Exercise 11E: Comparing groups of nominal data defined by two grouping variables in Prism

Go to parent GraphPad Prism statistical analyses

We used a number of perforin-deficient and wild type mice and used flow cytometry to count T-cell subpopulations in these mice. We counted the number of CD8+ naive cells, CD8+ central memory T cells (TCM) and CD8+ effector memory T cells (TEM). All variables are nominal: wt/mutant and CD8+ naive/TCM/TEM.
The question is: Is there an effect of the mutation on the distribution of CD8+ T cells?

The data are available as a csv file.

Which table type should you use for these data?
Unranked categorical data defined by two grouping variables should be stored in a contingency table.

Contingency tables are used to tabulate the actual number of subjects that fall into the categories defined by the rows and columns of a table, so you have to use counts as inputs. GraphPad advises against entering normalized values or percentages.

Enter the data in Prism
Create a new contingency table Import the csv file

Look at the graph that is generated by Prism
Go to the Graphs folder and click the corresponding graph. The default graph that is generated by Prism (Interleaved bar chart) is OK.

Remove Xtitle from the graph
In the Change section of the top toolbar click the Format axes button. On the Titles and fonts tab deselect Show X axis Title and click OK.

Add counts as Ytitle
Just place your cursor on YTitle and type counts.

Change the colors of the bars of both groups
In the Change section of the top toolbar click the Format graph button. In the Bars and boxes section change Fill, Pattern and Border color and click OK. To do the same for the other group, click the Format graph button again and switch the Data set to TwoFactors_nominal; perforin mt Now change the color of the bars.

Which test are you going to use to compare wt and mutant ?
Since it's a 3x2 table we have to use the chi square test.

Is there a difference between wt and mutant ?

Go back to the data table. In the top toolbar click the Analyze button in the Analysis section. In the Contingency table analyses section select Chi square (and Fisher's exact) test and click OK Take default parameter setting: you have 3 categories so you need a ChI Square test. You want a two-tailed test since you want if there is a difference between the groups: you're not stating anything about the direction of that difference. A confidence level of 0,05 is fine. The p-value is smaller than 0,05 meaning that we have sufficient evidence to reject the null hypothesis. Since H0 states that there is no difference we can state that wt and CD8+ deficientv mice have different CD8+ cell populations.