Skip to content
AP Statistics · Inference for Categorical Data: Proportions

One sided test

One sided test appears in 4 reviewed AP Statistics questions and sits in Inference for Categorical Data: Proportions. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How One sided test is tested

Example 1Medium
A company tests whether a training program increases the average number of widgets produced per hour. Before training, the mean was mu_0 = 45. After training, a random sample of 35 workers has x-bar = 48.3, s = 7.5. The test statistic is t = (48.3 - 45)/(7.5/sqrt(35)) = 2.60 with df = 34. The p-value for the one-sided test Ha: mu > 45 is approximately 0.007. At alpha = 0.01, what is the conclusion?
  1. AReject H0 because the sample mean of 48.3 is greater than 45
  2. BReject H0; there is convincing evidence that the mean production rate after training exceeds 45 widgets per hourCorrect
  3. CFail to reject H0 because the observed increase of 3.3 widgets per hour is too small to be practically significant
  4. DFail to reject H0 because 0.007 is very small

Why B is right

Since the p-value (0.007) is less than alpha (0.01), we reject H0. There is statistically significant evidence at the 0.01 level that the mean production rate after training exceeds 45 widgets per hour. The small p-value indicates the observed increase is very unlikely to have occurred by chance alone if the training had no effect.
Why the other 3 choices miss
A
Incorrect. While the conclusion (reject) is correct, the reasoning is wrong. We reject because the p-value < alpha, not simply because x-bar > mu_0. Any sample could have x-bar > 45 by chance.
C
Incorrect. While practical significance is a valid consideration, the question asks about the statistical conclusion at alpha = 0.01. The p-value is less than alpha, so we reject H0.
D
Incorrect. A small p-value leads to rejection, not failure to reject. The smaller the p-value, the stronger the evidence against H0.

Studied alongside

Other concepts from Inference for Categorical Data: Proportions.