# Optimal regression model for categorical independent variables and continuous dependent variable

I have a continuous dependant variable (engagement rate) and two categorical independent variables (type and type2) with 4 categories each, trying to regress the relationship. I want to know what is the optimal way of doing this, is it linear regression or logistic regression or some other way, was hoping for some insights into this.

Dependent variable

• Engagement rate

Independent variable

• Type(Competition, Promotional, Message, Recruiting)
• Type2(Video, Photo, Link, Album)

## 2 Answers

Have you tested to see if engagement rate is normally distributed? It's probably easiest to check this by plotting the data in a histogram or alternatively by using the shapiro wilk test for normality.

Assuming that your data is normally distributed, a two-way anova should do the trick.

• How would you use the Shapiro-Will test? What if the p-value is low? What if the p-value is high?
– Dave
Commented May 12, 2021 at 12:32

So I assume your rate is between 0 and 1. You can use beta regression for variables that are continuous between 0 and 1.

• Yes its percentage and its very low values, from (almost zero) to around 8%, thanks for the reply! Commented May 12, 2021 at 12:22