Assessing Psychological Symptoms and Well-Being

Subjective Well-Being

Subjective well-being is one increasingly studied construct that incorporates positive indicators of mental health beyond the absence of negative factors. It can be defined as an individual’s currently experienced quality of life and includes both cognitive and affective components. Positive and negative affect are considered the emotional components of subjective well-being; specifically, subjective well-being involves experiencing a high frequency of positive emotions and low frequency of negative emotions. Life satisfaction is the cognitive component of subjective well-being and involves an individual’s overall appraisal of his or her life quality (Diener, 1984).

An increasing body of evidence indicates that subjective well-being facilitates a number of desirable outcomes. Specifically, positive emotional well-being consistently predicts higher quality social relationships and increased relationship satisfaction (Lyubomirsky, King, & Diener, 2005). Occupationally, individuals with high subjective well-being are more likely to have successful job interviews, and their performance on the job is rated more positively by supervisors (Burger & Caldwell, 2000; Wright & Cropanzano, 2000). Subjective well-being even facilitates better physical health and increased longevity (Abel & Kruger, 2010; Graham, Eggers, & Sukhtankar, 2004). Thus, subjective well-being is not just an advantageous outcome in itself but also a facilitator of optimal functioning (Diener, 2012).

Dual-Factor Models of Mental Health

Although well-being and distress are often conceptualized as opposite ends of a single continuum, an alternative model proposes that positive well-being and psychological distress are two related but unique constructs contributing to overall mental health (Greenspoon & Saklofske, 2001). Keyes (2005) tested this model empirically using confirmatory factor analysis in which one model included well-being and illness on a single dimension, whereas another included the two factors as distinct but correlated constructs. Results indicated a significantly better fit for the two-factor model, supporting that well-being and distress are best conceptualized as two separate constructs, rather than opposite poles on one dimension (Keyes, 2005).

Several studies have investigated the application of a dual-factor mental health model in children and adolescents (Antaramian, Huebner, Hills, & Valois, 2010; Greenspoon & Saklofske, 2001; Lyons, Huebner, Hills, & Shinkareva, 2012; Suldo & Shaffer, 2008). In these studies, participants’ mental health was classified based on whether their subjective well-being was high or low and whether their psychopathology was high or low, yielding four distinct groups. First, some youth had optimal mental health, with high subjective well-being and low psychopathological symptoms. The second group of youth had the most significant mental health issues, with high psychopathology and low subjective well-being. These are the two groups most likely to be distinguished using a traditional one-factor mental health model. However, a third group was identified as having low psychopathology but also low subjective well-being, reflecting less than optimal mental health. The final group had high levels of psychopathology but also high subjective well-being, despite their psychological symptoms. Thus, a dual-factor model was usefully applied to distinguish children and adolescents with different mental health profiles. Moreover, evidence indicated that these four groups also differed on key individual characteristics and desirable outcomes, including self-esteem, self-efficacy, locus of control, extraversion, neuroticism, student engagement, academic achievement, motivation, perceived social support, and physical health (Antaramian et al., 2010; Greenspoon & Saklofske, 2001; Lyons et al., 2012; Suldo & Shaffer, 2008).

The evidence is much more limited for the application of this model in college students; however, Eklund, Dowdy, Jones, and Furlong (2011) found support for this approach in one college sample. Using measures of positive well-being and clinical symptoms, the researchers were able to classify a sample of college students into the same four mental health categories described above. In addition, results demonstrated that these groups differed in their levels of hope, gratitude, locus of control, and attention problems. Thus, indicators of both positive well-being and clinical symptoms were needed to categorize the functioning of college students in this study (Eklund et al., 2011).

The Current Study

The current study seeks to extend previous research by providing further evidence for the utility of the dual-factor mental health model with college students. Although early research supports its use in the college population, more evidence is needed to fully establish whether this model is appropriate for understanding the unique mental health needs of college students. The present research seeks to build on this evidence by examining the model with a larger, demographically distinct college sample compared with that used in prior research. Furthermore, the current study extends existing research by examining the dual-factor model in relation to different outcomes. Previous research with college students investigated the dual-factor model only in relation to other measures of psychological functioning (Eklund et al., 2011). The current study adds to this work by examining the model in the context of students’ educational outcomes and investigating whether students differed not only in their mental health profiles, but also in their academic success. One frequently measured indicator of academic functioning is student engagement. Although exact definitions vary, student engagement can be conceptualized as a broad construct that includes investment in academic tasks as well as social involvement and participation in the university community (Krause & Coates, 2008). Engagement is related to several important academic factors, including better critical thinking, higher grade point averages (GPAs), increased likelihood of returning to the university, and greater success in attaining a degree (Carini, Kuh, & Klein, 2006; Kuh, Cruce, Shoup, Kinzie, & Gonyea, 2008; Svanum & Bigatti, 2009). In fact, students with high levels of engagement are approximately 1.5 times more likely to graduate than their low-engagement peers, and they graduate approximately one semester earlier (Svanum & Bigatti, 2009). Accordingly, engagement is an important indicator of students’ involvement in their college experience and their likelihood of academic success. The current study, therefore, investigated several dimensions of student engagement, as well as GPA, using the dual-factor model to determine how students’ mental health related to their educational performance. The aim was to answer two specific questions. First, can this model be used to identify distinct groups of college students based on subjective well-being and psychological symptoms? Second, do these mental health groups differ in educational outcomes, including student engagement and GPA? It was hypothesized that four unique groups would be identified, consistent with prior research, and that individuals with high subjective well-being and low symptoms would have the highest engagement and achievement.

Method

Results

Descriptive Statistics and Intercorrelations

Descriptive statistics and correlations for all study variables are presented in Table 1. Moderate correlations among the mental health variables are consistent with the conceptualization of positive well-being and psychopathology as distinct but related constructs (Keyes, 2005). The significant positive correlations between GPA and the engagement dimensions were also expected based on prior research demonstrating that student engagement facilitates academic success (Carini et al., 2006; Kuh et al., 2008; Svanum & Bigatti, 2009).

OPEN IN VIEWER

Table 1.

Descriptive Statistics and Correlations Among Mental Health and Educational Outcomes (N = 561).

	1	2	3	4	5	6	7	8	9	10	11
1. Life satisfaction	—
2. Positive affect	.44**	—
3. Negative affect	−.34**	−.19**	—
4. Internalizing (depression)	−.50**	−.43**	.63**	—
5. Externalizing (aggression)	−.25**	−.13**	.29**	.35**	—
6. Academic engagement	.28**	.32**	−.12**	−.15**	−.19**	—
7. Peer engagement	.26**	.23**	−.04	−.14**	−.12**	.35**	—
8. Faculty engagement	.16**	.25**	−.11*	−.16**	−.14**	.38**	.17**	—
9. Intellectual engagement	.28**	.32**	−.10*	−.15**	−.13**	.50**	.25**	.57**	—
10. Beyond-class engagement	.35**	.23**	−.14**	−.29**	−.22**	.30**	.39**	.31**	.25**	—
11. GPA	.16**	.07	−.03	−.06	−.12**	.38**	.09*	.22**	.23**	.10*	—
M	26.26	3.32	2.32	16.27	2.31	3.08	2.97	4.05	3.41	3.91	3.04
SD	5.48	0.71	0.72	9.92	0.62	0.75	0.86	0.62	0.88	0.95	0.48

Note. GPA = grade point average.

*

p < .05. **p < .01.

Mental Health Group Classification

Participants’ mental health was classified based on their self-reported subjective well-being and psychological symptoms. First, following procedures used in previous research (Antaramian et al., 2010; Suldo & Shaffer, 2008), a composite subjective well-being score was calculated by adding standardized life satisfaction and PA scores and subtracting standardized NA scores. Second, scores on the mental health variables were dichotomized to classify each participant as high or low on subjective well-being and high or low on symptoms. The decision points were similar to those used by Eklund and colleagues (2011). Participants were considered to have high subjective well-being if they scored at or above the mean on the subjective well-being composite, and they were classified as having low subjective well-being if they scored below the mean. For symptoms, a T-score of 60 was chosen as the decision point based on commonly used guidelines for other clinical scales (Reynolds & Kamphaus, 2004). For the scales in the current study, a T-score of 60 was determined by calculating the score that was one standard deviation above the mean on each scale. Participants were classified as having high levels of symptoms if they met or exceeded this score on either internalizing or externalizing. Participants below this score on both scales were categorized as having low levels of symptoms.

Using these procedures, four distinct mental health groups were identified, and subsequent nomenclature for the groups is consistent with that of Eklund et al. (2011). The distribution of participants in the groups is presented in Table 2. Nearly half of the participants were considered well-adjusted because they had above-average subjective well-being and low levels of symptoms. Approximately 21% of the participants were classified as the most distressed with elevated symptoms and low subjective well-being. More than one quarter of the study participants were classified as at-risk, with no significant symptoms but also low subjective well-being. Finally, a small minority of the participants were classified as ambivalent because they had high symptoms but also above-average subjective well-being.

OPEN IN VIEWER

Table 2.

Mental Health Groups Yielded From a Dual-Factor Model.

	Subjective well-being
Psychological symptoms	High	Low
Low	Well-adjusted	At-risk
	n = 266	n = 146
	47.4%	26.0%
High	Ambivalent	Distressed
	n = 31	n = 118
	5.5%	21.0%

To determine whether the mental health groups differed in demographic characteristics, several chi-square tests were conducted. Group membership was not significantly related to ethnicity or class year. However, there was a significant relationship between gender and group membership, χ²(3, N = 561) = 15.321, p = .002. Male participants were overrepresented in the highly symptomatic groups, whereas female participants were overrepresented in the low-symptom groups. Accordingly, gender was included as a covariate in subsequent analyses.

Group Differences in Engagement and Achievement

A multivariate analysis of covariance (MANCOVA) was used to determine whether there were group differences in student engagement and GPA. The MANCOVA assumptions were tested and were within acceptable ranges. Results of the analyses indicated an overall effect of mental health status on the educational outcomes, Wilks’s Lambda = .85, F(18, 1559) = 4.95, p < .001. Subsequent univariate analyses of covariance (ANCOVAs) indicated significant group differences in all six outcomes, including academic engagement, F(3, 556) = 13.39, p < .001; peer engagement, F(3, 556) = 7.63, p < .001; faculty engagement, F(3, 556) = 8.97, p < .001; intellectual engagement, F(3, 556) = 13.70, p < .001; beyond-class engagement, F(3, 556) = 16.10, p < .001; and GPA, F(3, 556) = 3.03, p = .03.

Pairwise comparisons using the Bonferroni correction were used to determine which groups were significantly different. Results of these comparisons and the descriptive statistics for each group are presented in Table 3. On all five engagement dimensions, the well-adjusted group (high subjective well-being, low symptoms) scored significantly higher than the distressed group (low subjective well-being, high symptoms). The effect sizes for these differences were moderate to large, with Cohen’s d ranging from 0.52 to 0.74. The well-adjusted group also had significantly higher levels of engagement in comparison with the at-risk group (low subjective well-being, low symptoms). Effect sizes for these differences were moderate, with Cohen’s d equal to 0.50, 0.31, 0.40, 0.58, and 0.32 for academic, peer, faculty, intellectual, and beyond-class engagement, respectively. The ambivalent group (high subjective well-being, high symptoms) was not significantly different from the other groups on most of the engagement dimensions. However, they did score significantly higher than the distressed group on beyond-class engagement, d = 0.73.

OPEN IN VIEWER

Table 3.

Group Differences in Student Engagement and GPA.

	Mental health group
	Well-adjusted		At-risk		Ambivalent		Distressed
	M	SD	M	SD	M	SD	M	SD
Academic engagement	3.28a	0.70	2.94b	0.68	3.00a,b	0.89	2.83b	0.78
Peer engagement	3.14a	0.81	2.88b	0.88	3.00a,b	0.97	2.71b	0.87
Faculty engagement	4.19a	0.57	3.96b	0.62	4.06a,b	0.64	3.87b	0.69
Intellectual engagement	3.64a	0.83	3.17b	0.78	3.48a,b	1.15	3.16b	0.91
Beyond-class engagement	4.12a	0.88	3.84b	0.92	4.14a,b	0.71	3.44c	1.03
GPA	3.09a	0.46	3.03a,b	0.50	2.99a,b	0.50	2.93b	0.48

Note. Means having the same subscript are not significantly different. GPA = grade point average.

Regarding GPA, the well-adjusted participants had the highest scores, followed by the at-risk and ambivalent participants, and the distressed students had the lowest scores. The only statistically significant difference in GPA, however, was between the well-adjusted and distressed groups. This difference had a moderate effect size of 0.36.

Discussion

The current study investigated a dual-factor model of mental health that includes indicators of both positive well-being and psychological symptoms. Building on prior research (Eklund et al., 2011), results of the current study provide additional evidence for this dual-factor model in college students. Four distinct groups were identified based on varying levels of subjective well-being and psychological symptoms. The individuals who had high subjective well-being and low levels of symptoms were considered to have the most favorable mental health profiles and were therefore classified as well-adjusted. A second group of participants, in contrast, was identified as distressed because they exhibited the opposite mental health pattern, with low subjective well-being and high levels of symptoms. These groups are consistent with the classifications that would be made using a traditional one-factor mental health model. However, two additional groups were identified also, which provided support for the dual-factor model by demonstrating that positive well-being and psychological symptoms are not opposite ends of a single continuum. The third group of participants, classified as at-risk, did not have significant symptoms, but their level of subjective well-being was low. Thus, the absence of symptoms was not sufficient for optimal mental health. Using a traditional approach, these individuals would be considered mentally healthy, but their low subjective well-being distinguished them from their well-adjusted peers. The percentage of individuals in this group is larger than that found in studies of children and adolescents, suggesting that college students may be especially likely to lack positive well-being, even if they do not have psychological symptoms. Finally, a small group was identified as ambivalent, with elevated symptoms but also high levels of subjective well-being, which demonstrated that the presence of significant psychological symptoms did not preclude the experience of life satisfaction and positive emotions. Overall, these findings are aligned with those of the prior work of Keyes (2005) that identified well-being and distress as two distinct mental health constructs.

Results of the current study also demonstrate the usefulness of a dual-factor model in college students as it relates to student engagement. Overall, the well-adjusted individuals tended to report the highest levels of student engagement and had the highest GPAs. This finding suggests that both the presence of positive well-being and the absence of psychopathological symptoms are important for facilitating academic success. The well-adjusted group had higher levels of all types of engagement than the participants in the at-risk group. Thus, even though none of these individuals had elevated symptoms, those who had high subjective well-being had better study habits, were more engaged with faculty and peers, had higher intrinsic motivation, and felt a greater sense of belonging in the university community than the individuals with low subjective well-being. Accordingly, positive well-being seems to be an important component of an optimal college experience. Participants in the ambivalent group scored in the middle range on most engagement dimensions but reported especially high levels of beyond-class engagement. This finding suggests that positive well-being may be a protective factor that helps promote a sense of social connectedness and involvement in the university, despite psychological symptoms. Compared with their distressed peers who also had symptoms but low subjective well-being, the ambivalent individuals were significantly higher in beyond-class engagement.

Group differences in GPA were in the expected direction; the well-adjusted individuals had the highest GPAs, the distressed students had the lowest, and the at-risk and ambivalent groups scored between the other two. The only statistically significant difference, however, was between the well-adjusted and distressed groups, suggesting that positive subjective well-being may have a stronger impact on students’ engagement than on their course grades. However, because students reported their cumulative GPAs, these scores reflected academic performance across several semesters for the students in higher class years. Accordingly, it is possible that current mental health status would not have as much impact on these cumulative scores as it would on students’ current grades, resulting in the mostly nonsignificant findings.

The current study results have important practical implications. First, these findings demonstrate that practitioners working with college students should include measures of positive well-being in addition to traditional evaluations of psychopathology. Brief assessments of positive well-being like those in this study could be used as screening tools to identify students who have well-being deficits. This screening may be administered to students who are seeking counseling services or given to all students as part of a university-wide program. The results of these assessments would inform treatment planning by identifying students who may benefit from interventions targeting positive well-being in addition to treatment of clinical symptoms. For example, research indicates that interventions such as regularly recording positive life experiences, using character strengths in new ways, and expressing gratitude to others have lasting positive benefits, including increased happiness and decreased depression (Seligman, Steen, Park, & Peterson, 2005). Based on the current results, practitioners should consider utilizing such interventions not only to promote overall mental health but also to facilitate students’ academic success. Furthermore, because of the link between mental health and student engagement, academic support centers may also wish to monitor positive mental health and implement workshops targeting subjective well-being to promote optimal educational functioning. Finally, repeated administration of positive well-being assessments over time would allow practitioners to monitor students’ progress and potential improvements in mental health.

There were several limitations in the present study. First, data were collected from one sample at a single university, so the generalizability of the results is uncertain. The current study also relied entirely on self-report measures of both mental health variables and educational outcomes and did not incorporate any peer-reports or direct observations of behavior. Furthermore, to assess mental health efficiently, measurement of subjective well-being was limited to one affect scale and one life satisfaction scale, whereas measurement of symptoms was limited to one internalizing and one externalizing scale. Different results may be obtained with more comprehensive assessment of positive well-being and psychological symptoms. The use of students’ cumulative GPA is also a limitation because it does not accurately reflect current academic performance. Finally, the cross-sectional design of the study limits conclusions about changes in mental health over time or the long-term impact of mental health on educational performance.

Future research is needed to extend the current findings. Longitudinal studies can help determine the lasting effects of mental health status and identify the directionality of the relationship between mental health and academic outcomes. Additional studies may also incorporate other positive mental health factors, such as self-acceptance and optimism, into a dual-factor model. Finally, further research may yield a greater understanding of the at-risk and ambivalent individuals who would not be distinguished using a one-factor approach. For example, studies could examine thought patterns, coping skills, or social resources that may contribute to their unique mental health profiles.

Although more research is needed to further investigate a dual-factor mental health model, preliminary results support the applicability and utility of this approach in a college population. Building on the work of Eklund et al. (2011), this study provides further evidence that consideration of both positive well-being and psychological distress is important for understanding the mental health of college students. Examination of both of these factors allowed for a more comprehensive characterization of individuals’ functioning. Thus, assessment of both positive well-being and psychopathology is essential to understand students’ mental health and facilitate optimal academic success.