Recognize the different types of dependent samples.
In dependent samples, also called paired samples, cases are pairs of data points. Examples include (a) cases measured at two points in time (repeated measures, longitudinal, or pre-post designs); (b) cases measured in two different conditions (within-subjects designs); or (c) cases matched on some variable, so they are similar (matched-pairs designs).
Calculate and interpret a paired-samples t test.
A paired-samples t test is used to compare the means of two dependent samples. Because paired-samples t tests control for the effects of individual differences, they have more statistical power, need a smaller sample size to reject H0, and are useful for studying rare phenomena or those for which cases are hard to come by.
To conduct a paired-samples t test, the assumptions must be met, hypotheses stated, and decision rule set before calculating the actual t value.
In interpreting the results of a paired-samples t test, first apply the decision rule to determine if the null hypothesis was rejected, then write the results in APA format. If the null hypothesis was rejected, comment on the direction of the difference between the population means.
Next, to find the effect size, calculate the 95% confidence interval for the difference between population means. Interpret the confidence interval, paying attention to (a) whether it captures zero, (b) how close it comes to zero, and (c) how wide it is. Do not calculate r2 or d.
Finally, in a four-point interpretation: (1) explain what the study was about, (2) present the main results, (3) interpret them, and (4) make suggestions for future research.
316
Is there a difference in the ages at which men and women get married? Has there been a change in this statistic over time?
Go online or get a newspaper from a recent Sunday and find the wedding section. Note the ages of each bride and groom. Are these paired data?
Check the assumptions and complete a paired-samples t test. Is there a difference? In what direction? What do the results tell us?
Want to do some extra work? Go to your school library and have the reference librarian help you find a Sunday newspaper from a decade or two ago. Collect and analyze the data for the ages of the brides and grooms back then. Has a change in the age difference occurred over the years?