Example of Mantel-Haenszel-Cochran Test
main topic     interpreting results     session command     see also 

You want to assess the association between gender and smoking preference at each activity level (A lot, Moderate, or Slight) for individuals in your study, assuming there is no three-way interaction.

1    Open the worksheet EXH_TABL.MTW. If you have not already done so, set the value order for the variable Activity.

2    Choose Stat > Tables > Cross Tabulation and Chi-Square.

3    Choose Raw data (categorical variables).

4    In Rows, enter Gender.

5    In Columns, enter Smokes.

6    In Layers, enter Activity.

7    Under Display, check Counts.

8    Click Other Stats. Under Tests for 2x2 tables, check Mantel-Haenszel-Cochran test for multiple tables.

9    Click OK in each dialog box.

Session window output

 Interpreting the results

The cells in the table contain the counts for Gender (row) and Smoke (column). A two-way table is displayed for each Activity level (Slight, Moderate, and A lot).

The common odds ratio of 1.85449 combines data from all three tables to estimate that the ratio of smokers to non-smokers among males is 1.85449 times the same ratio among females. The Mantel-Haenstzel-Cochran (MHC) test assumes that the odds ratio is the same for all three levels of activity. The MHC test has a p-value of 0.318927, indicating that the observed difference between males and females is not statistically significant.