The model incorporates the possibility of endogenous trial costs, differential quality of information obtained by the disputants, differential bargaining strengths, differential expectations regarding both the probability of the award and the amount of the award given a verdict favoring the plaintiff, and differential stakes in the outcome. Results indicate that certain case characteristics influence the selection of cases for trial. For example, cases with potentially large awards are more likely to end in trial than those with smaller awards. If trial data show a negative relation between the probability of a plaintiff verdict and the size of the expected award given a plaintiff verdict, then there is likely to be a positive correlation between these two variables in settlement data. In another case, if the asymmetry in the litigants' ability to estimate a trial outcome produces trial outcomes below the average of the final demand by the plaintiff and the final offer by the defendant, then, for those cases that are settled, the expected trial outcome will be above the actual amount settled on. Numerous other results are generated, some of which are counterintuitive. The model provides guidelines for making inductions from trial data to settled cases that have implications for policy prescriptions regarding administration of the courts, compensation for injuries, and dispute settlement costs. Appendixes and 63 footnotes.