These groupings allow comparison of 'current', 'previous', and 'no history of self-injury' groups on all focal variables.
The most important variables according to this criterion are marked with † The second value corresponds to the gain of a model fitted using only the focal variable.
For each variable, we fitted the model 2000 times - once with the original data, and 1999 times with the focal variable replaced by a random permutation of itself.
From our previous experience [ 28, 29], and as recently demonstrated by Laakso et al. [ 25, 26], it has been apparent that whereas in some tumours virtually every cell expresses the basal marker (CK14, in this instance), others may show a focal or variable percentage of positive cells.
In other words, the focal variable was the induction of a positive interpretational tendency via mental imagery, and as such, the control condition was designed with this in mind.
The two-tailed permutation P-value is twice the proportion of these fits (including the original) whose coefficients for the focal variable are greater (respectively, less than) or equal to a positive (negative) coefficient from the original fit.
Because we also wanted to investigate the effect of body mass in our analyses, however, we repeated each sequential regression with body mass as the focal variable and brain mass included as the residuals from a PGLS log-log regression of brain mass on body mass.
In detail, we use the focal variable, social media interactions, as well as additional contractual (e.g., advertising consent, upselling behavior, and acquisition channel) and demographic covariates (e.g., age, gender, credit score, and purchasing power).
We initially included interactions between the focal variable and study type in the models, but the interaction term was never significant, so we removed it from the final models.
For continuous focal variable (household wealth), we tested this using bi-serial correlations.
