SDM calculations

Linear model analyses: comparing groups

This procedure calculates the differences between two or three groups, with their statistical significance based on Monte Carlo randomizations. You can specify covariates to remove potential confounds, and a filter for subgroup analyses.

Notes

Before conduting this analysis, all groups other than the 'first' must be coded as 'indicator' (i.e. binary) variables:

- To compare 2 groups, a filter must be created for the 'second' group, i.e. a value of '0' must be set to studies from the first group and a value of '1' to studies from the second group.

- To compare 3 groups, a filter must be created for the 'second' group, i.e. a value of '0' must be set to studies from the first and third groups and a value of '1' to studies from the second group, as well as a filter for the 'third' group, i.e. a value of '0' must be set to studies from the first and second groups and a value of '1' to studies from the third group.

If one variable is selected in the dialog (e.g. a filter for a two-group comparison), '0' estimates the value at the minimum value of the variable (e.g. the 'first' group), '1' estimates the value at the maximum value of the variable (e.g. the 'second' group), and '1m0' estimates the difference between the maximum and the minimum values of the variable (e.g. the difference between the 'second' and the 'first' groups). If two variables are selected, '10' relates to the maximum value of the first variable and the minimum of the second, while '01' relates to the minimum value of the first variable and the maximum of the second. And so on...

If two variables are selected, the two-variable Q is also computed, with a meaning similar to the F of an ANOVA (see Radua et al. Arch Gen Psychiatry 2010 for details).

Variables are automatically scaled to have values between 0 and 1.

To conduct a comparison between groups

Press the button [Linear model] and [Compare 2 groups] or [Compare 3 groups]

or:

Select [Linear model] in the Statistics menu, to open the following dialog:

Command-line and batch usage

lm formula, filter

Example:

AdultsAgeEffect = lm disorder

Log

The p = 0.005 thresholds obtained after each randomization are saved in a text file called (name_of_the_comparison)_z.htm, e.g. MyThreeGroups_z.htm. This is useful for studying the stability of the threshold.

References

(comparisons and linear models): Radua J, van den Heuvel OA, Surguladze S and Mataix-Cols D. Meta-analytical comparison of voxel-based morphometry studies in obsessive compulsive disorder vs other anxiety disorders. Arch Gen Psychiatry 2010; 67:701-711. .