**Expectations** *Describe the outcomes you expect/suspect/guesstimate for this study and explain why. Estimate the effect sizes you expect to find, for both the mean reaction time difference and the ex-Gaussian fitted RTV.* Consistent with the **Official Protocol-49TQVF** and with previous findings from similar relevant experiments, including Sripada et al. (2014) and Hagger et al. (2010), we expect the following: - **Analysis of RTV during incongruent trials of the MSIT:** Participants assigned to the difficult Letter “e” task (i.e., ego-depletion, experimental group) should show MORE reaction time variability than participants assigned to the easy letter “e” task (i.e., no-depletion, control group). - **Analysis of mean RT from the MSIT:** Participants assigned to the ego-depletion condition (i.e., experimental) should show LONGER reaction times than participants assigned to the no-depletion (i.e., control) condition. Although the differences between groups described above are anticipated to be similar to those found in previous similar experiments, it is expected that with *n* ≈ 50 per condition the likelihood of detecting statistically significant differences (if they exist) may be smaller than if the sample size were larger. However, even studies with relatively small sample sizes have demonstrated significant differences in RTV and RT (although not always statistically significant differences) between ego-depletion and no-depletion groups (e.g., Sripada et al., 2014). To further elaborate, according to the Official Protocol-49TQVF, with a sample size of *N* = 168 (*n* =84 in each of the two conditions), setting alpha at .01 and beta at 0.95, the overall expected effect size is *d* = 0.62. - **Analysis of RTV during incongruent trials of the MSIT:** Note that Sripada et al. (2014) found significantly greater RTV for the ego-depletion group than for the no-depletion group with an effect size of *d* = 0.69, sample size of *N* = 47, and *p* = .02. Using these data, G*Power 3.1.9.2 yields a post-hoc achieved power estimate of 0.49. Given that our lab will be using the same task/ methodology, if we assume the same error probability and power, the expected effect size would be *d* = 0.47. However, if we set the effect size to be the same as that in Sripada et al. (2014), *d* = 0.69, as well as *p* = .02, and set beta at 0.90, then the necessary estimated sample size per condition is *n* = 48 per group. - **Analysis of mean RT from the MSIT:** The average corrected effect size of studies incorporating the Stroop colour-naming task as the ego-depletion task has been estimated to be *d* = 0.77 (Hagger et al., 2010), with the recognition that the size of the effect varies according to the manner in which the Stroop task is modified. However, not all studies have incorporated a measure of RT in analyzing performance on the Stroop task. Webb and Sheeran (2003) measured mean RT, and found the depleted group performed significantly slower (*M* = 13.91; *SD* = 1.23) than did the non-depleted/ control group (*M* = 10.88; *SD* = 1.29) with *n*s = 28 and 29 in each of the groups and *p* < .001. Based on the data reported, the estimated effect size is quite large, *d* = 2.40 (not corrected). Sripada et al. (2014) reported change in RT across task completion and found no statistically significant change in RT in the placebo group regardless of whether they had been depleted or not. However, they did find that RT became slower over time. Therefore, it is challenging to predict the effect size for this comparison. Nevertheless, if we assume the effect size will be *d* = 0.77, consistent with the Hagger et al. (2010) meta-analysis, and set alpha at .01 and beta at 0.90 (as above), the necessary sample size per condition is *n* = 46 (according to G*Power 3.1.9.2).