0
$\begingroup$

What I have is a hiring analysis where we have a Dependent Variable for pre test and post test for a paired sample (same location's hiring success measure before and after treatment). I also have a host of other characteristics such as the location's region, the type of business, etc. What I aim to do is determine the best way to predict how successful hiring will be for a given region/business type etc. In my head, basically multiple regression, except that I have paired results as well. What would be the best approach to this? I typically lean on Excel/R Statistics.

Example of data

Location IV - Montana

Business_Type IV - Sales

PreSuccess DV - 46%

PostSuccess DV - 88%

Thanks for your time folks!

Improve this question
$\endgroup$
1
  • $\begingroup$ Have you considered logistic regression? Logistic regression outputs a probability of success in a binary outcome (0 = no hire, 1 = hire) $\endgroup$
    – jros
    Commented Jul 8, 2021 at 19:42

1 Answer 1

Reset to default
0
$\begingroup$

Multiple regression can work on each dependent variable, so you would have to do 2 regressions: 1 for pre-success and 1 for post-success. If you are interested in the question "can these (indepedent) variables (e.g. location, business type, etc.) predict a change in the success rate?", you will need to do another regression. First, create a new variable Y=post-pre. This operation must be pairwise. Positive values of Y indicate increased performance, whereas negative values indicate decreased performance. Then, run the multiple regression on Y.

Improve this answer
$\endgroup$

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.