**Data stopping rule**
We will aim to collect usable data that meet all criteria for inclusion in the confirmatory tests from 120 participants (60 couples - see power analysis below). We will check some variables of the data during data collection to ensure an accurate count of eligible participants, however, to prevent 'data peeking' concerning the confirmatory tests we will NOT merge the temperature data at any point prior to the full data analysis (thus, making it impossible for us to run the primary analyses and allow that to affect our judgment)
We will schedule sessions such that we end right around 118 participants. However, if we hit 118 earlier than expected any sessions already scheduled will still be administered.
**Power analysis**
We aimed to provide the focal hypothesis test (difference in temperature in response to seeing sad vs neutral partner’s face) with a 95% probability to detect an effect equal to 1/20°C. For the expected variability, that roughly translates to a standardized mean difference of d = .07. To compute the required sample size, we simulated the data taking the Wagemans & IJzerman study estimates as parameter values. After defining the smallest effect size of interest (1/20°C) as the target parameter value, we simulated new values for the dependent variable (temperature) and back-fitted the mixed-effects model in 1000 iterations employing the simr package (Green & MacLeod, 2016). The simulations showed that under the given design (controlling for duration of the trial and communal strength; including a by-participant random intercept), we would need 118 participants for the likelihood ratio test having a 95% probability of detecting the smallest effect size of interest if H0 is false. The required sample size for this hypothesis represented our pre-defined stopping rule, except we rounded up to 120 (60 couples) for simplicity.
Script available in this component and as .rmarkdown here: http://rpubs.com/ivanropovik/427553
Green, P., & MacLeod, C. J. (2016). SIMR: an R package for power analysis of generalized linear mixed models by simulation. Methods in Ecology and Evolution, 7(4), 493–498. doi:10.1111/2041-210x.12504
