2
$\begingroup$

A common ask I get is to compare demographics for 2 businesses. However, the data is nested (hierchichal). Each business has a unique set of locations, and the customer data comes from each location.

So, how can we compare demographics while also controlling for location? I considered a multilevel generalized linear model, but there is no clear outcome variable here.

How do y’all handle this?

Improve this question
$\endgroup$
4
  • 1
    $\begingroup$ How many demographic variables do you have ? Indeed a multilevel (mixed model) does spring to mind. You could try a multivariate multivariable mixed model or several univariate multivariable mixed models (making sure to handle multiple testing appropriately) $\endgroup$ Commented Oct 2 at 22:45
  • 1
    $\begingroup$ Also look out for answers by Ben Bolker, Erik Ruzek, Shawn Hemelstrand, EdM, Dimitris Rizopoulos, and Peter Flom. $\endgroup$ Commented Oct 3 at 0:04
  • 1
    $\begingroup$ Those are on CrossValidated, the much bigger sibling of this Data Science Stackexchange site (in case you were unaware), on the mixed-model tab,. I also have answers on the same mixed-model tab over there over there. $\endgroup$ Commented Oct 3 at 0:09
  • $\begingroup$ @RobertLong, I'll definitely take a look. However, it seems like you are suggesting I treat the location as another predictor variable? In my understanding, this would normally be treated as a 3 level model (lv 3 = the two businesses, lv 2= the locations for each business, and lv=1 marginal effects). Please correct me if I misinterpreted. $\endgroup$ Commented Nov 26 at 5:52

0

Reset to default

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.