Small Sample Properties of the Mantel-Haenszel Test Statistic for Density Follow-Up Studies

摘要: For the analysis of combinations 2×2 non-contingency tables as obtained from density follow-up studies (relating a number events to person-years follow-up) an analogue Mantel-Haenszel test for contingency is widely used. In this paper small sample properties test, both with and without continuity correction, are evaluated. Also improvement test-statistic by using first four cumulants via Edgeworth expansion was studied. Results on correction agree similar statistic tables: Continuity gives p-value which approximates exact better than correction; its approximations show considerable conservatism in samples; uncorrected that agrees more nominal significance level, but can be anti-conservative. The based approximation corrected especially when distribution has marked skewness.