On Test of Association Using Attributable Risk for a 2x2 Contingency Table

• Activity Number: 613
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Tanweer Shapla* and Khairul Islam
• Companies: Eastern Michigan University and Eastern Michigan University
• Keywords: Attributable risk ; Independence ; Power of the test ; Monte Carlo simulation

Abstract

Attributable risk is a widely used measure for assessing risk of a factor in public health and biostatistics. It provides the proportion of disease reduction due to the elimination of the risk factor from the population of interest. It is to be noted that attributable risk is rarely used for test of association or independence. The chi-square test investigates if a certain factor is independent of any outcome. However, if the null hypothesis of independence or no association is rejected, this test cannot provide any insight on whether the factor is associated positively or negatively. This paper considers test of hypothesis of independence or no association using attributable risk and discusses sensitivity of power analysis, theoretically and by Monte Carlo simulation.