This dataset example introduces readers to McNemar's test. This technique allows researchers to test whether there is a difference between two related groups of measurements of a binary (i.e., dichotomous) variable. Hence, it is commonly used to study any change in the subjects from one time point to another, or to compare the effects of different treatments/conditions. This example uses a subset of data from the College Scorecard Data from the U.S. Department of Education. It examines any change over time in the number of U.S. public colleges whose two-year default rates are 0 (i.e., no borrower of federal student loans defaulted within the first two years since entering repayment). Results from an analysis like this could provide useful information on the debt burden of college students, and help students make knowledgeable decisions when borrowing for college. In this example, readers are introduced to the basic theory and assumptions underlying this technique, the type of questions this technique can be used to answer, and how to produce and report results. The sample dataset has been cleaned and organized to make this example easier to follow. Interested readers should read the full documentation for the dataset before using it for research (https://collegescorecard.ed.gov/data/).

Description based on XML content.

Licensed e-book