Scales.
Note. POMS 2= Profile of Mood States (2nd Edition).
Scoring System
The POMS 2 scales are scored via the MHS Online Assessment Center. The system uses double entry to reduce the possibility of omitted responses. After that, if only one omitted item exists in each subscale, the prorated raw score method is used. If any subscale has more than one item omission, that subscale’s score will not be calculated.
The TMD score equals the sum of six subscales and does not include the F subscale. Raw scores are converted into T-scores with a mean equal to 50 and a standard deviation equal to 10. Confidence intervals and percentile ranks are included. Scores can be compared among different subgroups based on different ages and gender.
Test Materials and Stimuli
A standard kit consists of the POMS 2-A Form, the POMS 2-A Short Form, the POMS 2-Y Form, the POMS 2-Y Short Form, and the POMS 2 manual. The kit has online and paper versions and is user-friendly. Every POMS 2 item uses 0 to 4 to indicate the intensity of mood. The POMS 2 forms are orderly and easy for respondents to mark. The POMS 2 manual provides separate parts for theoretical background, administration standards, scoring and report, interpretation, norms, and psychometric properties. Each part is well introduced with elaborate instructions.
Test Construction
The POMS 2 revised the POMS by (a) providing more normative and psychometric samples, (b) adding more positive mood states, (c) modernizing outdated items, (d) deleting culturally specific items, and (e) creating the youth version and the short versions for the POMS 2-A and the POMS 2-Y. Based on those general goals, 89 items formed the candidate item set for the POMS 2-A and 69 items built the candidate item set for the POMS 2-Y. Those candidate items were then retained only when they had strong psychometric properties: no floor or ceiling effect, high variance, good item-total correlation, high discrimination among subgroups, strong face validity, and reasonable item characteristic curve in Item Response Theory (IRT) analyses. Thus, for the full version, 65 items were adopted to form the POMS 2-A and 60 items were used for the POMS 2-Y. For the short version, high item-total correlation and high regression prediction ability were set as criteria for selecting 5 items from each subscale of the full versions, resulting in a total of 35 items for the POMS 2-A Short and the POMS 2-Y Short.
Standardization Sample
All standardized information was gleaned by 62 site coordinators across the United States and Canada. Samples for the POMS 2-A were recruited from 2004 to 2009 and samples for the POMS 2-Y were recruited from 2006 to 2009. Samples’ gender, age, race/ethnicity, test-takers’ level of educational attainment (or parents’ highest level of education), and geographic region were collected. Statistical weights were applied to minimize the discrepancy of race/ethnicity and education level between samples and the Census. Like the POMS, the POMS 2 included normative samples and clinical samples diagnosed with anxiety disorder, major depressive disorder, or other psychiatric disorders. In all, 1,000 normative respondents comprised the POMS 2-A and 500 normative respondents were used to generate the POMS 2-Y norms; data from 215 clinical participants were collected for the POMS 2-A and 133 clinical participants were used for the POMS 2-Y.
Reliability
The Cronbach’s alpha coefficient was used for evaluating the internal consistency across each subscale’s scores with normative and clinical samples. Generally, coefficients ranged from .82 to .96 for the POMS 2-A, .80 to .95 for the POMS 2-A Short, .78 to .96 for the POMS 2-Y, and .81 to .95 for the POMS 2-Y Short. Results from alpha values indicated that items in all versions of the POMS 2 were correspondent with each other and with the test as a whole.
Test–retest reliability was calculated using the adjusted correlation coefficient (Cohen, Cohen, West,& Aiken, 2003) between the initial assessment (Time 1, T1) and a retest after a week (Time 2, T2), and between T1 and another retest after 30 days (Time 3, T3). Between T1 and T2, the adjusted correlations for each subscale’s scores ranged from .48 to .72 (all ps < .01) for the POMS 2-A, .52 to .80 (all ps < .01) for the POMS 2-A Short, .45 to .75 (all ps < .01) for the POMS 2-Y, and .46 to .81 (all ps < .01) for the POMS 2-Y Short except for the F subscale (r = .31). The adjusted correlations for the domain scores of four POMS 2 versions decreased for the time interval between T1 and T3. For the two POMS 2-A versions, the adjusted r ranged from .32 to .73 for each subscale (all ps < .01). For the two POMS 2-Y versions, only 3 and 5 out of 7 subscale scores had test–retest reliability over .30 for the long and the short versions, respectively. In addition, the F subscale showed the poorest test–retest reliability (r = .02). Results from the adjusted correlations show that the test scores were moderately consistent over a 1-week period. However, for youth and adult participants, test–retest reliability decreased for the 30-day interval.
Validity
Validity is examined in four parts: factor structure, discriminative validity, convergent validity, and generalization across race/ethnic groups. Factor structure evidence was derived from the results of confirmatory factor analysis (CFA), and hierarchical models were tested by the authors; the F subscale was not included in the analysis. The results show the data were adequately fitted to the hierarchical model for the POMS 2-A (normed fit index [NFI] = .92, nonnormed fit index [NNFI] = .91, comparative fit index [CFI] = .93, root mean square error of approximation [RMSEA] = .10) and the POMS 2-Y (NFI = .92, NNFI = .92, CFI = .94, and RMSEA = .10).
Discriminative validity was examined by conducting ANCOVAs/MANCOVAs. For the adult versions, participants in the clinical group reported higher negative scores and lower positive scores than participants in the normal group (all Fs significant, p < .001), which shows good discriminative validity. For youth versions, participants in the clinical group reported higher negative scores than participants in the normal group (all Fs significant, p < .001); no differences were reported for positive moods. Therefore, for youths, only subscales regarding negative mood states of the POMS 2 had adequate discriminative validity.
Convergent validity was examined via the Pearson correlations between scores on the POMS 2 and the Positive and Negative Affect Schedule–Expanded Form (PANAS-X; Watson & Clark, 1994). Correlations between selected scales on the POMS 2-A and PANAS-X ranged from .57 to .84 (all ps < .001), indicating adequate convergent validity. Convergent validity evidence for POMS 2-Y, POMS 2-A Short, and POMS 2-Y Short were not provided in the manual.
Generalizability across ethnicity was tested by conducting ANCOVA/MANCOVA. Follow-up one-way ANOVAs were conducted when the ANCOVA/MANCOVA results were significant. For the adult version, positive and negative scores were not different among different ethnicity groups, suggesting that the results of adults can be generalized across ethnicity groups. In contrast, White youths reported higher negative scores and lower positive scores than their Hispanic and African American peers.
Commentary and Recommendations
The POMS 2 was developed to assess affective traits, mood, and emotion. The standardization samples included children and adults across many areas of the United States and Canada, and included normative and clinical samples. Reliability and validity evidence has been provided from several analyses.
The POMS 2 has a number of strengths. First, the POMS 2 was designed to be widely used based on the following characteristics: (a) It can measure diverse users, including normative or clinical youths and adults; (b) administration time is short; and (c) there is an online version. Second, the examination of the psychometric properties was strictly conducted. Reliability analyses (internal consistency and test–retest reliability) and most validity analyses (factor structure, discriminative validity, and generalizability across ethnicity) were examined among adult and youth versions of the POMS 2. Third, instructions of this assessment were clearly given in the manual. For example, the manual recommends 1 week as the time interval for testing stability. Also, it pointed out that White youths reported slightly higher negative feelings. Such information is useful and could prevent future users from misusing the POMS 2.
Several weaknesses and recommendations, however, should be noted. The first category relates to the F scale. Theoretically, some items of the F scale were highly correlated with the VA scale, which indicates that the construct overlapped between them. From the practical aspect, scores from the F subscale were isolated from the composite score TMD, and the factor structure of the POMS 2 was examined using all subscales except the F scale. These factors hamper the clear interpretation of the F and the TMD scores. From the psychometric view, the test–retest reliabilities for the F scale were poor, especially for the youths in a longer period (i.e., r = .02), showing that the precision of the F scores was dubious. The test authors are encouraged to reselect items that are less related with the other constructs and redesign the “Friendliness” scale.
Second, the protocol of neglecting scale scores with more than one item missing might be too strict. According to this protocol, the subscales’ scores and the related TMD scores would fail to be calculated for some users, which further causes the loss of special test information and specific subgroups. Thus, to make sure this test has no bias for specific subgroups, test users may compare the demographic characteristics and test information between the subgroups that have complete mood scores and the subgroups whose scores are deleted.
Finally, some information was overlooked in the POMS 2 Manual. First, extra graphs and tables for explaining the results of the “Friendliness” scores are missing. Second, detailed information about the sampling procedures was lacking. Not addressing the sampling method could prevent readers from making fair judgments of the representativeness of the samples. Third, the manual provides convergent validity evidence for only the POMS 2-A; no such evidence is provided for the other three versions of the POMS 2.