Revisit the data in the "crime" sheet. Divide the newly generated

variable number of police officers per 1000 inhabitants into two

categories - high and low (determine the threshold value of the

variable reasonably, using descriptive statistics). Based on the new

categorical variable, divide the sample of cities into two groups - with

high and with low number of police officers per 1,000 inhabitants.

Plot a polygon of relative crime rates per 1000 inhabitants for each

subsample into one common graph. How can an image be interpreted

in practice?