Fertility Preferences

CHAPTER 6

FERTILITY PREFERENCES

This chapter addresses three questions that allow an assessment of the need for contraception. Does the respondent want more children? If so, how long would she prefer to wait before the next child? If she could start afresh, how many children in all would she want? Two further issues are examined. To what extent do unwanted or mistimed pregnancies occur? What effect would the prevention of such pregnancies have on the fertility rates? Bearing in mind that the underlying rationale of most family planning programs is to give couples the freedom and ability to bear the number of children they want and to achieve the spacing of births they prefer, the importance of this chapter is obvious.

Interpretation of data on fertility preferences has always been the subject of controversy. Survey questions have been criticized on the grounds that answers are misleading because: a) they reflect unformed, ephemeral views, which are held with weak intensity and little conviction; and b) they do not take into account the effect of social pressures or the attitudes of other family members, particularly the husband, who may exert a major influence on reproductive decisions. The first objection has greater force in non-contracepting societies where the idea of conscious reproductive choice may still be alien; preference data from these settings should be interpreted with caution. In societies with moderate to high levels of contraceptive use, greater interpretive weight can be attached to the findings. The second objection is correct in principle. In practice, however, its importance is doubtful; for instance, the evidence from surveys in which both husbands and wives are interviewed suggests that there is no radical difference between the views of the two sexes.

The inclusion of women who are currently pregnant complicates the measurement of views on future childbearing. For these women, the question on desire for more children is rephrased to refer to desire for another child after the one that they are expecting. To take into account the way in which the preference variable is defined for pregnant women, the results are classified by number of living children, including the current pregnancy as equivalent to a living child.

The table allows the potential need for contraceptive services for spacing as well as limiting births to be examined. Until recently, concern for providing appropriate contraceptive methods to couples who wish to have no further children has overshadowed contraception for child spacing purposes. The interest in spacing has been reinforced by recent evidence that: a) short birth intervals are harmful to the welfare of children and mothers; b) large numbers of couples wish to postpone childbearing by using contraception; and c) there is a potential demand for contraception for spacing births in some countries where demand for limiting family size has not yet emerged.

In Tables 6.2.1 and 6.2.2 the percentage of respondents who want no more children is shown for each parity by selected background variables. This tabulation provides information about group variations in the potential demand for fertility control.

The working tables (Table 6.2.1working table: Number of currently married women by number of living children and Table 6.2.2working table: Number of currently married men by number of living children) should be consulted to determine whether the denominator for each cell requires that the percentage in the cell be suppressed (less than 25 unweighted cases) or placed in parentheses (25-49 unweighted cases).

Thus far in this chapter, interest has focused on the respondent’s wishes for the future, implicitly taking into account the number of sons and daughters she/he already has. In ascertaining the total ideal number of children, the respondent is required to perform the more difficult task of considering abstractly and independently of her/his actual family size the number of children she/he would choose if she could start again.

There is usually a correlation between actual and ideal number of children. The reason is twofold. First, to the extent that respondents implement their preferences, those who want larger families will tend to achieve larger families. Second, respondents may adjust upward their ideal size of family as the actual number of children increases (i.e., rationalization). It is also possible that respondents with large families, being on average older than those with small families, have larger ideal sizes because of attitudes they acquired 20 to 30 years ago.

Despite the likelihood that some rationalization occurs, it is common to find that many respondents state ideal sizes lower than their actual number of surviving children. The use of ungrouped variables in Table 7.4 permits the classification of respondents at each parity into three categories: ideal size is greater than actual size; ideal size is less than actual size; ideal size equals actual size.

The second category is of particular interest, because it is an indicator of surplus or unwanted fertility, which is also a topic in a later table.

The mean should not be shown in Table 6.3 if more than 30 percent of respondents have a non-numeric response.

Mean desired family size by age should be discussed here prior to interpretation of the other differentials.

Greater interpretive emphasis should be placed on the results for women age 2029 than for women under age 20 or age 30 and over. For some subgroups, women under age 20 may be a small selective group whose views are atypical; while for older women, the dangers of rationalization are greater.

If more than 30 percent of women have a non-numeric response to ideal number of children, Table 6.4 should not be used.

Women are asked a series of questions for each child born in the preceding five years and any current pregnancy to determine whether the particular pregnancy was desired at the time (“planned”), not desired at the time but wanted at a later time, or unwanted at any time. These questions form a potentially powerful indicator of the degree to which couples successfully control childbearing. In addition, the data can be used to gauge the effect of the prevention of unwanted births on period fertility.

The questions are demanding. The respondent is required to recall accurately her wishes at one or more points in the last five years and to report them honestly. The danger of rationalization is present; an unwanted conception may well become a cherished child. Despite these potential problems of comprehension, recall and truthfulness, results from previous surveys have proved surprisingly plausible. Respondents are clearly willing to report unwanted conceptions, although some postfactum rationalization probably occurs; therefore the result is probably an underestimate of unwanted fertility.

In DHS surveys, these retrospective questions are asked independently of the questions on the desire for more children and total desired family size and have not been crossedited at the data processing stage. Investigation of consistency of attitudes at the individual level is thus possible but is not attempted in this report. However, broad consistency at the average or aggregate level between the total ideal family size and actual fertility and wanted fertility can be examined.

Table 6.5 is a birthbased rather than a womanbased table. It provides a useful indicator of the degree of successful reproductive control exercised by couples in the recent past.

Wanted fertility rates are calculated in the same manner as the conventional agespecific fertility rates presented in Chapter 4, except that births classified as unwanted are omitted from the numerator; the remainder can be cumulated to form a total wanted fertility rate (TWFR), which is analogous to the conventional total fertility rate (TFR). The total wanted fertility rate provides another indicator of fertility aspirations and may be interpreted as the number of wanted births that a woman would bear by age 50, if she experienced the wanted fertility rates observed for the past three years.

The Lightbourne method of calculating a "wanted" birth is used for this table: a birth is considered wanted if the number of living children at the time of conception was less than the ideal number of children reported at the time of the survey.

Wanted fertility rates express the level of fertility that theoretically would result if all unwanted births were prevented. Comparison of actual rates with wanted rates indicates the potential demographic impact of the elimination of unwanted births. This calculation is highly relevant for countries that have official policies to reduce the birth rate and thus the rate of population growth.

There is a difference between ideal family size and the wanted fertility rate in that the wanted fertility rate takes observed fertility as its starting point and can never be larger than the actual TFR; ideal family size can be and often is larger than the number of children born. This characteristic of the wanted fertility rate has an advantage and a disadvantage. It may be the more realistic measure, because it takes into account the fact that fecundity impairment prevents some women from having wanted births and from achieving their desired family size. But it has the disadvantage of interpretive complexity and, like any period measure, is vulnerable to temporary influences on the level of recent fertility.

If more than 30 percent of women have a non-numeric response to ideal number of children, Table 6.6 should not be used.

1