Most of these pilot programs run for a set number of months or years, and are then subject to evaluation. After the evaluation, the government and/or organization behind the program decides whether or not they'll try to continue or expand the program. So, what does it take to consider a pilot program successful and worthy of expansion? It usually seems pretty obvious: If the program is set to help families from falling into homeless and most of the families stay off the streets, it can probably be deemed successful. But who's to say the families were really going to become homeless? Perhaps they could have found a way to make ends meet even without the program.
The New York Times reports that the Big Apple is adding a control group to one of its homeless-prevention programs to see if people who aren't given the service really do end up homeless. The program, called Homebase, provides job training, counseling and rental assistance to help participants stay off the streets. The Homebase study will monitor a total of 400 participants, but only offer its services to 200. The remaining 200 participants who are rejected aid will be the study's control group. This structure will help officials determine if the absence of Homebase actually leads to homelessness, or if Homebase merely helps those who can already help themselves.
Looking at this study of a city program with a set of utilitarian-tinted sunglasses on, it makes a whole lot of sense. With government budgets in need of immense spending diets, figuring out what programs can be cut is more important than ever. But looking at this study with more a compassionate, slightly rosy set of sunglasses on, it seems awfully wrong. With household income levels being slashed from rampant unemployment, figuring out a way to have a roof over your head is more important than ever.
Now, I leave the tough question to you guys: How can state and local governments make sure their homelessness-prevention programs are really doing what they're supposed to be doing? You know, while remembering that the participants are people, not just test subjects?
