Investigating the effects of changes in legislation

Dear all,

I am planning a study to investigate the effect of a change in the law on crime rates in different countries.

Each country introduced the same change at different times (different years).

As for the data, for each country I have the annual crime rate for at least 3 years before and after the change of legislation and some control variables.

Is there a good statistical technique to find out whether the crime rate has changed significantly since the introduction of a particular law? In addition, I would like to calculate a kind of overall effect across all countries.

In some studies I have seen so far, the values of the years before the change and the years after the change have been averaged and a t-test has been calculated, but in my opinion this does not take into account the many control variables that also have an influence on the crime rate.

Maybe you can help me?



No cake for spunky
You can try segmented regression. It will generate a test if the slope changed after the intervention. But it will be difficult if the change took a long time (gradual change). You can control for various factors as with any regression as long as you can measure it.

ARIMA style intervention analysis is similar, although very complex in practice. See Hays et el for a discussion of it from SAGE. I would warn you that I personally could not figure out how you would use this approach in practice, but you might understand it better than I. :)