STATISTICAL AND POWER ANALYSIS

Statistical analyses computed for internal consistency relies on the calculation of Omega (ω) with 95% confidence intervals (CIs) with the MBESS package in R studio (Dunn, Baguley, & Brunsden, 2014; Kelley, 2018; Kelley & Pornprasertmanit, 2016). Omega was computed instead of the Cronbach alpha because Omega makes the assumption of multidimensionality nature of the scale which is more realistic given that only a few scales seem to be characterized by unidimentionality (Béland, Cousineau, & Loye, 2018; Socan, 2000). Omega is also associated with a lower risk of overestimation or underestimation of reliability in comparison to alpha (Dunn et al., 2014; Kelley & Pornprasertmanit, 2016).

Statistical analyses computed for test-retest reliability depend on the calculation of Pearson correlation coefficients. A correlation of between .10 and .30 corresponds to a small effect, between .30 and .50 to a medium effect and above .50 to a large effect (Cohen, 1988). To measure the responses convergence and equivalence of the two translated versions, Pearson coefficients were first computed followed by equivalence paired t-tests based on the two one-sided hypothesis tests with the TOSTER package in JASP (Lakens, 2017; Team, 2020). T-tests on mood scores at both evaluation times were computed to control for the effect of mood on test-retest analysis.

Descriptive analyses and correlational analyses were performed using STATISTICA 10 software (TIBCO Software Inc., 2017) with a significant criterion of .05. The de-identified data can be downloaded on the Open Science Framework on the following link: https://osf.io/8ct7f/. The required sample size for the Haccoun approach was a priori determined based on the recommendations of at least 5 subjects per item of the instrument (Nunnally & Bernstein, 1999).