Because these three strategies produce similar data, they also tend to use
similar statistical techniques. For example, t tests and analysis of variance are
used to evaluate mean differences and chi-square tests are used to compare
proportions.
Correlational studies do not involve comparing different groups of scores.
Instead, a correlational study measures two different variables (two different
scores) for each individual in a single group and then looks for patterns within the set of scores (see Figure 6.2). If a correlational study produces numerical
scores, the data are usually evaluated by computing a correlation (such as the
Pearson correlation). If the data consist of nonnumerical classifications, the
statistical evaluation is usually a chi-square test.
Descriptive studies are intended to summarize single variables for a specific
group of individuals. For numerical data, the statistical summary usually consists
of a mean, or average, score. If the data are nonnumerical classifications, the summary
is typically a report of the proportion (or percentage) associated with each
category. For example, the average student sleeps 7 hours a day and eats two pizzas
a week. Or, 58% of the students report having failed at least one course.