Skip to content
AP Statistics · Exploring One-Variable Data and Collecting Data

Comparing groups

Comparing groups appears in 3 reviewed AP Statistics questions and sits in Exploring One-Variable Data and Collecting Data. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Comparing groups is tested

Example 1Harder
A sports physiologist measures the VO2 max of 30 male athletes (mean = 55, SD = 6) and 30 female athletes (mean = 48, SD = 5). She constructs separate 95% confidence intervals: Males: (52.76, 57.24), Females: (46.13, 49.87). She notices the intervals do not overlap and concludes the means are significantly different. A statistician warns this method is:
  1. ACorrect; non-overlapping intervals always indicate a significant difference, so no further test is needed
  2. BIncorrect; overlapping intervals always indicate no difference between the two means
  3. CConservative; non-overlap implies significance, but overlap does not rule it out, so a two-sample test is preferredCorrect
  4. DInvalid; separate confidence intervals for each group can never be used, even informally, to compare two population means

Why C is right

While non-overlapping confidence intervals do imply a significant difference between the means, the converse is not true. Overlapping intervals do not necessarily mean there is no significant difference. A formal two-sample confidence interval or hypothesis test is the proper way to compare two means. Comparing individual CIs is conservative (may miss real differences).
Why the other 3 choices miss
A
Incorrect because while the conclusion is correct in this case, the method is not fully reliable: overlapping intervals can still correspond to a significant difference, so a formal test is still warranted.
B
Incorrect because overlapping intervals do not always indicate no difference; the overlap must be substantial relative to the margin of error.
D
Incorrect because separate CIs can provide some informal comparison, but the method has limitations and should not replace formal tests.

Studied alongside

Other concepts from Exploring One-Variable Data and Collecting Data.