The data are yearly skin cancer rates (cases per 100,000 people) in Connecticut, with a code identifying those years that came two years after higher than average sunspot activity and those years that came two years after lower than average sunspot activity. (Data from D. F. Andrews and A. M. Herzberg, Data, New York: Springer-Verlag, 1985.)
Is there any reason to suspect that using the two independent sample t -test to compare skin cancer rates in the two groups is inappropriate?
