(e) Since the Durbin-Waston statistic may not be appropriate, a non-parametric test, such as the
     runs test, should be used. For the runs test, on theTheil-nagar adjusted model, n = 14, n₁ = 8,
     n₂ = 6, and k = 10. The 95% confidence level values of runs are 3 and 12, as shown in the
     Swed and Eisenhart tables. Since k falls within this range, there is no evidence of autocorrelation
     in the Theil-Nagar corrected model.