Dear all, I am planning to analyze the effect of a change in legislation on behavior in different countries. All countries of interest adapted the same legislation at different points of time (in different years). I now want to investigate whether changing the legislation had an effect on certain behaviors (e.g. crime rate) in the respective country. However, there are many other factors that might influence the behavior of interest that I would like to control for. If possible, I would also like to assess an effect across all countries. The data I have available is the absolute crime rate per year for at least 3 years before and 3 years after the change in legislation. How would one go about to answer this question? Is there something like a "standard procedure" that is being used for example in political sciences?
The first idea that came to mind was conducting a time series anaylsis.
Thanks for your help!
You want to use pre-post comparisons, treating the legislation as an intervention. Look up quasi-experimental design strategies. You will want to make the # of crimes into rates (divide crime by population; criminologists often then multiple by 100,000, so you have crimes per 100,000 people). Also, read cross-national research to find the major variables you need as controls (see, for example, studies by William Pridemore).
Now, these comparisons could get complex. For example, let's say the change in law reduces crime in some places on average. That's doesn't mean the effect lasts -- it could be large in the first year, and then dissipate and disappear. Also, since crime is going down in many places, and the decline in crime is so widespread across countries that it cannot be due to certain things (e.g., people didn't change biologically, and given that psychological theories state that personal is rather fixed, those kinds of measures are not going to predict changes in crime over a short period of time). So, you have to come up with other things that are changing.
This study sounds very interesting.
I hope you expand it to include things like:
-Alterations in behaviors, crime rates, divorces, etc.
--- After women receive the vote
--- After the voting age is lowered
--- After abortions become legal
--- After old-age pension or Social Security programs become available
--- After other social welfare programs are adopted.
etc.
The data period 3 years before and 3 years after--- is too small to engage a conventional pre- and post T Test. All T tests assumed that the data follows normal distribution. Small data set may not allow us to see normality or may seem unreasonable to so assume.
Do simple 2x2 chi squared test. See 100 Statistical Tests by Gopal Kanji linked below for details on how to do the test (Test 40; p. 86).
==========================
BEFORE AFTER
EFFECT ? ?
NO EFFECT ? ?
==========================
If you are looking into many factors and wants to test for the effect of legislation on more than one factor, use Test 41 (Kx2), see p. 86.
LINK:
https://www.pdfdrive.com/100-statistical-tests-e97177.html
Thank you all very much for your interesting replies! I will read into them and come back with further questions (if necessary)!
- Badges
- Science topic
- Similar topics
- Political Science