TPC Journal Volume 11 Issue 2

The Professional Counselor | Volume 11, Issue 2 255 multiple imputation (MI), a method to allocate missing data without causing inflated bias even when there is a large portion of missingness in the data (Osborn, 2013). Next, the assumptions of normality, linearity, homoscedasticity, and multicollinearity were tested. The residuals were linear and did not deviate from normality as evidenced by the residuals lying reasonably in a straight, diagonal line. The assumption of homoscedasticity was also supported, as most of the residuals were concentrated along the zero point. All variance inflation factor (VIF) values were less than 10 and tolerance values were greater than .1, indicating absence of multicollinearity (Tabachnick & Fidell, 2019). Therefore, the data were deemed appropriate for hierarchical regression and path analysis (Tabachnick & Fidell, 2019). Analytic Strategy Hierarchical regression models of stress-related growth were employed using SPSS version 27. First, gender, age, education status, sexual identity, and help-seeking experience were entered in Model 1 as the control variables. In Model 2, the first independent variable of subtle and blatant racism was added. In Model 3, the second independent variable of ethnic identity was entered. Finally, the remaining two independent variables of resilience and coping strategy were added as key predictors that may function as potential moderators in Model 4. To examine potential moderating roles of resilience and coping strategy in the relationship between racism and stress-related growth, Hayes’ (2018) PROCESS macro version 3.5 was conducted. Specifically, 10,000 bootstrapping resampling was conducted to produce 95% percentile confidence intervals (CIs) for the moderating effect. If the CIs excluded zero, moderating effect was considered to be significant. Furthermore, the moderating effects were examined utilizing three conditional values of moderators (Hayes, 2018; Preacher et al., 2017), which included low (the mean score of the moderator − 1 SD ), moderate (the mean score), and high values (the mean score of the moderator +1 SD ). Bodner’s (2017) formula was used to calculate effect size across moderator values. All predictors and moderators were mean-centered for more meaningful interpretation of moderating effect (Hayes, 2018). Results Preliminary Analyses Descriptive characteristics are found in Table 1. Male and female participants reported similar mean scores on all measurements, except the SABRA-A 2 . Female participants reported experiencing significantly higher levels of racism ( M = 29.10, SD = 6.25) than their male counterparts ( M = 26.75, SD = 7.59), with a small effect size ( d = 0.34; Cohen, 1998). Participants who had sought mental health services since COVID-19 reported significantly higher resilience scores ( M = 138.78, SD = 20.59), experiences of subtle and blatant racism ( M = 29.99, SD = 6.38), coping strategy ( M = 84.34, SD = 12.61), and stress-related growth ( M = 81.13, SD = 14.25) than participants who either did not seek professional mental health services or who considered seeking services, but had not used them. Correlations Correlational analyses among all study variables were conducted. Table 1 presents the correlations among the predictive and outcome variables assessed in the study as well as the mean and standard deviations for each variable and internal reliability for each measurement. As expected, ethnic identity, resilience, coping strategy, and stress-related growth were positively and moderately correlated with each other. Interestingly, subtle and blatant racism were also positively related to ethic identity, resilience, coping, and stress-related growth.