Parameter limits determine which distributions can be fitted. Non-negative lower limits are needed for the gamma, lognormal and log-t distributions, and both limits must be finite for to fit a beta distribution. If a histogram is fitted without specifying finite limits, endpoints are chosen based on fitting a normal distribution.As an example, if the elicited judgements are P(X<15)=0.25, p(x<20)="0.5)" and="" p(x<40)="0.75," specify="" the="" parameter="" values="" as="" 15,20,40="" cumulative="" probabilities="" 0.25,0.5,0.75.<="" p="">
Press Esc in the R console window to exit the elicitation session.