Which statement best differentiates a null hypothesis from an alternative hypothesis?

Which statement best differentiates a null hypothesis from an alternative hypothesis?

• A. Null states there is an effect; alternative states no effect.
• B. Null states no effect; alternative states there is an effect.
• C. Null is tested with descriptive statistics.
• D. Alternative always has a p-value less than 0.05.

Answer: (B)

In hypothesis testing, the foundational idea is how we frame competing claims about a population. The null hypothesis represents no effect, no difference, or no relationship. The alternative hypothesis represents that there is an effect, a difference, or a relationship. We use data to see whether the observed result is unlikely under the null; if it is, we favor the alternative. This makes the statement that the null states no effect and the alternative states there is an effect the best fit. The other choices miss the standard setup: the null is not about “there is an effect,” it’s about “no effect”; descriptive statistics by itself isn’t the tool used to test hypotheses (inferential tests do that); and having a p-value below 0.05 is not guaranteed for the alternative—the p-value depends on the data and the test, and you can fail to reject the null even when an effect exists.

In hypothesis testing, the foundational idea is how we frame competing claims about a population. The null hypothesis represents no effect, no difference, or no relationship. The alternative hypothesis represents that there is an effect, a difference, or a relationship. We use data to see whether the observed result is unlikely under the null; if it is, we favor the alternative.

This makes the statement that the null states no effect and the alternative states there is an effect the best fit. The other choices miss the standard setup: the null is not about “there is an effect,” it’s about “no effect”; descriptive statistics by itself isn’t the tool used to test hypotheses (inferential tests do that); and having a p-value below 0.05 is not guaranteed for the alternative—the p-value depends on the data and the test, and you can fail to reject the null even when an effect exists.