I have a data-set, and the details are not crucial, but if it helps: The variables are (A) country, (B) investment, (C) campaign type, (D) external event. Obviously A is a categorical variable, B is a continuous quantitative variable, C is categorical, and D is a Bernoulli variable. I also have information about these for each year stretching back decades. I would like to know whether A, C, and D cause changes in B and exactly what that causal relationship is if it exists.

So my question is, how should I analyze this data, in broad terms? Should I use Time Series Analysis? It sounds like this should be a time-series problem but given that I've never studied the subject, I'm not certain. Is it possible to derive causal information from this data given that it's observational? It would be particularly great if there were some textbook or other resource that I could read to learn about the tools I ought to be using in a case like this.


Definitely try regression to understand some sort of relationship between the variables first.

Regression with categorical variables is relatively simple in R using lm() with factors() to convert categorical variables to leveled factors.

Here you would regress using something a model like this... $B \sim \alpha + \beta_1 A - \beta_2 C + \beta_3 D + \epsilon$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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