One hundred adults who had either Type 1 or Type 2 diabetes were divided into 2 groups to determine if the inhalable version of insulin is able to manage blood sugar levels just as well as injected insulin. At the end of the trial period we count the number of Type 1 diabetics who reacted better to the inhaler than to the injections. We do the same for the Type 2 diabetics. If we wish to determine if there is a difference in the percentages of diabetics who reacted better to the inhaler between Type 1 and Type 2 diabetics, which inference procedure should we use?
A) A t-test about the difference between two means for independent samples.
B) A z-test about the difference between two proportions.
C) A confidence interval for the difference between two means for independent samples.
D) A confidence interval for the difference between two proportions.

Question

Quizplus · Accepted Answer

DThe correct inference procedures for comparing the percentages of diabetics (Type 1 vs. Type 2) who reacted better to the inhaler are a z-test about the difference between two proportions (B) and a confidence interval for the difference between two proportions (D). Both methods are appropriate because the data involve comparing proportions (percentages) from two independent groups.

One Hundred Adults Who Had Either Type 1 or Type