Information on reproductive attitudes is now routinely collected in fertility surveys in developing countries, and has become very important for understanding fertility behaviour. The quality of this information, however, is rarely assessed, partly due to lack of necessary data. In this paper, the recently completed panel survey in Morocco by the Demographic and Health Surveys (DHS) was used to investigate the consistency of reporting of ideal number of children, reproductive intentions and the planning status of the last birth. The validity of reproductive intentions for subsequent fertility behaviour was also examined. The findings indicate that the three measures of reproductive attitudes are subject to different degrees of measurement error. The measure of reproductive intentions is the most consistent of the three, followed by ideal number of children. The latter is much more consistent at the aggregate level than at the individual level. The reliability of the planning status of the last birth is marred, mainly by an unwanted child being reclassified as 'wanted' with time since the birth of the child. The reproductive intentions of women were also found to be positively related to their subsequent fertility.
PIP: This study examined the validity of responses on ideal family size, reproductive intentions, and planning status of the last birth in Morocco. Data were obtained from a sample of 1664 married women aged 15-46 years who were originally interviewed in 1992 and reinterviewed in 1995 for the Demographic and Health Survey. Findings indicate that family size desired was a consistent measure at the aggregate level, but was inconsistent at the individual level. For example, the mean number desired was 3.85 in 1992 and 3.82 in 1995. However, only 36% gave the same numerical response in both surveys. The number with different responses were distributed equally between women desiring more or less children. 30% differed by at least 2 children. Education and urban residence were directly associated with consistent responses. Planning status of the last birth had serious inconsistencies regardless of level. Unwanted births tended to be rationalized as wanted births after the fact. This tendency increased with time since the birth of the last child. Of the 140 children reported as unwanted, only 38% were reported unwanted in 1995. The most valid measure was reproductive intent, which was especially consistent at the individual level. Desire to have more children was related strongly to the likelihood of having another birth in the 3 years between surveys. Desire to stop childbearing was significantly, negatively related to having another birth in the 3-year interval. Urban women who were uncertain were more likely to behave like stoppers, while rural women who were uncertain were more likely to behave like childbearers.