Confidence interval for proportions in Contingency Analysis

[bayesfactor]

We often use a contingency analysis to show outcomes across groups. However, there is no visual indication on the plot of confidence intervals, giving users no idea about whether the differences are significant or not. What I'd suggest is more like the 2nd image below. Confidence intervals should be calculated using, for example, Agresti-Coull, Clopper-Pearson, etc (https://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval#Agresti%E2%80%93Coull_interval)

bayesfactor_0-1587491835626.png

bayesfactor_1-1587491983679.png

[bayesfactor1]

+1.  The number of times we make a poor decision because someone glances at these charts without confidence intervals really sets us back. I suggest everyone use python or R because of this deficiency.

[Sarah-Sylvestre]

Hi @bayesfactor, thank you for your suggestion! We have captured your request and will take it under consideration.