Skip to main content
. 2018 Jan 17;8:974. doi: 10.3389/fphar.2017.00974

Table 1.

Repeated measures ANOVA; OBN and DED predict changes in QIDS-SR over different time points up to 5 weeks.

Within-subjects effects F(3, 48) p Partial η2
Time * OBN Sphericity Assumed 5.563 0.002 0.258
Time * DED Sphericity Assumed 15.39 0.003 0.252
Within-subjects contrasts F(1, 16) p Partial η2
Time * OBN 1 Day vs. Baseline 13.143 0.002 0.451
1 Week vs. Baseline 7.237 0.016 0.311
5 Weeks vs. Baseline 13.29 0.002 0.454
Time * DED 1 Day vs. Baseline 9.941 0.006 0.383
1 Week vs. Baseline 6.828 0.019 0.299
5 Weeks vs. Baseline 15.298 0.001 0.489

Time is the within-subject factor (independent variable), QIDS-SR is the within-subject dependent variable in baseline, 1-day, 1-week, 5-weeks. OBN and DED are independent variables. Time-by-OBN and Time-by-DED interactions are the primary outcomes of interest (within-subjects effects). Within-subjects simple contrasts are calculated for each time point compared to baseline. A significant within-subject contrast suggests a linear relationship between OBN or DED and the difference between time points in QIDS-SR. Sphericity assumed: Mauchly's W = 0.71 (p = 0.411). partial η2 > 0.25 = large effect size. To inform directionality see Figure 1 for Pearson's correlations: high OBN and low DED predict reductions in QIDS-SR.