Use the Matching Column option to specify a matching (ID) variable for a matching model analysis. The Matching Column option addresses the case when the data in a one-way analysis come from matched (paired) data. Matched data can occur when observations in different groups come from the same participant. The Matching Column option is not available when a Block variable is specified in the launch window. See Example of the Matching Column Option.
Note: A special case of matching leads to the paired t test. The Matched Pairs platform handles this type of data, but the data must be organized with the pairs in different columns, not in different rows.
The Matching Column option performs two primary actions:
• It fits an additive model (using an iterative proportional fitting algorithm) that includes both the grouping variable (the X variable in the Fit Y by X analysis) and the matching variable that you select. The iterative proportional fitting algorithm makes a difference if there are hundreds of participants, because the equivalent linear model would be very slow and would require huge memory resources.
• It draws lines between the points that match across the groups. If there are multiple observations with the same matching ID value, lines are drawn from the mean of the group of observations.
The Matching Column option automatically activates the Matching Lines option connecting the matching points. To turn the lines off, select Display Options > Matching Lines.
The Matching Fit report shows the effects with F tests. These are equivalent to the tests that you get with the Fit Model platform if you run two models, one with the interaction term and one without. If there are only two levels, then the F test is equivalent to the paired t test.
Note: For more information about the Fit Model platform, see Model Specification in Fitting Linear Models.