How were hypotheses to be tested chosen? For each study we have selected one statistically significant within-subject or between-subjects treatment effect or interaction effect for replication, using the following criteria:
Randomly pick a MTurk experiment reporting a statistically significant (p < 0.05) treatment effect or an interaction with a treatment effect for papers reporting more than one MTurk experiment.
Include the most central statistically significant result of this study/experiment (as highlighted in the original study).
If more than one equally central result was reported, randomly pick one of the results for replication.
If an original study included more treatments than the treatments selected for replication, we focus on replicating the treatments used for the result selected for replication.
We will use the same statistical test in the replication as in the original study.
studies to replicate
The table below summarizes the 41 studies selected to be included in the #MTRP (based on the inclusion criteria sketched above). Note that the table columns are sortable: clicking on a column header once will sort the table (based on the particular column's values) in ascending order; clicking on the header twice will sort the table in descending order. Hovering over the -symbol in the "Hyp." column, will show a pop-up window containing the paper's title and the hypothesis to replciate and bet on. The -symobl in the right-most column will direct you to the replication report (in PDF format), outlining all relevant details.