A total of 193 white undergraduate college students (60 percent female and 40 percent male) at Louisiana State University volunteered for the study. All subjects participated in small groups ranging in size from 12 to 18. Subjects completed demographic information forms and instruments that measured belief in a just world, religious dogmatism, and locus of control. Each subject read a written summary of a murder, with variances in defendant SES, defendant race, and type of murder. Each subject decided how many years of imprisonment to recommend for the defendant. Only defendant SES predicted sentencing severity; low SES defendants were given significantly longer sentences than high SES defendants. None of the measures of juror characteristics correlated with sentencing severity. Failure to replicate significant relationships between most of the variables and sentencing severity suggests that studies which manipulate only one variable dimension may overestimate the influence of these variables in mock jurors' decisions. 2 tables and 35 references.