Question

In 2001, the federal government enacted a law that forbade any student with a drug offense from receiving financial aid. Lovenheim and Owens (2014) study the effect of this law on college enrollment behavior by examining how enrollment among those with convictions changes relative to those without convictions in a difference-in-difference analysis. A difference-in-difference analysis takes two successive differences. The first difference is the difference between the pre and post-treatment values of a group that underwent treatment and the difference between the pre and post-treatment values of a group that did not undergo treatment. Once these two values are calculated, the difference between these differences is computed.

a) Using the provided information, calculate the difference-in-difference estimate

	No Conviction	Conviction
Pre-Change	0.623	0.358
Post-Change	0.652	0.269

b) Does the fact that college enrollment was rising throughout the United States for all students over the period cause a problem for their study?

Answer 1

A. This has been calculated below

	No Conviction	Conviction	Difference
Pre-Change	0.623	0.358	0.265
Post-Change	0.652	0.269	0.383
Change	0.029	-0.089	0.118

The DID estimate is .118.

B. Yes this fact causes problems for the study as the factors used in study are not constant. We now need to know the increase across different types of student too to estimate the correct impact of the new law. If rates of all students were rising, we need to know how much were the rates rising for them to adjust for that organic increase in rates.