StatFas emblem: a philosopher beside a rising bar chart StatFas.comAn Easy Way to Do Statistics

Paired Proportions Chart Maker: Before and After

Use this tool when the same people or items are classified twice, for example pass or fail before and after training, or positive or negative on two tests. Paste or upload your paired yes/no data below, read the paired comparison, edit colours and labels, and download the figure as a picture or as an APA 7 Word file.

About the paired proportion plot

A paired proportion plot compares the share of cases with an outcome at two matched time points or conditions. Each subject appears in both, so the data form a 2 × 2 table of agreement and change rather than two independent groups.

The standard test for this design is McNemar's test, which looks only at the pairs that changed. Plotting the two proportions side by side, ideally with confidence intervals, shows both the size and the direction of the change.

When to use it

Use a paired proportion plot for before-and-after binary outcomes, for matched case-control data, or when two tests or raters classify the same subjects. For independent groups use a standard comparison of proportions instead, and for more than two time points consider Cochran's Q.

What this tool offers

Assumptions to check

How to report it in APA7

Report the two proportions, the number of pairs that changed in each direction, the McNemar test statistic and p-value, and cite the figure as Figure 1 with a note about the confidence intervals. The generator writes the text, caption and note for you.

Frequently asked questions

Is the paired proportion plot generator free?

Yes. You can create and edit the plot without an account. Create a free account or log in only when you want to download it.

When should I use McNemar's test?

Use it when the same subjects are classified twice with a yes or no outcome, such as before and after an intervention.

How is this different from comparing two independent proportions?

With paired data the two measurements are correlated, so only the pairs that changed carry information about the difference. Independent-group tests would ignore this.

Is my data uploaded to a server?

The plot is built in your browser. Your data stays on your device while you work.

Related tools