Statistics Resources

T-tests are used to compare two groups on some numerical measure. The results of the test will tell you if there are any significant differences between those two groups. Here are some examples of what types of research questions and/or hypotheses may indicate a t-test is appropriate:

Single Sample t-test:

• RQ: Does the Read with Me program improve reading scores compared to previous years?
  • H0: There is no improvement in readings scores after participating in the Read with Me program.
  • Ha: There is significant improvement in reading scores after participating in the Read with Me program.

Here, the study would give the intervention to one group of individuals and compare their performance to historical averages for reading scores. Because the comparison data is the historical average, this is a single sample t-test.

Independent Samples t-test:

• RQ: Is Lowfya more effective at reducing depression symptoms than DepLo?
  • H0: There is no difference in depression symptoms between individuals taking Lowfya and those taking DepLo.
  • Ha: Participants taking Lowfya have significantly lower depression symptoms than those taking DepLo.

This study would be constructed with two separate groups of individuals: one group taking Lowfya and one group taking DepLo. Because the comparison is between two different, unrelated groups, this is an independent samples t-test. If participants are randomly assigned to a group, this would be an experiment.

Dependent/Paired Samples t-test:

• RQ: Do people recall more words after learning a memorization strategy?
  • H0: There is no difference in word recall frequency after learning a memorization strategy.
  • Ha: There is a significant difference in word recall frequency after learning a memorization strategy.

This study would be constructed with one group of individuals. Word recall would be assessed before and after the intervention. Because the comparison is between two scores obtained from the same group of individuals, this is a dependent samples t-test.