In Landon et al. (2009), 958 women with mild gestational diabetes were randomized to usual care or treatment. Among the control group, 68 babies were LGA; among the treatment group, 29 were LGA. Which test would compare these proportions?

In Landon et al. (2009), 958 women with mild gestational diabetes were randomized to usual care or treatment. Among the control group, 68 babies were LGA; among the treatment group, 29 were LGA. Which test would compare these proportions?

• A. 2-proportion z-test
• B. Chi-square test of independence
• C. Fisher's exact test
• D. Chi-square test of homogeneity

Answer: (D)

The main idea is to compare proportions across two independent groups using a contingency-table approach. Here the outcome is LGA (yes or no) and the groups are usual care vs treatment, so you get a 2x2 table. The chi-square test of homogeneity asks whether the distribution of a binary outcome is the same across these two populations. It tests the null that the proportion of LGA babies is identical in both groups, with the alternative that they differ. Because the study has a large sample size, the chi-square approximation is appropriate and reliable. A 2-proportion z-test would also address the same question in large samples, but the chi-square test of homogeneity is the standard way to compare distributions of a categorical outcome across populations. Fisher’s exact test would be used only if cell counts were very small; that isn’t the case here.

The main idea is to compare proportions across two independent groups using a contingency-table approach. Here the outcome is LGA (yes or no) and the groups are usual care vs treatment, so you get a 2x2 table. The chi-square test of homogeneity asks whether the distribution of a binary outcome is the same across these two populations. It tests the null that the proportion of LGA babies is identical in both groups, with the alternative that they differ. Because the study has a large sample size, the chi-square approximation is appropriate and reliable. A 2-proportion z-test would also address the same question in large samples, but the chi-square test of homogeneity is the standard way to compare distributions of a categorical outcome across populations. Fisher’s exact test would be used only if cell counts were very small; that isn’t the case here.