Figure 1.Cumulative Binomial Distribution Plots for Various Values of the Parameter θ. The figure shows that we can find values of θ_L and 〖 θ〗_H for which 1-p_1-p_2 is close to the desired confidence level, where p_1=F_T (t_0;θ_H) and p_2=1-F_T (t_0-1;θ_L) and to is the observed number of successes in n trials. For the plot n = 20 and t_0 = 4, and we were seeking a 95% confidence interval. For the values of θ shown we were able to find a 93.3% confidence interval by choosing θ_L to be .05 and θ_H to be .40. These values are marked on the plot. The width of this 93.3% interval is .35. By choosing other values of θ it is possible to find an exact 95% confidence interval with minimum width.