I have a research question for an assignment where I am trying to investigate if a certain DV can be predicted by 4 continuous IV's and a categorical IV (gender)
I'm extremely confused on how to approach the question, initially, I just used the lm() function with the relevant equation:
reg.mod <- lm(self.efficacy ~ sex. + swb + extr + self.concept + inte, data = data.reg) coef(reg.mod)
However, afterwards when i needed to check the statistical assumptions underlying the linear regression model such as: linearity, constant variance of residuals and normality of residual scores, i ran into problems because of the included gender factor variable.
For example, i had to look at a scatter plot matrix, residual plots, marginal model plots, marginal conditional plots, a Breusch-Pagan test [using ncvTest()] , and an influence index plot using Cook's d value and studenised residual values.
Should i approach this question as two separate questions? One with just self.efficacy ~ sex. and self.efficacy ~ swb etc.. without sex?
Any thoughts would be greatly appreciated.