Hello everyone,
I am working on linear and logistic regression models, I want to determine the relationship between an ordinal variable called Social Support, with three categories being: Strong, Average, Weak/No existent and mental and physical outcomes after trauma. My dependent variables are 4: Post-Traumatic Stress Disorder (0=No, 1=Yes), Pain (0=No, 1=Yes), SF-12 Mental Compose score (0-100), SF-12 Physical Compose score (0-100). I am running four different and independent models. Each model has been adjusted for sociodemographics (age, sex, education level, Injury Severity Score).
For each outcome I have run three different regressions, let's say PTSD, so I run a Logistic Regression Analysis, first. by using Strong Social Support as reference group, second. by using Average Social Support as reference group, third. by using Weak/Inexistent Social Support as reference group. I am interested in seeing how much change these associations when reference group is changed. Problem with this approach, Multiple Testing. This is the reason I need now to perform a Bonferroni correction either for logistic or linear models.
I've been reading some materials but it makes examples after running ANOVA.
My question is: how do I do a Bonferroni correction in the context of Multivariate Logistic and Linear Regression Models? What are the codes? Any consideration? Interpretation of p values?
There is not too much out there in the Website addressing this specific topic.
Thank you for any advice and recommendation,
Claudia
I am working on linear and logistic regression models, I want to determine the relationship between an ordinal variable called Social Support, with three categories being: Strong, Average, Weak/No existent and mental and physical outcomes after trauma. My dependent variables are 4: Post-Traumatic Stress Disorder (0=No, 1=Yes), Pain (0=No, 1=Yes), SF-12 Mental Compose score (0-100), SF-12 Physical Compose score (0-100). I am running four different and independent models. Each model has been adjusted for sociodemographics (age, sex, education level, Injury Severity Score).
For each outcome I have run three different regressions, let's say PTSD, so I run a Logistic Regression Analysis, first. by using Strong Social Support as reference group, second. by using Average Social Support as reference group, third. by using Weak/Inexistent Social Support as reference group. I am interested in seeing how much change these associations when reference group is changed. Problem with this approach, Multiple Testing. This is the reason I need now to perform a Bonferroni correction either for logistic or linear models.
I've been reading some materials but it makes examples after running ANOVA.
My question is: how do I do a Bonferroni correction in the context of Multivariate Logistic and Linear Regression Models? What are the codes? Any consideration? Interpretation of p values?
There is not too much out there in the Website addressing this specific topic.
Thank you for any advice and recommendation,
Claudia
Comment