a trend towards significancep = 0.141
Here we found that the generation effect was significantly higher than the aha effect in solving old items during testing [ t (19) = 4.40, p < 0.001, Cohen’s d = 1.457], whereas—although descriptively double the size—only showing a trend towards significance in recognizing old items [ t (19) = 1.54, p = 0.141, Cohen’s d = 0.442].