I'm currently carrying out research for my masters dissertation and i've run into a critical issue while trying to analyse my data. I'm a psychology student looking into predictors of male rape myth acceptance. My predictors are Sex, Age, Ethnicity, Estimated monthly income and Homophobia. My criterion variable and Homophobia predictor are both scale data, while Sex, Age, Ethnicity and Estimater monthly income are all categorical data.

I have successfully recoded Sex into a 0= female and 1= male binary variable and also dummy coded all of my other categorical variables. I've now come to do the analysis and I'm completely clueless on how I should enter the variables into the hierarchical model and then how to interpret the results after. I've only ever conducted a multiple regression using scale predictors before. I am also only familiar with dummy variables when they are entered into a model alone without any other variables as this is as far as our teaching went regarding them. If anyone can give me any help I'd appreciate it!


1 Answer 1


Basically, you've got three kinds of predictors:

  • Binary (e.g., gender sex coded female = 0, male = 1): The coefficient is the degree to which males are higher than females on the outcome variable, holding all other variables constant.
  • Numeric (e.g., age in years): The coefficient is the effect of being one year older on the outcome variable holding all other variables equal
  • Nominal (3 or more categories; e.g., ethnicity). Here you create k - 1 dummy variables (where k is the number of categories). Each dummy category is the effect of being in that category relative to the reference category holding all other variables constant.

Once you understand how to interpret the coefficients, hierarchical regression really doesn't change anything. Hierarchical regression is really the same as running a set of regression models with different predictors (they just happen to involve incrementally adding predictors).

  • $\begingroup$ Hi, thank you for your answer. I understand how the constant B value is used in the regression coefficients table as the reference to compare the dummy variables to. I don't however understand how to interpret the results when more than one nominal variable (with multiple dummy variables each) is present, how is it possible to know what the value for the reference is for each nominal variable, especially when scale variables are also present in the same model? $\endgroup$
    – Ellie
    Commented Aug 25, 2020 at 12:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.