Perhaps an element of that...but they are talking about changes in rate at the specific schools.
So here are my many issues with this study... Long term study (1992-2014), 'control' group is 20-24 year olds, 'teen' includes 13-19 year olds and measured pregnancies are those as measured of the general population which includes many urban anchor baby cities (which of course also happen to have the largest concentration of non-counseling distributions).
If this study, say, tracked birth rates of those in school who likely received condoms from the school, that would be useful data. Heck, if you simply did the sensible thing and excluded large urban centers where many arrive illegally to give birth to their 'citizen' babies and collect the benefits, that might give some indications as to usefulness of the program.
But measuring general population, most especially including heavy immigrant anchor baby destinations, extending 'teen' out to 19 year olds who aren't in school -- well, you have something that is nifty for the cocktail circuit but useless for actual management information.