Testosterone and Antisocial Behavior

David C. Howell

Introduction

Dabbs and Morris (1991, Psychological Science) examined the military records of 2736 American males. These records contained information on testosterone levels (Lord Knows Why!) and incidence of adult delinquency. Dabbs and Morris were interested in the relationship between delinquency and testosterone. Below are the data separated by High and Low SES subjects.

		a.  Low SES



					   Delinquency    

					Yes		No

			Normal		190		1104

			High		 62		 140







		b. High SES



					   Delinquency    

					Yes		No

			Normal		53		1114

			High		 3		  70

Collapse across the two SES conditions to create one 2 X 2 table.

• What model would you reasonably test here?
• Run that analysis using chi-square and draw the appropriate conclusion.
• What does this analysis tell you about the rule of testosterone in adult delinquency?

Test the relationship between Testosterone and Delinquency within each SES level.

• What models are you testing?
• What is the resulting chi-square?
• What would you conclude?
• How has this approach told you more than you knew before?

Calculate the odds ratios for the two analyses in the second problem.

• What do these results tell you?

For those people who have made it to the chapter on log-linear models, create and test models for all three variables (Testosterone, Delinquency, and SES).

• What are suitable models?
• What would you conclude?

Return to Dave Howell's Statistical Home Page  

Last revised: 7/11/98