Is it acceptable (publishable) to run a multilevel model with only level-2 and level-3 predictors? Example: Looking only at the effect of school resources, size or location, and at the district level (budget etc.). The dataset does not contain individual level variables (e.g. gender, SES etc.)
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
Yes, it's acceptable. Sometimes you don't have level 1 predictors. If you don't have them, you can't use them.
answered Sep 7, 2018 at 20:11
$\begingroup$
$\endgroup$
1
For whatever reason, the bot bumped this post. While Jeremy is correct, I would also add that while there may not be level-1 predictors or the data for it, in many cases, aggregate data exists at higher levels. E.g., SES at level one might be captured by median income, %male, or %white for a school or for a district and would still be relevant. So while it is not necessary to have level-1 predictors, you should consider adding aggregate predictors. Though they may not be a perfect substitute, they would be the next best option.
-
1$\begingroup$ Id forgotten about this answer. Thanks for the clarification. $\endgroup$ Commented Jun 3 at 14:15