reported that total public expenditures on contraceptive services (including Medicaid, Title X, and state funds) per woman at risk of unintended pregnancy had no apparent effect on adolescent pregnancy rates, but did seem to have a variable effect on birth rates, differing somewhat by race and age. Singh (1986) also found lower teenage birth rates in states with higher proportions of adolescents served in family planning clinics, but did not find an association with lower pregnancy rates. Similarly, Anderson and Cope (1987) found that publicly funded family planning programs in the United States could be linked to lower birthrates; this analysis did not assess effects on pregnancy rates. Olsen and Weed (1986) also concluded (using 1978 data) that overall enrollment in family planning clinics is associated with reduced teen birth rates, but suggested that such enrollment may also be associated with increased teen pregnancy rates. In a subsequent analysis, these same investigators (Weed and Olsen, 1986:190) seemed to soften their earlier finding by concluding that "greater family planning program involvement does not result in a reduction in teenage pregnancy rates." It is important to add, however, that all of these analyses have some unresolved methodological problems that suggest their conclusions should be viewed carefully. None of them, for example, has been able to control for varying levels of sexual activity, nor do they factor in such other dynamics as the growing use of condoms—widely available outside of organized clinic systems—to prevent pregnancy.3

One of the most recent such investigations is that of Meier and McFarlane (1994). They conducted a state-level analysis to measure the effectiveness of publicly funded family planning during the mid-1980s in influencing a variety of outcomes. The analysis focused on several indicators of effectiveness: the state-level abortion rate, the age-specific fertility rate for adolescents, the incidence of low birthweight and premature births, the proportion of pregnant women receiving late or no prenatal care, and the neonatal and infant mortality rates. The principal measure of public funding was the level of family planning


As noted, these studies rely on state-level data, and this choice of analysis unit has both advantages and disadvantages. The disadvantage is that the outcome indicators of effectiveness are not directly linked to program activities or to the behavior of program clients. Viewed from another perspective, however, the use of state aggregates may be a potential strength. As is argued in Appendix G, the evaluation of any given program is greatly complicated when a number of programs coexist in the same geographic area. For instance, the information provided by one program may encourage a potential contraceptive user to seek out the services that are provided by a different program or by the private sector. These cross-program and spillover effects cannot be captured using program data alone; they require histories of program contacts and service utilization on the part of clients. A state-level analysis is implicitly concerned with the net effects of all publicly funded activities, and may therefore provide a truer picture than would emerge from consideration of any one program in isolation.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement