I am currently looking at the effect of a policy change on some states in the US that implemented the change and some that didn’t with the outcome being the proportion of people of a certain age group receiving vaccinations before and after the policy change. The policy change occurred at the same time for all states in 2010. I have data for most states ranging from 2006 to 2014 so I am interested in seeing if the policy change had an effect. I have set up my regression to account for time and state effects like so:

Y = B0 + B1(Treatment) + B2(yeardummies) + B3(state)

Where
-Treatment is 1 if the state was treated and the data point occurred after the intervention (commonly noted as an interaction term time*treat in 2 group, 2 period models where Treatment is 1 if the state is treated and time is if the data point occurred after the intervention)
- Yeardummies is a categorical variable of years where the data points are from 2006, 2007, 2008 etc…
- State categorical variable of state where the data points are from “Alabama, Arkansas etc…”

After running the regression, the coefficient of the Treatment term, B1 is 1.5. Ignoring the significance of the p-value is the correct interpretation of this that for every observed point after the intervention, the difference in treatment group and control due to the intervention is 1.5? Say in the baseline period the treatment group was already 5% higher. After the treatment it is now 6.5% higher at every post observation point?


Thanks to everyone for reading this!