checking for significantly below-chance performance!

[mcfrank]

We can use a simple binomial test (conservative as many of our tasks are 4AFC) to find out whether participants are below chance. For example, getting 3 / 12 trials is p = .07 - 2/12 is p = .01. I think one of those might be good to flag. This kind of checking will help us find kids for whom there are bugs... @mikabr suggested this.

[ezhang61]

@mcfrank, Mika and I have talked about this new feature. I'm going to apply the method here: https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.binomtest.html
with params as binomtest(3, n=12, p=0.25, alternative='less'), any other column I should add to runs table besides the pvalue?