fertility rate can be estimated by deleting recent births to women who report an ideal family size lower than their actual number of living children (Cochrane and Sai, 1993; Bongaarts, 1990). An alternative is to construct a "synthetic estimate" of desired total fertility rates, summarizing estimates of proportions of women at each parity who report that they want no more children. These indirect methods do not allow estimation of mistimed (as opposed to entirely unwanted) pregnancies, and they require a hypothetical recasting of respondents' lives that may not be especially meaningful. But they avoid one weakness of the direct questions about specific pregnancies: "[U]nderreporting is apparently common and is presumably caused by a reluctance of women to classify their offspring as unwanted" (Bongaarts, 1991:223).
Since desired family size has fallen in almost all developing countries for which it has been measured, it has been conventional to focus on only one aspect of failure to achieve reproductive goals, namely, unwanted childbearing. Yet as Kingsley Davis (1967) noted three decades ago, "family planning" literally construed should include the notion of couples planning to have a family. Infertility can lead to loss of social status, divorce, and other negative consequences, as well as being a cause of tremendous unhappiness.
Larsen and Menken (1989) provide useful evidence on the prevalence of infertility from a combination of demographic surveys and microsimulation of populations with varying amounts of deliberate fertility control. Larsen (1994) estimates that the proportion of women sterile by age 34 in 17 sub-Saharan African countries varied from a low of 11 percent in Burundi to a high of 31 percent in Cameroon. She concluded that "the true prevalence of sterility in sub-Saharan Africa is so substantial that it ceases to be a merely individual problem and has become a public health issue" (Larsen, 1994:469). As in previous reviews, Larsen found great geographic variation in apparent infertility, which she plausibly ascribes to geographic variation in the incidence of reproductive tract infections. In Chapter 3 we discuss the needs for programs to control sexually transmitted diseases (STDs), which is the most effective way to prevent infertility in the countries most affected.
A crucial question in all societies is whose fertility intentions count in deciding if a birth is "intended." The potential mother is obviously most directly affected, and any definition of intention that did not include her wishes and interests would be unacceptable. Potential fathers also have, or are expected to have, a lot at stake in fertility decisions. A widely shared ideal would call for good spousal communication and socially