marginal significancep -value < 0.1
Overall differences in Likert-type data were tested by a multivariable analysis of variance (MANOVA) for all questions with marginal significance ( p -value < 0.1) at the individual question level.
Overall differences in Likert-type data were tested by a multivariable analysis of variance (MANOVA) for all questions with marginal significance ( p -value < 0.1) at the individual question level.
Then, using any factors that were marginally significant ( p < 0.1), we built MANOVA models, using gender, amount of debt, or prior urology clerkship as the explanatory variables.