In the criminal justice system, one is presumed innocent until proven guilty. Unfortunately, the same logic is misapplied in the social service sector, where social service initiatives are presumed effective until proven pointless.

Part of the problem is that performing social service evaluations is complex and can be costly. Agencies don't always have the expertise or funds to perform proper evaluations. The bigger problem, though, is that agencies don't necessarily have an incentive to conduct evaluations in the first place because of a common and wrongheaded assumption that if a social service agency is doing a nice thing that sounds intuitive, then it must be working. This is a dangerous assumption that is perpetuated in media portrayals of social services.

Take for example a recent article in the LA Times about a pilot program in the LA County jail system to distribute condoms to inmates. The intended purpose of the program was to reduce the spread of HIV amongst inmates. On face, it sounds like a compelling idea. We know that the use of condoms reduces the spread of HIV and other STD's in the general population.

However, the jail environment is different. There may be a stigma effect in the jail that reduces substantially the number of inmates willing to opt-in to the condom program. Also, since the pilot program only supplied a limited number of condoms, it’s possible that inmates reused the condoms, used them improperly, or even engaged in riskier sexual behavior on account of the presence of condoms.

A prison is a perfect place to do a meaningful evaluation. The jail has medical information on inmates and could easily track which inmates opted-in to the voluntary condom program. Also, since the pilot was not conducted at all LA County jails, the non-participating jails could act as control groups.
While the article mentions a group of health advocates who said

that a successful review of that program could lead to widespread distribution of condoms in prisons throughout the state

the evaluation is at best framed as a necessary formality to confirm what is believed to already be known. Instead of casting a critical eye on the effectiveness of the condom program at reducing HIV transmissions, the author implicitly presumes the effectiveness of the pilot program by failing to challenge at any point the assertion that the condom program reduces HIV transmissions amongst inmates.

My point here is not to say that the pilot was ineffective at reducing the spread of HIV. Rather, my point is that without an evaluation, we simply don't know. If we want to get real about helping people, we have to know whether we really are helping people.