combination of components that is effective. In addition, longitudinal follow-up of participants is difficult in general, but is particularly challenging in reproductive health programs because of confidentiality issues.
Another problem faced in many program evaluations is that outcome measures are limited to self-reported sexual activity and contraceptive use. Such reports may be unreliable, but there are often no alternative outcome measures available, save the most conspicuous consequences such as sexually transmitted diseases (STDs) and pregnancy. Even these obvious outcome measures can be difficult to assess precisely. For example, although births can be verified through the vital registration system, there is no universal system for reporting abortions or miscarriages, a fact that leads, among other things, to chronic problems in documenting the actual number of abortions performed annually in the United States, as discussed in Chapter 2.
These considerations argue in favor of evaluating only a few large, multisite, model programs relying on experienced evaluators having resources sufficient for the task. Stahler and DuCette (1991) suggest that individual programs should focus attention on process evaluation (i.e., the careful collection of data on client characteristics and service utilization) and that third parties should undertake well-funded impact evaluations (i.e., outcomes and long-term follow-up) of various program models that target different subpopulations.
Evaluation of local programs has also been impeded by the prevailing societal environment. The past 10 or 15 years has not been an era hospitable to research that might be seen as sex-related and therefore controversial. For example, very little survey and ethnographic research on sexuality has been done in the past two decades. Not only is it controversial politically to conduct research on sex-related issues, but involving adolescents in such research, particularly without parental consent, can raise legal issues as well.
During the 1980s especially, the federal government severely curtailed systems of data collection that had been used to monitor a wide variety of programs related to fertility, such as the national family planning reporting system. The view seemed to be that because such programs were seen by some public officials as objectionable, it was best to down play or ignore them altogether by, among other things, collecting little information on their activities or effects. Thus, the fact that only 23 programs met the committee's evaluation criteria may reflect more the political climate within which pregnancy-related programs have recently operated than a disinclination among program leaders to evaluate their activities. In addition, during the 1980s, the withdrawal of much federal funding from all but abstinence-only programs may have had a chilling