167 The Professional Counselor | Volume 13, Issue 3 Data Analysis We first explored descriptive statistics of demographic variables to understand the sample characteristics, including gender, race, age, employment status, work setting, specialty, and years of experience. Missing values were imputed using predictive mean matching, which estimates missing values by matching to the observed values in the sample (Rubin, 1986). To examine relationships among the five dimensions, we conducted a correlation analysis. After identifying the correlation matrix, which confirmed significant relationships among the five dimensions, we tested the hypothesized process model with a serial mediation using a path analysis. The ratio of response-to-parameter was 18:1 for the data, which satisfies the minimum amount of data for conducting a path analysis (Kline, 2015). Because the variables did not follow normality according to the Shapiro-Wilk test, skewness and kurtosis, and plots, we used weighted least squares. We started with a saturated model in which all possible direct paths were identified (Figure 1). Using the trimming method introduced by Meyers and colleagues (2013), we successively removed the least statistically significant path from the previous model until we found a model with all significant paths. To assess the model goodness of fit, model fit indices, including root mean square error residual (RMSEA), standardized root mean square residual (SRMR), comparative fit index (CFI), and Tucker-Lewis index (TLI) were evaluated. The RMSEA and SRMR values ≤ .06 indicate excellent fit, and CFI and TLI values ≥ .95 indicate excellent fit. All analyses were conducted using R Statistical Software (R Core Team, 2022). Results Pairwise Correlation Analysis Table 1 depicts the results of the pairwise correlation analysis. Significant positive correlations were found among the five dimensions of burnout. The relationship between Deterioration in Personal Life and Exhaustion was the largest, while that between Deterioration in Personal Life and Devaluing Client was the smallest. Also notable was that Devaluing Client, which is the dependent variable in the serial mediation model, displayed weak relationships with Negative Work Environment, Deterioration in Personal Life, and Exhaustion, but a moderate relationship with Incompetence. Table 1 Descriptive Statistics and Correlation Matrix Variable n M SD 1 2 3 4 5 1. Negative Work Environment 359 9.91 3.75 — 2. Deterioration in Personal Life 359 9.46 3.52 .35** — 3. Exhaustion 359 11.84 3.54 .43** .58** — 4. Incompetence 359 8.67 2.56 .23** .35** .33** — 5. Devaluing Client 359 5.52 1.94 .23** .22** .27** .40** — **p < .01
RkJQdWJsaXNoZXIy NDU5MTM1