Hi!
I have a repeated measures dataset (180 days of data x 60 neighbourhoods = 10,800 data points). I want to analyze the interactive effect of neighbourhood census-level data (e.g., neighbourhood median household income) by a daily count (e.g., # of rainy days in the city) on a count DV (e.g., # of complaints).
I believe it should be a negative binomial regression, as my DV is a count variable and the variance in the DV > M. But I'm not sure whether this would require MLM or some sort of repeated measures approach... Any suggestions would be greatly appreciated!
Not sure what you mean by MLM, but if mixed/multilevel model, then this is a common repeated measures approach (so not an OR question). You would need to decide how to deal with responses near each other being correlated. By interactive effect, I assume you mean how neighborhoods that are next to each other "interact." Can you say more what you mean by this? Do you just mean taking into account that nearby neighborhoods will be similar, of do you want something like complaints at t1 in neighborhood 1 predicts complaints at t2 in neighborhood 2? The rain effects are presumably just an association (so if it rains at t1 in n1, it is likely also raining in neighborhoods next to it)?
Negative binomial should be ok, and it's more user-friendly than a mixed poisson model, which is the alternative.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression