Redo Plots that show switch rates, high variance chosen etc. over trials with cumulative points on x -axis

[mdsteiner]

To have a better measure of when participants were closer to the goal, don't use trials as measure but the cumulative points. Maybe use bins of 5 points or so to make sure to have enough datapoints to get a meaningful mean value.

[mdsteiner]

I address this in 05_issue11.R.

Here are some plots about this. I created bins to be able to aggregate data somehow over the point values. The bins are always on the x axes. Results don't differ too much from the analyses over trials. Here's a summary of what the plots show:

• Most of the time, predictions of RSF are the least accurate.
• Proportion of high variance chosen don't differ much between no goal and goal but they do between the environments. The curves over bins are also different for the environments.
• Proportion of switches tend to be highest for LN and HG conditions and decreasing over bins.
• Response times decrease over bins. Those for the high EV is high variance conditions (HN, HG) are the highest.

Predictions of the different models over the bins and separated for the conditions

Proportion of high variance chosen over bins and conditions

Proportion of switches over bins and conditions

Median response times over bins and conditions