search

ndphillips / EE-Goals

0 stars 0 forks source link

redo the analyses with the proportion of high var chosen predicted by when it's rational to choose high var. #15

Closed mdsteiner closed 7 years ago

mdsteiner commented 7 years ago

I've found some errors with these analyses where I didn't pay enough attention to the within - between subject issue. So there might be some changes.

mdsteiner commented 7 years ago

I address this in 06_issue15.R

So using the actual (objective) distributions to calculate whether it is rational to choose the high variance option or not, there is no significant effect on the participant level (t(22) = 1.77, p = .1, d = 0.36). Also on the game level there's no significant effect (I used a mixed effects model with participant and game as random effect for this, see script for details). grafik grafik

By using the subjective distributions to calculate whether it is rational to choose the high variance option I found a pretty strong effect on the participant level, t(21) = 4.07, p < .001, d = 0.87. Also in the mixed effects model whether it was rational to choose the high variance option or not was a signigicant predictor b = .19, t = 4.78.

grafik

grafik