The Professional Counselor | Volume 13, Issue 2 86 In the second data analysis, we used the Pearson correlation to assess an association between data sets from the AFA and Composite Weight Bias. Composite Weight Bias was calculated by summing the scales of the revised PWS. Before computing a correlation, we examined the scatter plot between the independent and dependent variables to check the linearity between the two variables and the existence of outliers. We found less than three outliers on all four graphs and identified negative linearity. Using G*Power (Faul et al., 2007), we obtained the estimated sample size necessary to run the correlation analysis, which was 64 with a medium effect size of .3, an alpha level of .05, and power of .8. Lastly, we explored the relationships between demographics and weight bias toward counselors using a one-way ANOVA or an independent t-test. We selected statistical methods based on the number of categories of each demographic variable. See Table 2 for descriptive statistics of the outcome variables. Results Areas of Trust, Advice, and Selection We found significantly different levels of trust, advice following, and counselor selection behaviors among participants assigned to hypothetical counselors of different weights. Welch ANOVA test results indicated a statistical significance in all three areas between groups, F(2, 120.60) = 12.89, p < .001 with a medium effect size (ꞃ² = .11). Post hoc comparisons using Games-Howell showed the following results at the significance level of α = .05. Counselor Trust for average-weight counselors (M = 47.48, SD = 7.75) was significantly higher than Counselor Trust for underweight counselors (M = 41.15, SD = 10.72) at p = .001. Counselor Trust of overweight counselors (M = 49.70, SD = 9.89) was also significantly higher than Counselor Trust for underweight counselors at p < .001. There was no statistical significance for Counselor Trust between average and overweight counselors. Advice Following for average-weight counselors (M = 31.93, SD = 4.65) was significantly higher than Advice Following for underweight counselors (M = 24.72, SD = 7.28) and Advice Following for overweight counselors (M = 25.75, SD = 8.72) at p < .001 for both. Finally, Counselor Selection for an overweight counselor (M = 32.41, SD = 5.73) was statistically higher than Counselor Selection for an underweight counselor (M = 30.00, SD = 5.59) with p = .028. There was no statistical significance in the Counselor Selection of average-weight counselors (M = 31.41, SD = 4.91) compared to overweight or underweight counselors. See Table 3 and Figures 1–3. Next, we conducted a Welch ANOVA between overall composite scores and the three weight groups (see Table 4). Again, Welch test results indicated a statistical significance between groups, F(2, 118.73) = 11.71, p < .001 with a medium effect size ( ² = .10). Post hoc comparisons using GamesHowell showed statistical significance between overweight counselors (M = 107.87, SD = 21.68) and underweight counselors (M = 95.88, SD = 20.00). Underweight counselors were also significantly lower on the overall composite than average-weight counselors (M = 110.83, SD = 19.97). There was no statistical significance between overweight and average-weight counselors for their overall composite scores, which include all three variables.
RkJQdWJsaXNoZXIy NDU5MTM1