Related Experiment Video
Updated: Jul 17, 2026
07:29
Online Explorative Study on the Learning Uses of Virtual Reality Among Early Adopters
Published on: November 22, 2019
The analysis of stratified 2 x 2 contingency tables
Vance W Berger1, Catalina Stefanescu, Yan Yan Zhou
1University of Maryland Baltimore County and National Cancer Institute, Executive Plaza North, Suite 3131, Bethesda, MD 20892-7354, USA.
Biometrical Journal. Biometrische Zeitschrift
|January 24, 2007
Summary
This study evaluates statistical tests for binary outcomes in clinical trials. Novel tests are developed to improve power when treatment effects vary across patient groups, outperforming the standard Cochran-Mantel-Haenszel test in such scenarios.
Area of Science:
- Biostatistics
- Clinical Trial Methodology
- Statistical Inference
Background:
- Binary response variables are common in clinical trials, necessitating robust statistical methods for comparing treatments across strata.
- The Cochran-Mantel-Haenszel (CMH) test is frequently used but assumes a common odds ratio, which may not hold in practice.
- Issues arise regarding data pooling and the performance of standard tests when treatment effects (odds ratios) differ across strata.
Purpose of the Study:
- To investigate the performance of statistical tests for independence against consistent treatment superiority with binary outcomes in stratified randomized clinical trials.
- To address challenges related to small/zero margins, common odds ratio assumptions, and the impact of differing odds ratios across strata.
- To develop and propose novel statistical tests that are more appropriate when odds ratios may vary.
Main Methods:
- Analysis of the power profile of the Cochran-Mantel-Haenszel test under varying odds ratios.
- Development of new statistical tests analogous to Smirnov, modified Smirnov, convex hull, and adaptive tests.
- Evaluation of test performance in randomized clinical trial settings with binary outcomes and multiple strata.
Main Results:
- The commonly used Cochran-Mantel-Haenszel test can exhibit a poor power profile when odds ratios differ across strata, despite its optimality under common odds ratios.
- Novel tests analogous to those for ordered categorical data demonstrate potential for improved power and appropriateness in situations with heterogeneous treatment effects.
- The study highlights the limitations of assuming a common odds ratio and the need for flexible analytical approaches.
Conclusions:
- Standard tests assuming a common odds ratio may be suboptimal in clinical trials where treatment effects vary.
- Newly developed statistical tests offer a more powerful and appropriate alternative for analyzing binary outcomes in stratified settings with heterogeneous treatment effects.
- This research provides valuable insights for biostatisticians and clinical researchers designing and analyzing trials with binary endpoints.
Related Concept Videos
01:29Contingency Table
A contingency table provides a way of portraying data that can facilitate calculating probabilities. It is a method of displaying a frequency distribution as a table with rows and columns to show how two variables may be dependent (contingent) upon each other; The table helps determine conditional probabilities quite quickly and can help systematically organize, analyze and quantify data. The table displays sample values concerning two variables that may be dependent or contingent on one...
01:21Introduction to Test of Independence
In statistics, the term independence means that one can directly obtain the probability of any event involving both variables by multiplying their individual probabilities. Tests of independence are chi-square tests involving the use of a contingency table of observed (data) values.
The test statistic for a test of independence is similar to that of a goodness-of-fit test:
The test statistic for a test of independence is similar to that of a goodness-of-fit test:
01:08Determination of Expected Frequency
Suppose one wants to test independence between the two variables of a contingency table. The values in the table constitute the observed frequencies of the dataset. But how does one determine the expected frequency of the dataset? One of the important assumptions is that the two variables are independent, which means the variables do not influence each other. For independent variables, the statistical probability of any event involving both variables is calculated by multiplying the individual...
01:16Hypothesis Test for Test of Independence
The test of independence is a chi-square-based test used to determine whether two variables or factors are independent or dependent. This hypothesis test is used to examine the independence of the variables. One can construct two qualitative survey questions or experiments based on the variables in a contingency table. The goal is to see if the two variables are unrelated (independent) or related (dependent). The null and alternative hypotheses for this test are:
H0: The two variables (factors)...
H0: The two variables (factors)...
01:23McNemar's Test
McNemar's Test is a nonparametric statistical test used to determine if there is a significant difference in proportions between two related groups when the outcome is binary (e.g., yes/no, success/failure). It is beneficial when we have paired data, such as pre-test/post-test designs, where the same subjects are measured under two different conditions. The test is named after the statistician Quinn McNemar, who introduced it in 1947. It is commonly used in situations where subjects are...
01:21Friedman Two-way Analysis of Variance by Ranks
Friedman's Two-Way Analysis of Variance by Ranks is a nonparametric test designed to identify differences across multiple test attempts when traditional assumptions of normality and equal variances do not apply. Unlike conventional ANOVA, which requires normally distributed data with equal variances, Friedman's test is ideal for ordinal or non-normally distributed data, making it particularly useful for analyzing dependent samples, such as matched subjects over time or repeated measures from...