I randomly generate the scores for each group with a condition that the MEAN of all of the 56 scores for the treatment group is at least 1 point greater than the mean of 28 placebo scores. This condition eliminates the need for 1000s of simulations and consistently shows this as the threshold for p- value hovering around 0.05. You can also calculate the sd and do a simple t or z test since the nominal data from 0 to 12 has some degree of normality.
Once you generate the data you copy paste into an online statistic calculator. The non parametric tests are based on the rank of data point in the sample.
Market Data 
Markets 
AI
