Statistical Analysis

The current study aimed to (1) assess whether a bimodal therapy for chronic subjective tinnitus consisting of the combination of TRT and EMDR results in a clinically significant different efficacy in comparison with the prevailing bimodal TRT and CBT therapy, and (2) evaluate the effect of the bimodal treatment on the primary (i.e., TFI) and secondary outcome measurements.

Data were analyzed using SPSS statistical software version 25 (SPSS Inc., Chicago, IL, United States). There were three repeated measurements for each individual. To test the first research question, a linear mixed model was constructed, with the treatment group as fixed effect and the interaction between treatment group and time added to the model. The second research question was investigated using a similar linear mixed model, which included time as fixed effect and a random intercept, accounting for the non-independence between the observations taken from the same individual. If the effect of time was significant, a post hoc analysis with Bonferroni correction for multiple testing was conducted.

The analyses were carried out using linear mixed model analysis, which is robust against data missing at random. To assess the goodness of fit for the mixed models, we carried out a Levene’s Test for equality of variances. In addition, homoscedasticity was visually inspected using boxplots.