Social science research often seems to raise at least as many questions as it answers, perhaps because so many types of factors—many of them unmeasurable—can affect human behavior. The challenge for the investigators, then, is to figure out what their findings mean in the real world. Several articles in this issue of International Family Planning Perspectives look at situations in which the investigators' conclusions were based not simply on odds ratios from a multivariate analysis, but on interpreting them in the context of other available data and the setting from which they were drawn.

In the lead article, Juan Schoemaker uses data from the most recent Indonesia Demographic and Health Survey to look at the influence of poverty on use of modern contraceptive methods among married women [see article]. In this study, extremely poor women (those in the bottom socioeconomic decile) had the lowest level of contraceptive use and the highest level of unmet need, which would seem to indicate that these women wanted to limit their family size but could not afford contraceptives. A multivariate analysis that controlled for social and demographic characteristics supported this interpretation, but also suggested that attitudes might play a role. Indeed, further analysis showed that lower percentages of extremely poor women than of better-off women approved of family planning, believed their spouse approved and considered their last pregnancy mistimed or unwanted. Their mean ideal number of children was higher, as was the percentage with two or more children who wanted another child. The author concludes that increasing contraceptive use among poor women will require changing attitudes toward smaller family sizes and family planning use as well as increasing access.

Soon after the government of Bangladesh transferred family planning services in rural areas from doorstep delivery to static community clinics, levels of fertility—which had been decreasing since the 1970s—stabilized. The timing suggests that—in a culture that traditionally restricts women's movements outside the home—the change in service delivery left many women without access to contraceptive supplies, thus stalling the decline in fertility. As Alex Mercer and colleagues demonstrate, however, things aren't that simple [see article]. By the time the government changed hands at the end of 2002, they note, only about half of the planned clinics were operational, and none of the country's subdistricts had fully implemented the clinic system; the subsequent administration switched back to doorstep delivery of contraceptives. Their analysis, which examines patterns of contraceptive use in two rural subdistricts between 1998 and 2002—thus encompassing periods before and during the transition—shows that in areas of these subdistricts that had operational community clinics, three-quarters of women used one at some time and substantial proportions obtained contraceptive supplies from the clinics. Moreover, contraceptive prevalence increased in one subdistrict and remained stable at a high level in the other during the short period when community clinics were in operation. The authors conclude that women in these areas do not appear to have been dependent on home delivery for access to contraceptives.

When a couple in Sub-Saharan Africa has no children, the wife is likely to bear the blame. Because of the high value placed on fertility, a woman who does not bear children early in her marriage may be beaten, divorced or forced to share her husband with other wives. But it appears that having a large family does not always bring a woman approval. According to an analysis of intimate partner violence in an urban district of Tanzania by Laura Ann McCloskey and colleagues, both women who had trouble conceiving and those who had five or more children had an elevated risk of being physically abused, even after the effects of demographic characteristics were accounted for [see article]. The authors comment that with the ideal number of children decreasing in urban areas of Tanzania, having more than four children places a couple outside of family-size norms and may create economic strains.

Most studies in developing countries have found that men are the main decision makers in their households, even on issues such as contraceptive use and family size. These studies, however, rely on reports from women only. Ilene Speizer and colleagues use two nationally representative surveys—one of men and one of women—conducted in Honduras in 2001 to look at who respondents thought should make reproductive decisions and who in their household actually made them [see article]. The samples for the study were restricted to men and women who were in consensual unions or married (but not to each other). Interestingly, the majority of women and men believed that partners should make joint decisions on family size (58% and 67%) and on contraceptive use (54% and 63%); similar proportions of women and somewhat larger proportions of men said this was the case in their own household. Still, when respondents' attitudes toward male decision making were compared to their reports of decision making by the man in their household, one major difference appeared between male and female respondents: Some 15% of women said that men should make neither family size decisions nor contraceptive use decisions on their own, but that the man in their household did make one or both decisions; only 5% of men gave that response.

Abortion is legally restricted in the Philippines, yet thousands of women are hospitalized each year for treatment of abortion complications. Using indirect estimation techniques, Fatima Juarez and colleagues calculate that more than 473,000 women had induced abortions in 2000, an increase of more than 70,000 over the estimate for 1994 [see article]. Further calculations suggest that this trend reflects an increase in the proportion of contraceptive users who are relying on traditional methods—which have high failure rates—and a resulting increase in the proportion of pregnancies that are unwanted. —The Editors