Linking Research Hypotheses to Analyses

From Clinicaltools.com

For each of the statistical tests, the researcher sets alpha (usually at .05) and finds the value of the statistic (i.e., r, t, or F). With statistical computer programs (e.g., SPSS or SAS) the statistic will be shown with a probability. If the probability shown on the printout is equal to or less than the alpha level, the statistic is said to be "significant."

Example Hypotheses	No. Groups	No. Measures	Statistic
Correlation: A systematic, linear relationship between two variables (e.g., as one goes up, the other goes up) Hypothesis: There is a relationship between post-test scores and the amount of time spent on a module.	1	2 (post-test and amount of time spent)	Correlation Coefficient (r)
Paired t-test: Difference between two means Hypothesis: Participants do better on the post-test than on the pre-test.	1	2 (pre-test and post-test)	Paired t-test (t)
t-test for independent samples: Difference between the means of two groups Hypothesis: The treatment group does better than the comparison group on the outcome measure	2	1 (outcome measure)	t-test for independent samples (t)
Analysis of Variance (ANOVA): Difference among means - three or more groups Hypothesis: Group 1 does better than Group 2 or Group 3 on the outcome measure.	3 or more	1 (outcome measure)	Analysis of Variance (F)
Analysis of Covariance (ANCOVA): Difference among means - three or more groups, controlling for a confounding variable Hypothesis: Group 1 does better than group 2 or group 3 on the outcome measure, controlling for previous achievement.	3 or more	2 (outcome measure and covariate measure)	Analysis of Variance (F)