Comparing the change in behaviour (from 2014 to 2015) for two groups of taxpayers: one group was audited, the other – not. Audited group reported median increase in taxable income for $100, the control – for $50. I am pretty clear with median test and Mann Whitney U test. However, I noticed that although we might have taxpayers reporting similar increases in income for $100, for one this amount represents 50% increase, for the other – only 0.01%. Now the question is: if I calculate % changes from 2014 to 2015, can I use those values to do median test and/or Mann Whitney test? Cannot find anything in literature L