We conducted negative binomial regressions for each sample to understand which subfactor of the SPQ-BR (interpersonal, cognitive-perceptual, and disorganized) best predicted the Pattern Glare score. Negative binomial regressions were used to adjust for overdispersion in Poisson regressions and to accommodate the skewed distribution of the outcome variable (number of reported illusions in the Pattern Glare). All assumptions were met when conducting regressions. Additionally, there were no influential outliers as assessed by the Cook’s distance values being less than one (all Di <.04).
All tests were corrected for multiple comparisons with a Bonferroni adjusted alpha level of .02. Statistical analyses were carried out using Statistical Package for the Social Sciences (SPSS Version 27).