Interpretation der Effektstärke
Verfasst: Mi 19. Feb 2020, 00:24Hallo,
in einer Studie sind Effektstärken als Eta-Quadrat angegeben. Die Effekte 0,40 und sind dort als mittlere Effekte interpretiert worden. Ich verstehe leider nicht weshalb, da die Werte >0,14 sind, müssten es doch sehr große Effekte sein!?
Danke schon einmal und viele Grüße
Results
Inspection of the data indicates that the mean pretest scores on the SSRS, CDI, and ChIA are similar for both groups in each treatment condition. This was substantiated by independent samples t-tests, thus, the two groups in each condition were combined for data analysis. At post-treatment, both conditions yielded noticeable improvement on the SSRS, CDI and ChIA (see Table 1). Paired samples t-tests were used to compare students in the AM/REBT group at time 1 to students in the AM/REBT group at time 2; similarly, the scores of students in the AM group were compared at time 1 to their scores at time 2. The Bonferroni correction was applied to manage the error due to multiple comparisons and effect sizes were computed. These data analysis procedures were chosen because some assignments to groups were made based on convenience; moreover, a no-treatment control group was not used, which increased the appeal of the interventions in a school-setting. No significant differences were observed on any scale for the AM (only) group, once the Bonferroni correction was applied. Significant differences were observed on the SSRS (t =-4.42; p\.008) and the ChIA (t = 10.02; p\.008) for the combined AM/REBT intervention. Effect sizes (Cohen 1990; Schmidt 1996), however, were medium, for the CDI (η² = .40) and SSRS (η² = .68), and large for the ChIA (η² = .80).
in einer Studie sind Effektstärken als Eta-Quadrat angegeben. Die Effekte 0,40 und sind dort als mittlere Effekte interpretiert worden. Ich verstehe leider nicht weshalb, da die Werte >0,14 sind, müssten es doch sehr große Effekte sein!?
Danke schon einmal und viele Grüße
Results
Inspection of the data indicates that the mean pretest scores on the SSRS, CDI, and ChIA are similar for both groups in each treatment condition. This was substantiated by independent samples t-tests, thus, the two groups in each condition were combined for data analysis. At post-treatment, both conditions yielded noticeable improvement on the SSRS, CDI and ChIA (see Table 1). Paired samples t-tests were used to compare students in the AM/REBT group at time 1 to students in the AM/REBT group at time 2; similarly, the scores of students in the AM group were compared at time 1 to their scores at time 2. The Bonferroni correction was applied to manage the error due to multiple comparisons and effect sizes were computed. These data analysis procedures were chosen because some assignments to groups were made based on convenience; moreover, a no-treatment control group was not used, which increased the appeal of the interventions in a school-setting. No significant differences were observed on any scale for the AM (only) group, once the Bonferroni correction was applied. Significant differences were observed on the SSRS (t =-4.42; p\.008) and the ChIA (t = 10.02; p\.008) for the combined AM/REBT intervention. Effect sizes (Cohen 1990; Schmidt 1996), however, were medium, for the CDI (η² = .40) and SSRS (η² = .68), and large for the ChIA (η² = .80).