52.
The program DATASIM is used to simulate the classic “horn-honking” study by Doob and Gross (1968). In a 2×2 field experiment, Doob and Gross investigated the effects of status of frustrator—a low- or high-status car blocking an intersection—on the latency to honk among male and female drivers. The present paper illustrates how to extract the values of simulation parameters from the published study, how to initialize the simulation in DATASIM, and how to generate and analyze the simulated data. Certain complications arise because the latency data collected by Doob and Gross were nonnormally distributed, cell variances were heterogeneous, and sample sizes were unequal. DATASIM is able to incorporate these features in the simulation, and several methods for assessing the quality of the simulation are illustrated. In addition, sampling experiments are reported, which were performed to assess the joint and individual effects of nonnormality and heterogeneity on the Type I and Type II error rates of the
F test. The paper concludes with some practical suggestions regarding how researchers can evaluate, and adjust for, the effects of such violations.
相似文献