This section describes how I implemented the procedures stipulated by the Replicability Project. The general guidelines are available on the main project website [here][1].
**Effect of interest**
The main effect of interest was identified as the chi-squared test reported in Study 1 and cited below:
"Only 28% of subjects in the achievement condition said the behavior was intentional, but a full 76% of
subjects in the immoral condition thought the behavior was intentional. This contrast is statistically significant, chi^2 (1, N=37) = 7.7, p < .01."
**Power analysis**
A power analysis was conducted using R, alpha of .05 and power of .95.
*Safeguard power analysis* [(Perugini, Gallucci & Constantini, 2014)][3] was employed, which uses the uncertainty in published effect size estimates to approximate the population effect size.
This approach is particularly useful in avoiding underpowered replication of low-N studies with wide confidence intervals surrounding effect size.
See code and output [here][2]. This power analysis yielded a target sample size of 72 participants per cell.
**Materials**
The Qualtrics (*.qsf) file is available [here][4].
[1]: https://sites.google.com/site/thexphireplicabilityproject/procedure
[2]: https://www.dropbox.com/s/4zhbkcldaufl3ct/power%20analysis.html?dl=0
[3]: http://pps.sagepub.com/content/9/3/319.abstract
[4]: https://www.dropbox.com/s/f1qjj8dnueowh44/knobe_2003philpsych_replication_ivar.qsf?dl=0