College Students' Reported Financial Stress and Its Relationship to Psychological Distress

Student loan debt has reached $1.3 trillion, surpassing credit card debt and second only to mortgages among debt that Americans hold (Federal Reserve Bank of New York, 2016). The Federal Reserve reported that college loans were the only form of debt that continued to rise during the Great Recession (Lee & Mueller, 2014), with two thirds of graduates reporting debt at an average of $29,400 (Erwin & Wood, 2014). As debt continues to rise, increased attention is being directed to the impact of financial anxiety and high relative debt (debt‐to‐asset ratio) on physical and mental health. Reports have associated financial distress with higher general anxiety, depression, sleep difficulties, self‐reported stress, and decreased physical health (Cooke, Barkham, Audi, Bradley, & Davy, 2004; Sweet, Nandi, Adam, & McDade, 2013).

For university and college counseling centers, the impact of financial strain on students' mental health is not negligible. 2015) indicated that 25.3% of students seeking mental health services described their financial situation growing up as either “often” or “always” stressful. From the same respondents, 37.9% of students rated their current financial situation as often or always stressful. Gault, Reichlin Cruse, and Román (2014) suggested that this increased financial strain can have an adverse impact on students, decreasing their time for schoolwork, self‐care, sleep, and integration into the campus environment, and having a generally negative impact on their well‐being. Overall, these factors can decrease students' ability to complete degrees (Chen & DesJardins, 2010).

The majority of research on financial stress is community‐based, from sociological or financial perspectives, and with non‐U.S. populations. Therefore, the present study is unique in examining the relationship of reported financial stress with psychological distress among students in the United States, using psychological measures commonly used in college counseling centers to examine the relationship between students' reported financial stress and psychological distress. Multiple factors influence financial stress. First, we will explore research that examines contributions to financial stress.

Financial Stress and Psychological Distress

Financial debt and the accompanying stress lead to adverse health and psychological outcomes. Studies have demonstrated that level of debt, such as the amount of debt associated with credit and other spending, has a positive association with anxiety (Drentea, 2000), and that among people who endorse financial distress, many report their health being negatively affected (O'Neill, Sorhaindo, Xiao, & Garman, 2005). Specifically, O'Neill et al. (2005) reported that their sample of 3,121 financially distressed consumers stated that stress, worry, nerves, anxiety, depression, insomnia, and sleep problems were among the most significant factors affecting health. Moreover, debt and financial strain can lead to a higher likelihood of developing a mental disorder (Meltzer, Bebbington, Brugha, Farrell, & Jenkins, 2012), such as depression and anxiety (Sweet et al., 2013). Sweet et al. (2013) observed that it may be an individual's attitude about debt, as opposed to the actual amount of debt, that predicts adverse health outcomes.

Concerning college students, Jessop, Herberts, and Solomon (2005) conducted a study on two cohorts of students. The first cohort was a group of British students who experienced the potential for greater college debt than was historically the case due to national policy changes. The second cohort was a group of students in Finland, where college tuition is free. Results showed that British students reported higher debt and greater financial concern than did the Finnish students, which resulted in worse health outcomes. Specifically, greater financial concern led to increased role limitation due to emotional problems; increased physical pain; and decreased social functioning, mental health, energy, and physical health. Likewise, in a longitudinal study of British undergraduate students, Richardson, Elliott, Roberts, and Jansen (2017) found that students' degree of financial stress predicted the severity of their anxiety 3 to 4 months later and alcohol dependence at 6 to 8 months. Similar to Sweet et al.'s (2013) findings, the authors suggested that students' stress level about their debt, rather than the amount of debt, predicted poorer outcomes.

Financial Stress and Academic Performance

From 2004 to 2014, average student debt at graduation increased from $18,550 to $28,950 (The Institute for College Access and Success [TICAS], 2015). In 2017, average student debt varied across states from $18,850 to $38,500, while the number of borrowers remained relatively stable (TICAS, 2018). Concurrently, the cost for tuition, fees, room, and board rose 34% at public and 25% at private institutions (Kena et al., 2016). Debt accrued while a student is pursuing a degree not only affects the borrower after graduation but can have a detrimental effect on productivity and academic performance while the student is enrolled in college (Pinto, Parente, & Palmer, 2001). Students working part‐time perform better academically than peers, but 74% of undergraduates work more than 20 hours per week, which is related to poorer academic outcomes (Dundes & Marx, 2006), including longer time to completion and decreased retention and graduation (Kuh, Kinzie, Buckley, Bridges, & Hayek, 2006).

Increased financial stress often co‐occurs with anxiety and depression and is predictive of lower engagement in college activities such as studying and seeking academic support, lower persistence, and reduced overall academic performance (Hunt, Boyd, Gast, Mitchell, & Wilson, 2012; Joo, Durband, & Grable, 2009; Ziskin, Fischer, Torres, Pellicciotti, & Player‐Sanders, 2014). Financial factors, including financial problems, amount of debt, lower availability of family financial resources, and increased need for employment, are particularly impactful with respect to attrition rates (Lombardi, Murray, & Gerdes, 2012).

Group Differences in Financial Stress

Although financial distress and debt affect many students, some populations are disproportionally affected, including female, low‐income, older, first‐generation, and racial minority students and students with children (Hunt et al., 2012; Joo et al., 2009). These students face more barriers to obtaining funding for college, but research also suggests differences in financial stress and college outcomes. For example, need‐ and gift‐based financial aid (e.g., grants, work‐study) relate to higher retention and graduation rates than loans for all students, with more pronounced positive effects for low‐income students and students of color (Chen & DesJardins, 2010).

Students of color are at increased risk of dropout in their first year attending college, influenced by family income, parental education, and amount of financial aid. Some low‐income students of color (e.g., Hispanic and Asian) are less likely to borrow for college, which can be detrimental to completion (Gault et al., 2014). For those who do graduate, Black and Hispanic graduates are more likely to accrue loans, and the total amount of loan debt is higher for Black graduates than others (Cataldi, Siegel, Shepherd, & Cooney, 2014). Although loans can lead to higher persistence and graduation rates, for some students this results in more substantial debt and default rates. For example, approximately 70% of Black students who borrow money do not complete their degree (Jackson & Reynolds, 2013).

For first‐generation students, lower parental income and education, unmet needs with room and board, and inability to pay tuition and fees can lead to dropout (Chen & DesJardins, 2010). First‐generation students are more likely to be low‐income students (Miller, Gault, & Thorman, 2011), and loans are the most common form of financial aid they receive (O'Brien & Shedd, 2001). First‐generation students, compared with continuing‐generation students, are apt to believe they cannot afford college without incurring debt and are therefore more likely to take out loans (Lee & Mueller, 2014).

Among women, there is an increased chance that they are balancing the financial demands of college and single parenting, particularly for women of color (Gault et al., 2014). These demands, as well as the likelihood of unmet need cited earlier, are further intensified by differing opportunities upon graduation. Compared with men, women are paid less on average, disproportionately major in fields that earn lower wages, and have higher debt upon graduating (Musu‐Gillette et al., 2016), with pronounced effects on women of color, resulting in differential return‐on‐investment of a college education.

Purpose of the Study

Financial stress can have deleterious psychological effects, lead to poorer health outcomes, and negatively affect academic performance. As noted earlier, research has found that financial stress is related to family income, race, gender, and first‐generation status. Therefore, the present study sought to identify how self‐report via intake data may predict psychological distress and may serve as a useful tool for college counselors. On the basis of the literature, we examined the following hypotheses:

Hypothesis 1: Group differences will exist on the level of financial stress based on gender, race/ethnicity, and generation status, with members in minority categories reporting higher financial distress.

Hypothesis 2: Financial stress is positively related to psychological concerns as measured by the Counseling Center Assessment of Psychological Symptoms–62 (CCAPS‐62; Locke et al., ).

Hypothesis 3: Financial stress is negatively related to grade point average (GPA).

Hypothesis 4: Financial stress will predict GPA as mediated by psychological distress.

Method

Participants and Procedure

The present research was designed to study how students' self‐reported level of financial stress is related to their reporting of mental health concerns (e.g., anxiety and depression). In total, the sample comprised 4,372 undergraduate and graduate students who attended an intake session for counseling at a midsize, midwestern, 4‐year university between 2010 and 2015. After we removed cases in which important data were missing (see details below), the final sample comprised 3,303 students. Students completed intake measures before the appointment. Deidentified data were exported from Titanium and examined using SPSS, Version 24.

A number of participants (n = 1,332) provided no response with respect to racial/ethnic identity. Among participants who did provide a response, a majority of the sample identified as European American or White (n = 1,386, 70.3%), followed by African American/Black (n = 282, 14.3%), Asian American/Asian (n = 82, 4.2%), multiracial (n = 71, 3.6%), Middle Eastern/North African (n = 54, 2.7%), Hispanic/Latino (n = 54, 2.7%), Native Hawaiian or Pacific Islander (n = 8, 0.4%), and American Indian/Alaska Native (n = 2, 0.1%). Twenty‐three participants (1.2%) indicated they preferred not to answer this item, and nine (0.5%) self‐identified. More than half identified as female (n = 1,888, 57.2%); 41.8% identified as male (n = 1,382); and the remainder identified as transgender, gender fluid, or gender queer (n = 21, 0.6%) or indicated a preference not to answer (n = 12, 0.4%). The age of the sample ranged from 18 to 62, with a mean age of 26.03 (SD = 6.78). The largest proportion of students identified as sophomores (n = 796, 24.1%), followed by juniors (n = 753, 22.8%), seniors (n = 621, 18.8%), freshman/first year (n = 582, 17.6%), and graduate‐ or professional‐degree‐seeking students (n = 477, 14.4%); the remainder indicated they were non‐degree‐seeking students (n = 74, 2.2%). The majority indicated continuing‐generation status (n = 2,358, 71.4%). (Percentages do not total 100 because of rounding.)

Measures

CCAPS‐62. The CCAPS‐62 (Locke et al., 2011) is a standardized measure of psychological symptoms, distress, and change in treatment developed specifically for college populations. It is a self‐report measure consisting of 62 items that independently load onto eight subscales, including Depression, Generalized Anxiety, Social Anxiety, Academic Distress, Eating Concerns, Family Distress, Hostility, and Substance Use. Items are rated on a 5‐point Likert‐type scale ranging from 0 (not at all like me) to 4 (extremely like me), with nine reverse‐scored items. Students are instructed to rate each item on the basis of the past 2 weeks. T scores and percentile ranks are provided for each subscale within the administrative manual. The CCAPS‐62 normative sample includes college students seeking counseling services at 97 participating institutions, consisting of 59,606 undergraduate and graduate students, ranging in age from 18 to 68 years (M = 22.74). Internal consistency (alpha) reliability coefficients range from .82 to .91.

Standardized Data Set (SDS). The SDS (Pennsylvania State University, 2012) is a set of questions completed before intake appointment that includes demographics, background, and mental health. The SDS is in use at college counseling centers throughout the United States. Items are updated over time, with input from participating centers, and can be edited by a center to meet the needs of their population. For this study, variables included in analyses were age, gender, race, GPA, and first‐generation college status. Two items assessing current and past financial stress and two items assessing family and social support were included in analyses, as well. Students are asked to rate these items on a 5‐point scale.

Preliminary Analysis

Initial examination of 4,372 total cases identified missing data on key variables, including GPA. These cases were subsequently removed. The final data comprised 3,303 responses. Data were examined and met requirements for normality, linearity, and homoscedasticity. Following guidelines for skewness and kurtosis, we screened residuals and residual plots, as well as bivariate scatterplots as outlined by Tabachnick and Fidell (2013). No univariate or multivariate outliers were detected. Next, multicollinearity was tested using regression in SPSS. Collinearity diagnostics were inspected across the planned variables for group differences. No violations were discovered.

We then conducted preliminary assumption testing. As is common in large data sets (Tabachnick & Fidell, 2007), we found our Box's test of equality of covariance was violated (Box's M = 92.801, p < .001), indicating our data set could violate the assumption of homogeneity of variance. Next, we examined Levene's test for equality of error variances; family distress, financial stress past, and family support had significance levels below .05, indicating they violated the assumption of equality of variances. In cases of unequal variances, we followed the recommendation for the transformation of the variables to correct for distribution. Upon inspection, the histogram revealed a peak of scores on the left tail of the distribution for both family distress and past financial stress and the right tail for family support. Transformation of the data distribution yielded the best results from square‐root transformations, but a visual inspection showed little improvement in distribution, and the variance tests were still violated. Given there was not an improvement from the untransformed distributions and that skewness may not make “substantiative difference” in samples greater than 200 (Tabachnick & Fidell, 2013, p. 80), the original distributions were retained. Instead, we followed recommendations to apply a strict alpha of .025 or .01 to the F test for univariate analysis (Tabachnick & Fidell, 2013). Although some violations were identified, recommendations were applied, and we proceeded to inspect results of the multivariate analyses of variance.

Results

Hypothesis 1 stated group differences will exist on the level of financial stress based on gender, race/ethnicity, and generation status, with members in minority categories reporting higher financial distress. We first investigated differences between first‐generation students and continuing‐generation students, using a one‐way between‐groups multivariate analysis of variance, on the following dependent variables: family distress (FD), academic distress (AD), depression (DEP), anxiety (ANX), financial stress now (FSN), financial stress past (FSP), and family support (FS). There was a significant difference between first‐generation and continuing‐generation students on the combined dependent variables, F(8, 3275) = 40.68, p < .001, Pillai's trace = .09, η2 = .09. In instances of multiple post hoc comparisons, Tabachnick and Fidell (2007) suggested using a Bonferroni adjustment, which consists of calculating a more stringent alpha value to avoid Type I error. The calculated Bonferroni adjusted alpha level (.0063) was applied to inspection of dependent variables. Significant group differences were found on FD, F(1, 3282) = 63.85, p < .001, η2 = .02; AD, F(1, 3282) = 14.78, p < .001, η2 = .004; distress, F(1, 3282) = 15.55, p < .001, η2 = .01; ANX, F(1, 3282) = 11.80, p < .001, η2 = .004; FSN, F(1, 3282) = 131.91, p < .001, η2 = .04; FSP, F(1, 3282) = 279.60, p < .001, η2 = .08; and FS, F(1, 3282) = 63.85, p < .001, η2 = .02. Among these, small effect sizes were found for FD (1.9%; M = 1.51, SD = 1.04 compared with M = 1.20, SD = 0.97; first generation to continuing generation, respectively), FS (1.1%; M = 2.23, SD = 1.38 vs. M = 2.53, SD = 1.32), and FSN (3.9%; M = 3.73, SD = 1.02 vs. M = 3.24, SD = 1.14). A medium effect size was found for FSP (7.9%; M = 3.33, SD = 1.30 vs. M = 2.53, SD = 1.22).

Next, we explored group differences on categories of gender (female and male). Other categories of identification—transgender, gender fluid, gender queer (n = 21, 0.6%) and prefer not to answer or not sure (n = 12, 0.4%)—constituted a small percentage of the sample. Therefore, we chose to not include them in our analyses on gender differences among our variables. Similar to what was described earlier, Box's test criteria were not met, and Levene's test was violated for FD and ANX, meaning a higher significance level was needed for these dependent variables. There was a significant difference between groups on gender on the combined dependent variables, F(13, 3256) = 38.77, p < .001, Pillai's trace = .13, η2 = .13. The suggested Bonferroni adjusted alpha level (.0063) was applied to inspection of dependent variables. Significant group differences were found on FD, F(2, 3300) = 27.78, p < .001, η2 = .02, female (M = 1.40, SD = 1.03) and male (M = 1.14, SD = 0.92). For DEP, female (M = 1.49, SD = 0.94) significantly differed from male (M = 1.36, SD = 0.96), F(2, 3300) = 8.86, p < .001, η2 = .01. Results showed for ANX, female (M = 1.55, SD = 0.97) was significantly different from male (M = 1.22, SD = 0.88), ANX, F(2, 3300) = 49.69, p < .001, η2 = .03. For distress, F(2, 3300) = 8.02, p < .001, η2 = .01, female (M = 1.50, SD = 0.85) compared with male (M = 1.40, SD = 0.86). Finally, both financial stress variables showed significant differences, FSN, F(2, 3300) = 7.74, p < .001, η2 = .01, female (M = 1.41, SD = 1.03) and male (M = 1.14, SD = .92); and FSP, F(2, 3300) = 5.33, p < .005, η2 = .003, female (M = 2.82, SD = 1.32) and male (M = 2.67, SD = 1.25). Among these, small effect sizes were found for FD (1.6%) and ANX (2.9%).

Finally, a one‐way multivariate analysis of variance was run for race across the dependent variables. Preliminary assumption testing was conducted, and no violations were noted. There was a statistically significant difference between groups (person of color, White, and no answer), across the dependent variables, F(16, 6584) = 3.84, p < .001, Wilks's Λ = .98, η2 = .01. With significant group differences in reporting of FD, F(2, 3300) = 8.30, p < .001, η2 = .01; FSN, F(2, 3300) = 5.05, p < .01, η2 = .003; FSP, F(2, 3300) = 15.43, p < .001, η2 = .01; and FS, F(2, 3300) = 3.08, p < .05, η2 = .002. Post hoc comparisons using Tukey's honestly significant difference test indicated that person of color respondents reported higher FD (M = 1.45, SD = 1.01), FSN (M = 3.49, SD = 1.14), and FSP (M = 3.03, SD = 1.32) than White (M = 1.27, SD = 0.99; M = 3.40, SD = 1.12; and M = 2.69, SD = 1.29, respectively) or no answer (M = 1.25, SD = 0.99; M = 3.32, SD = 1.13; and M = 2.70, SD = 1.27, respectively) respondents, and less FS (M = 2.32, SD = 1.36) than comparison groups (Whites, M = 2.48, SD = 1.32; and no answer, M = 2.46, SD = 1.35). The abovementioned results supported Hypothesis 1; all groups demonstrated differences on the financial stress variables.

Hypothesis 2 stated that financial stress would be positively related to psychological distress variables measured by the CCAPS‐62. Table 1 reports the correlations between the variables in the current study. Results supported Hypothesis 2; both FSP and FSN were positively related to CCAPS‐62 subscales, except substance use, which was not significantly related to FSP (r = .03).

Table 1

Intercorrelations Between the Major Study Variables

Note. FN = financial stress now; FP = financial stress past; FS = family support; SS = social support; EC = eating concerns; SU = substance use; AN = anxiety; HS = hostility; SA = social anxiety; FD = family distress; AD = academic distress; TD = total distress; DP = depression; GPA = grade point average (self‐reported).

*p < .05 (two‐tailed). **p < .01 (two‐tailed).

Hypothesis 3 stated that financial stress would be negatively related to GPA. Both FSP and FSN were significantly negatively related to GPA (r = –.09, p < .01, and r = –.16, p < .01, respectively). Thus, Hypothesis 3 was supported by our results.

Hypothesis 4, stating financial stress as predictive of CCAPS‐62 categories and mediated by psychological distress, was tested using path analysis in SPSS Amos (Arbuckle, 2016). We first inspected Pearson bivariate correlations among our variables (see Table 1). Given our large sample size, we determined variables with correlations of .30 as potential predictors of psychological distress. The initial model (see Figure 1) was run with the full sample and provided an excellent fit to the data χ2(8) = 22.002, p =.005; comparative fit index (CFI) = .998; goodness‐of‐fit index (GFI) = .998; root‐mean‐square error of approximation (RMSEA) = .023 (90% confidence interval [CI] [.012, .035], p close‐fit Ho = 1.00); Akaike's information criterion (AIC) = 78.002. Paths are labeled in Figure using standardized values. Given the group differences identified among first‐generation and continuing‐generation students on key variables, we tested models for each group. Across the three models (total sample, first generation, continuing generation), path coefficients were consistent regarding their relationships and predictive power. One notable exception was the paths from FSN to distress and from FS to AD, which were both nonsignificant in the first‐generation model. The two alternate models also demonstrated a good fit to the data, first generation, χ2(8) = 16.809, p =.032; CFI = .995; GFI = .996; RMSEA = .035 (90% CI [.010, .058], p close‐fit Ho = .85); AIC = 72.809; and continuing generation, χ2(8) = 28.402, p < .001; CFI = .995; GFI = .997; RMSEA = .033 (90% CI [.020, .046], p close‐fit Ho = .98); AIC = 84.402. The AIC offers a comparative measure of fit. The lower the AIC, the better fitting the model. Of the three aforementioned models, the first‐generation model demonstrated the lowest AIC (AIC = 72.809) and therefore can be inferred to be the best fitting model among the three.

Figure 1

Structural Path Coefficients for Model

Note. All coefficients are standardized. T = total sample; FG = first generation; CG = continuing generation; FSP = financial stress past; Family = family distress; FS = family support; FSN = financial stress now; Academic = academic distress; GPA = grade point average; SS = social support; Distress = total distress. *p < .05. **p < .01. All others (if not labeled NS), p < .001.

Among first‐generation students, FSP and FS accounted for 14% of the variance in FSN and, along with social support, 40% of the variance in family distress. Indirect and direct effects accounted for 11% of the variance in AD, 51% of the variance in distress, and 13% of the variance in GPA. The direct effect of FSN on GPA was β = –.09, SE = .035, p < .05 (90% CI [–.146, –.033]). The indirect effect was β = –.06, SE = .015, p < .01 (90% CI [–.086, –.037]), and a total effect of β = –.15, SE = .034, p < .001 (90% CI [–.207, –.097]).

Comparatively, among continuing‐generation students, FSP and FS accounted for 24% of the variance in FSN and, along with social support, 38% of the variance in family distress. Indirect and direct effects accounted for 10% of the variance in AD, 50% of the variance in distress, and 12% of the variance in GPA. For the continuing‐generation model, the total effect of FSN on GPA was β = –.14, SE = .021, p < .001 (90% CI [–.177, –.109]).

Finally, the total model accounted for 23% of the variance in FSN and, along with social support, 39% of the variance in FD. Indirect and direct effects accounted for 10% of the variance in AD, 51% of the variance in distress, and 12% of the variance in GPA. For the total model, the total effect of FSN on GPA was β = –.15, SE = .017, p < .001 (90% CI [–.183, –.126]). Thus, results across all three models supported Hypothesis 4.

Discussion

This study provides an exploratory analysis of the influence of perceived financial stress on psychological distress and academic performance. The results support hypotheses regarding group differences in reported financial stress, as well as self‐reported financial stress's relationship to psychological distress and GPA, and supply an exploratory model of interaction among the variables studied. Results align with outcomes from noncollege samples and thus link financial distress among college students to the larger body of findings on financial stress.

Results for first‐generation students showed that, compared with continuing‐generation students, they reported greater family distress, academic distress, total distress, and anxiety. Additionally, first‐generation students endorsed higher levels of current and past financial stress, as well as less family support. These results are consistent with previous findings that have shown that first‐generation students are more likely to have unmet needs, to have lower parental income (Chen & DesJardins, 2010), and to have low income themselves (Miller et al., 2011). Moreover, financial distress can cause increased stress and lower academic and campus engagement (Cooke et al., 2004).

Next, among the represented gender categories, women reported significantly different scores on the majority of dependent variables, with the most pronounced differences on levels of family distress and anxiety. For women, compared with men, these results may reflect the increased burdens of debt, and postgraduation debt, faced by women (Gault et al., 2014). Furthermore, anxiety may stem from concerns about paying back debt after college, as incomes for women are often less than those for male counterparts (Musu‐Gillette et al., 2016).

Among racial/ethnic groups, results showed students of color reporting greater family distress, current and past financial stress, and less family support. These results correspond with the literature that has shown students of color incur more significant amounts of debt (Hunt et al., 2012), and among students of color, Black students, in particular, are more likely to have more considerable debt in college (Jackson & Reynolds, 2013). Among students of color in our sample, those identifying as African American/Black were the largest population. Therefore, results among students of color are more likely to represent students from this racial/ethnic group.

The exploratory model (see Figure ) modeled three levels of relationships. First, we modeled exogenous variables (self‐reported past financial stress, social support, and family support) as predictors of current financial stress and psychological distress variables. As expected, past financial stress predicted both current financial stress and psychological distress. Next, family and social support both predicted academic distress and family distress. As well, social support was a strong predictor of total distress. Social support was a strong negative predictor of academic distress, suggesting that increasing social support would reduce academic distress. Results were similar for the relationship between social support and total distress. We posit within this model that low support in these domains is likely to generally exacerbate distress among the variables studied, ultimately having an impact on academic performance as measured by GPA.

Results suggest that as financial stress increases, academic and family stress also increase. Moreover, academic distress has a strong indirect influence on GPA. Across all three models, the combined variance explains approximately 12% of GPA with a total effect of FSN decreasing the GPA by 0.15 when financial stress increases by 1 point. These results support previous research that suggests financial stress has an impact on not only psychological health but academic performance as well (Pinto et al., 2001).

In summary, the results suggest an increased risk among students of color, women, and first‐generation students for financial stress. Furthermore, financial stress when combined with lower levels of family or social support can significantly contribute to reported family distress and overall distress (manifested in symptoms of anxiety and depression). Together, these can have an impact on academic performance. Lessening distress surrounding finances could lead to decreased distress and improved academic performance.

Implications for College Counselors

The results of this study have important implications for college counseling and university practitioners. We outline recommendations for the individual, group, and outreach practice that can aid in addressing financial stress related to psychological health. Furthermore, we propose suggestions for advocacy and policy‐based work at the university, state, and national levels.

Individual and group counseling. For most college counseling centers, the primary intervention modalities are individual and group counseling. Across all interventions, counselors should follow recommendations for multiculturally sensitive counseling approaches to assist counselors in increasing awareness and responsiveness to financial status and distress as important components of mental health distress. Discussing financial stress directly with students in individual and group modalities may assist with the building of a working alliance and the conceptualization and treatment planning process, allowing counselors to identify strategies to address financial stressors. It may also be a key component in the alleviation of mental health concerns, increased retention and graduation rates, and overall well‐being.

Although this study focused on self‐reported financial distress, rather than social class, counselors are also encouraged to consider social class identity and financial status as integral parts of identity using multiculturally sensitive practice (Hays, 2016). Liu (2011) developed two psychological models of social class, (a) the social class worldview model and (b) the social class and classism consciousness model. The first models the development of social class and classism consciousness, socialization messages, behaviors related to social class (e.g., lifestyle, material possessions), and experiences of upward, lateral, and downward classism. The model suggests that these factors contribute to the development of internalized classism, which can result in anxiety, depression, anger, and self‐harming behaviors. Theory suggests maintaining constancy of social class and minimizing internalized classism can prevent these adverse psychological effects. Liu's second model, the social class and classism consciousness model, functions similarly to other identity development models, exploring how an individual's sense of social class position and awareness of social class develop, eventually culminating in a full understanding of the impact of social class and classism. As Liu suggested the use of self‐reported social position is more predictive of psychological variables, such as well‐being, than traditionally used social class variables (i.e., income, education, and occupation), financial distress could be indicative of one's subjective awareness of class, access to capital, and experiences with classism leading to internalized classism. College counselors are strongly encouraged to familiarize themselves with Liu's models and incorporate these concepts and interventions into practice.

Outreach, consultation, and advocacy. Counselors on university campuses are uniquely positioned to engage in prevention and consultative interventions outside the counseling office. They are thereby expanding the focus of outreach and consultation to incorporate financial issues. Forming strong relationships between counseling centers and student service offices, such as financial aid and student employment, can help counselors to connect students with information and resources to improve financial means and decision‐making. Counselors will also be better equipped to advocate for students in need (Hunt et al., 2012). Consultative relationships with financial service offices on campuses can contribute to increased financial literacy for counseling staff and a better understanding of the connection between mental health and financial distress for financial service staff; such relationships can thus facilitate assisting students and their families in obtaining access to and navigating services, supporting students' long‐term financial literacy, and decreasing financial avoidance behaviors (Jackson & Reynolds, 2013; TICAS, 2015; Ziskin et al., 2014). Such partnerships could also lead to developing financial literacy outreach, such as presentations, information tables, and online educational resources.

As part of the institutional system, we encourage counselors to consider the impact of financial policies and practices on students. College counseling centers are often tasked with goals regarding retention and graduation rates and can advocate for changes that can assist students in gaining access to and completing education. Counseling center staff should seek involvement in university committees that contribute to the development and implementation of financial policies, as well as involvement in advocacy at state and federal levels for public policy changes.

Limitations and Future Research

There are several significant limitations to consider when interpreting the results of this study. First, our results were derived from one university counseling center and therefore are limited in their generalizability to other counseling centers and college campuses. Because many university and college counseling centers collect similar data, we encourage multisite collaboration to determine whether these results are consistent across sites and populations. Next, we conducted secondary data analysis on information derived from intakes at a counseling center. Therefore, our study lacked a comprehensive measurement of financial stress. Future research should consider multi‐item measures of financial stress, debt, and other financial and class‐related variables that would inform financial stress. Such research could more directly assess the causal links proposed in the current study and additionally provide opportunities to study other supplementary factors that influence the relationship between financial stress and psychological distress.

As our results were mostly exploratory, using defined measures of financial stress and financial anxiety with formulated counseling interventions would assist in increasing our understanding of how financial stress contributes to college students' psychological concerns. Using the models cited earlier, as well as outreach and collaborative interventions, could offer programmatic and quasi‐experimental research that could benefit psychological outcomes among college students at risk for financial hardship, debt mismanagement, and financial stress.