6. Patients on aspirin and dipyridamole discontinued trial medication more often than those on aspirin alone, 470 out of 1363 compared to 184 out of 1376. What method would we use to test the null hypothesis that discontinuation is unrelated to medication?
These are two categorical variables, so we could use a chi-squared test or Fisher’s exact test. These would give the same answer for so large a sample size.
We get chi-squared = 167.89 with d.f. = 1, P < 0.001. Fisher's exact test also gives P < 0.001.
Back to question.
Back to Biostatistics in Research Practice index.
To Martin Bland's home page.
This page maintained by Martin Bland.
Last updated: 5 April, 2007.
Back to top.