I would like to utilize the chi-squared built-in function to get
p-values for each experiment (pivoted, in this example aa00 and aa01),
when comparing expected vs observed values. However, if one experiment
has more variants then the other (aa01 only...