About Grouped tables (two-grouping variables)

What is a grouped (two-grouping variable) table?

The idea of grouped variables is best understood by example. One grouping variable might be "male vs. female"; the other grouping variable is "control vs. treated." Every data point you collect is categorized by both grouping variables: "male vs. female," AND "control vs. treated." A two-grouping variable table and graph is used to tabulate outcomes that are measurements such as weight or blood pressure, often with error bars.

Understand the difference between grouped tables and contingency tables where each value you enter is an actual count of the number of subjects that fall into a category.

Analyses from a grouped table:

•	Two-way ANOVA

•	Repeated-measures two-way ANOVA (mixed-model)

•	Bonferroni post test

Graph types from a grouped table:

Example of a grouped table:

•	Unlike other programs, Prism does not use separate columns to enter grouping variables. Instead, groups are defined by rows and columns.

•	The rows categorize the data by one grouping variable (men vs. women).

•	The columns (datasets) categorize the data by the other grouping variable (control vs. treated).

•	Blank cells represent missing values, which Prism handles properly. If you are doing repeated-measures ANOVA, however, you can't have any missing data values unless an entire subcolumn is empty.

•	Blue italics values are “excluded” (ignored by analyses and graphs). Use Edit… Exclude values.

Error bars

When you create a new data table, one of the essential choices is the subcolumn format. If you choose to enter replicate (i. e. duplicate or triplicate) values of Y at each value of X, Prism will either plot these individually or compute and plot error bars. Or you can enter error values (i. e. SD or SEM) that you have already calculated.

•	If you want Prism to calculate your error bars, tell Prism how many replicates you have for each subject.

•	Prism will create a data table with that number of subcolumns for you to enter your data. Enter your replicates in side-by-side columns.

•	Prism can create subcolumns formatted for error values computed elsewhere.