1

I am working on a project where 8 municipalities have participated in a program, in which they received consultancy on how to restructure the social sector. As a part of the program the municipalities have been able to follow their development through seven different parameters. My job is to analyze the progression of each municipality but also make a comparison between the progression of the municipalities. However, the municipalities did not start in the program at the same time, and they did not have the same starting point (some municipalities had already been working on the restructuring, before they participated in the program).

My question is therefore: How do I compare the progression between the municipalities and also take their difference in starting points in to consideration? My own thoughts are that I need to establish a baseline of some sort, but how do I do this correctly?

BeeBee
  • 33
  • 5

1 Answers1

1

If you have repeated measurements you can fit some form of mixed effects model for each of your seven variables separately. To do this you need for each observation: which municipality, which time, and whether before and after the introduction of the restructuring. I would suggest making time zero the timepoint when the restructuring took place so you line them all up with negative times before and positive after. The fit the model with time and before/after and a random effect for municipality. You can then add an interaction between time and before/after to see if the rates of change vary.

One thing to think about here is when you expect the intervention to take effect and how. Will it be immediate or come in gradually over time? Will it be a sudden shift or will it continue to take effect over time?

mdewey
  • 17,806