I have a survey data which includes both my dependent and independent variables. I will do a principal component analysis to get rid of the multicolinearity problem. Then, I want to do a regression analysis. My question is, how do I include control variables in the model if they were not a part of the questionaire? I've heard from a colleague of mine that he inserted some control variables as dummy variables, but I am not sure how to technically do that in the process of plotting the regression.
Could you please give me some insight into the topic?
Thank you.
@Ivana Jovanovska, usually in a regression model we control for socioeconomic, demographic characteristics of subjects to see the association between outcome (dependent variable) and exposure (main independent variable). However, you have to have those variables in your dataset. In addition, types of regression will depend on types of outcome variable. For example, logistic regression for binary outcome variable. I assume you already have known this.
Just a suggestion here. I wouldn't include a variable that you haven't measured or asked about in your study. HTH would you know what to put in anyway? Best wishes David Booth
To account for the impact of certain factors on a dependent variable, just include them in the model. As an example, suppose you do a regression on a dependent variable y and an independent variable x. You believe that z has an effect on y as well and want to account for this impact.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size