The effects of individual- and community-level knowledge, beliefs, and fear on stigmatization of people living with HIV/AIDS in China.

This paper assesses the effects of individual- and community-level knowledge of HIV transmission, beliefs and fear of infection on individual attitudes toward and willingness to interact with people living with HIV/AIDS in China. Data are from a sample survey of 12,270 men and women age 15-49 from seven provinces/municipalities. The survey was conducted by the Population and Family Planning Commission in December 2003. Multilevel regression analyses show that stigmatization of people living with HIV/AIDS is strongly associated with community-level HIV knowledge and fear, even after taking into account HIV knowledge and fear at the individual level. The findings suggest that individual attitudes are partially shaped through social learning and social influence. Public-health interventions aimed at eliminating stigmatization of people living with HIV/AIDS will need to pay special attention to social and community influences while continuously working to enhance HIV knowledge and reduce inaccurate beliefs and fear.

Risk-taking behavior among high school students in South Korea.

In a national sample of high school students, 43 per cent had drunk regularly, 32 per cent had smoked regularly, and 11 per cent had experienced sexual intercourse. Boys were more likely to have experienced these behaviors than girls. Going to school in Seoul, not having two parents in the family, and having an experience of living away from immediate family were associated with higher livelihood of experiencing these behaviors. Students who had friends with sexual experiences were much more likely to experience sex themselves than those without.

Are the WHO / UNICEF guidelines on breastfeeding appropriate for India?

PIP: Using data from India's first National Family Health Survey (NFHS-1), this study examines breast-feeding practices and their effect on infant mortality in India. Conducted in 1992-93, NFHS-1 collected complete birth histories from a probability sample of 89,777 ever-married women aged 13-49 years. In the study, the breast-feeding status of infants at ages 2,7, and 11 months and the effects of exclusive and nonexclusive breast-feeding on mortality at ages 1-14 months were estimated using logistic regression models to NFHS-1 data. Separate analyses focus on three groups of states, distinguished according to levels of infant mortality, namely, high-mortality states, medium mortality states, and low-mortality states. Results show that breast-feeding reduces mortality in the first few months of life. The beneficial effects extend to older infants when combined with supplemental foods. At the same time, the results indicate that the effect of breast-feeding in lowering infant mortality is smaller the higher the level of mortality. The beneficial effects of both exclusive and nonexclusive breast-feeding are weakest in states where infant mortality is high and strongest in the medium-mortality states. Findings suggest that educational programs that promote breast-feeding with supplementation could help lower infant mortality in India.^ieng

Identifying children with high mortality risk.

PIP: This issue reports selected results from a comprehensive study of infant and child mortality based on the National Family Health Survey data. The analysis distinguishes between neonatal, postneonatal, infant and child mortality since mortality and its causes vary considerably among children of different ages. Hazard regression analysis was used to estimate the effects of each individual variable as the factors that affect infant and child mortality tend to be correlated with each other. The study involves adjusted effects of selected socioeconomic and demographic characteristics on neonatal, postneonatal, infant, and child mortality for children born during the survey. Short birth intervals have a great effect on infant and child mortality. A previous birth interval of less than 24 months increases child mortality by about 67%. Neonatal mortality is highest among children of very young mothers. Child mortality is higher for girls in all states except Tamil Nadu, Kerala, and Goa. Seven groups of children who are especially vulnerable to infant and child mortality were identified. Thus, intervention programs, such as efforts to provide supplemental nutrition and basic immunization to pregnant mothers, infants and young children need to focus on these high-risk groups. Results for many states show elevated mortality rates for girls after the neonatal period. Family health programs aimed at overall improvement in mortality levels should pay attention to providing basic health care and supplemental nutrition to girls.^ieng

State-level variations in wanted and unwanted fertility provide a guide for India's family planning programmes.

PIP: This article discusses trends in total marital fertility and wanted and unwanted fertility from eight states in India. Data were obtained from the 1992-93 National Family Health Survey for the large states of Uttar Pradesh, Bihar, Madhya Pradesh, and Rajasthan and for the more socially advanced states of Kerala, Maharashtra, Himachal Pradesh, and Punjab. The calculation of total marital fertility rates (TMFR) is based on the period parity progression ratio for 3 years before the survey. Wanted TMFR (WTMFR) indicates fertility preferences are met. Unwanted TMFR (UTMFR) indicates excess fertility. Desired fertility was 3-4 children in Uttar Pradesh and Bihar, about 3 children in Madhya Pradesh and Rajasthan, and about 2 children in the other states. Uttar Pradesh had the highest unwanted marital fertility of about 1.4 children more than desired. Other states had an unwanted fertility of about 1 child more than wanted fertility. Total fertility was 2-3 children higher in Uttar Pradesh and Bihar than in Kerala. Total marital fertility would be reduced by 20-28% if women had only the number of children wanted in Uttar Pradesh, Bihar, Madhya Pradesh, and Rajasthan. Total marital fertility would be reduced by 31-34% in Himachal Pradesh, Punjab, and Maharashtra if women had only wanted children. Unwanted childbearing was highest in Himachal Pradesh, Punjab, and Maharashtra, states that lagged behind in contraceptive use. The TMFR was highest among illiterates and declined with higher education. UTMFR varied irregularly by educational level. Populous states had high WTMFR among illiterate women. UTMFR was highest among illiterate women only in Uttar Pradesh. Findings suggest that when fertility began to decline, education was strongly related to fertility preferences. Family planning should be directed to promoting the small family and meeting the needs of women desiring small families.^ieng

A mixture model for duration data: analysis of second births in China.

In this paper we introduce a mixture model in which we combine logistic regression and piecewise proportional hazards models for analysis of duration data. The model allows simultaneous estimation of two sets of effects of covariates: one of the probability of an event and the other of the timing of the event. We illustrate the application of the model through an analysis of the effects of women's characteristics and of the acceptance of a one-child certificate on the birth of second children in China. Both factors affect the probability of having a second child, but only the acceptance of a one-child certificate has a significant and strong effect on the second-birth interval.

Is son preference slowing down India's transition to low fertility?

PIP: This report summarizes findings from the 1992-93 India National Family Health Survey among 90,000 Indian women. The study analyzes indicators of son preference in 19 Indian states and the impact of son preference on fertility decline. The authors offer some policy recommendations. Preference for sons was particularly strong in Punjab, Rajasthan, Uttar Pradesh, Bihar, and Gujarat; it was weakest in Kerala, Delhi, Assam, Goa, Karnataka, and Tamil Nadu. Son preference based on the desire to stop childbearing was strongest in Haryana and Rajasthan. Women in all states were more likely to practice family planning if they had two sons than if they had two daughters. This measure was strongest in Rajasthan and weakest in Kerala, Andhra Pradesh, Goa, and Tamil Nadu. Based on all three measures, son preference was strongest in the northern and central regions. In every state, with the exception of Tamil Nadu, women with two sons were less likely to have a fourth child than women with one son. Gender preference based on the Arnold (1985) method had a particularly strong impact on fertility in the northern states of Himachal Pradesh, Maharashtra, Punjab, and Haryana. The effect was very weak in the low-fertility states of Kerala and Tamil Nadu and in the high-fertility state of Uttar Pradesh. Gender preference tended to have the strongest effect in states with intermediate levels of fertility; however, there were wide variations. Multivariate analysis revealed that rural women had slightly higher fertility than urban women. Illiterate women had higher fertility than literate women. The influence of son preference was weakest among Muslims compared to other religious groups. The authors suggest that cultural and economic factors underlying attitudes about son preference must be changed. Social development for women is a solution.^ieng

Effects of gender, birth order, and other correlates on childhood mortality in China.

Using data from the 1988 Two-Per-Thousand Survey of Fertility and Birth Control, this paper examines the effects of gender, birth order, and other correlates of childhood mortality in China. Controlling for family-level factors, childhood mortality is found to be associated with the child's gender and birth order. Among firstborn children the difference between male and female childhood mortality is not statistically significant, but among others, female children between ages 1 and 5 experience higher mortality than male children. Childhood mortality is slightly higher for children who have older brothers only than for those who have older sisters only, and it is highest for those who have both older brothers and sisters. Other factors affecting childhood mortality in China include mortality of older siblings, birth interval, urban/rural residence, mother's level of education, and mother's occupation. All interactive effects between gender and family-level characteristics are found to be statistically insignificant.

PIP: Life table analysis of birth intervals and child survival in China reveals that discrimination against females is not apparent for first born children. Discrimination against females after the first child is generally uniform among different socioeconomic groups. Attitude change and changes in societal norms are expected to have the greatest impact on elimination of gender discrimination in China. Childbearing in China follows a pattern of a strong desire for at least one son and a small family. After the birth of a female child, the probability of having a short birth interval and having a next birth is much higher. The leading causes of death among children aged under 15 years are considered preventable and a result of the quality of parental care, such as shortened breast feeding periods among female infants. Female discrimination appears to start with second daughters. Both male and female first children are valued. Tests of covariance and the observed gender difference in child mortality departing from the usual male/female ratio confirm this finding. Mortality is lowest among children who have older sisters. Proportional hazard models reveal that all of the family level factors (urban residence, mother's education, and mother's occupation) are statistically significant in explaining differential childhood mortality. Differences in childhood mortality by gender are greater for higher birth order children. The sex ratio of mortality varies by birth order and age of children. Data were obtained from China's 1988 Two-Per-Thousand Survey of Fertility and Birth Control among almost 500,000 ever married women aged 15-57 years.

[Post-nuptial coresidence of married women in Korea].

PIP: In 1960, only 28% of the population of Korea was urban. This proportion rose to 74% by 1990. In 1960, primary industry engaged 66% of the labor force; however, this dropped to only 20% by 1989. In 1960, only 14% of women participated in the labor force; this rose to 46% by 1990. During 1956-60, senior high school or higher-level education was attained by 53% of Korean women; this increased to 86% by 1966-70. Before 1960, as many as 76% of the women lived with parental in-laws immediately after marriage; however, by 1980 or later, this proportion declined to 33%. Among rural residents, postnuptial correspondence reached 80% in 1960 and decreased to 58% in 1980. Among urban wives, the rate was 61% in 1960 and 26% in 1980. 78% of wives who were 20 years of age at marriage lived with in-laws in 1960; this dropped to 49% in 1989. AMong those marrying at age 20-24, the co-residence rate was 65% in 1960 and 34% in 1980; while among those marrying at 25 or over, coresidence was 30% in 1960 and 30% in 1980. When the wife's education was less than junior high school level; 77% lived with in-laws in 1960; this proportion declined to 41% in 1980. The respective figures for junior high school level education of the wife were 61% in 1960 and 37% in 1980; and for senior high school education, coresidence rates were 65% in 1960 and 31% in 1980. Among those whose husbands grew up in a rural area, 79% coresided in 1960 compared to 39% in 1980. Among those whose husbands grew up in an urban area, 64% lived with in-laws in 1960 compared to 29% in 1980. 77% of coresident wives in 1960 had husbands or parents who decided on coresidence as opposed to 442% in 1980. Among wives who made this decision themselves, 50% lived with in-laws in 1960 compared to 32% in 1980. If the husband was the first son, 92% of wives took up coresidence in 1960 compared to 47% in 1980. However, when the husband was not the first son, only 57% lived with parental in-laws in 1960, which proportion declined to 23% in 1980.^ieng

Determinants of contraceptive method choice in an industrial city of India.

PIP: This study examines the determinants of contraceptive method choice in Jamshedpur, an industrial city in Bihar State, one of the few areas in India that enjoys a "cafeteria approach" to family planning method selection. While contraceptive prevalence in India is about 35%, Jamshedpur has a prevalence rate close to 60%. One of Jamshedpur's special programs is the Tata Iron and Steel Company (TISCO) Family Welfare Programme, which provides an array of services to both employees and nonemployees. In 1983, TISCO commissioned the International Institute for Populations Studies (IIPS) to evaluate its programs. Based on the findings of an IIPS survey of 2376 currently married women between the ages of 15-44, this study attempts to identify determinants of contraceptive method choice. For its analysis, the study used a multinomial method choice. For its analysis, the study used a multinomial logit regression, a model appropriate for studying the relationships between a number of covariates and a dependent variable (the contraceptive method used) that has more than 2 possible outcomes (female sterilization, male sterilization, condom, female temporary method, natural methods, or no method). The survey examined the following covariates: religion/caste, mother tongue, husband's occupation, place of employment, age at consummation of marriage, husband's education, wife's education, marriage duration, survival status of last child, number and sex of living children, and attitude variables. The study found that religion, mother tongue, and educational levels are important determinants of acceptance and method choice. The number of children -- especially sons -- also affected contraceptive use, rising as family size increases. The study revealed that the Jamshedpur population did exercise choice in clear patterns when the choices were made available.^ieng

Why do Chinese women practice contraception? The case of rural Jilin Province.

This study focuses on the relationship between contraceptive behavior, family size preferences, and perceptions of the one-child policy among young Chinese women in rural areas of Jilin Province. In 1985, about 85 percent of rural married women with one surviving child were practicing contraception, although most of them reported two as their ideal number of children. Most women with one surviving child, including those with one-child certificates, were practicing contraception in response to the government campaign, while more than half of women with two or more children were doing so voluntarily. Most of the women with one child were using the IUD, whereas more than half of women with two or more children were sterilized. Through multivariate analysis of contraceptive behavior and method choice, additional factors were found to be associated with the contraceptive behavior of rural Jilin women; achievement of their ideal family size was a significant factor in the voluntary practice of contraception as well as in contraceptive method choice. Implications of the results are discussed.

PIP: Data from the 1985 Survey of Rural Fertility and Living Standards were used to analyze the contraceptive behavior of 5399 married women aged 15-60 in rural Jilin Province, China. Jilin had a population of 22,980,000 in 1985, and a high socioeconomic level and educational status. The Survey contained a multiple choice question on why women practiced or did not use contraception. First, ideal family size was analyzed by woman's age, number of living children and contraception. 85% of married women with 1 living child were contracepting, although 2 was their ideal family size. 78% of women with 1 living child, including those on the One-Child Certificate, use contraception because of the government campaign, usually with the IUD; the majority of those with 2 children were contracepting voluntarily, usually by sterilization. 2 multivariate analyses were performed: an analysis of covariates of contraception in response to the government campaign, and of contraception for other reasons (voluntary); and another analysis of covariates of contraceptive method choice for women with 1 child and for women with 2 or more. When duration of marriage was controlled, for women with 1 child, ethnicity and having at least 1 son significantly affected reasons for contraception. For those with 2 children, the ideal family size also significantly affected the reason for contracepting, i.e., whether for government campaign reasons for voluntarily. Choice of the IUD was linked with having 1 son, contracepting because of the government campaign, and acceptance of the One-Child Certificate. Having had an abortion was associated with choosing the IUD. For those with 2 or more children, having achieved ideal family size was significant for choosing sterilization. Having taken a One-Child Certificate did not guarantee choice of sterilization. Thus women may be hoping that restrictions may be further eased, as they were in 1988 for rural families with 1 daughter, so they can have another child.

Effect on famine on child survival in Matlab, Bangladesh.

"This paper examines how the level of child mortality and rate of out-migration, and their covariates were affected during the 1974 famine and the periods immediately following the famine in Bangladesh." The authors examine several variables affecting mortality and out-migration, including socioeconomic status, sex factors, age factors, educational status, and religion. Data are from the Matlab Demographic Surveillance System.

Labor force participation of married women in contemporary Korea.

"This paper presents some results from the 1986 Family Life Survey of Korea. This survey was the first that was conducted in Korea with the specific purpose of understanding women's life cycle beyond marriage and fertility....A probability sample of 3,400 households representing the whole country were selected, [including] interviews of ever-married women of ages 15 to 64.... The survey collected information on household composition, with basic characteristics of members of household including health status....marriage history, including how husbands were selected, pregnancy history including the age of children when they first left home and reason for leaving home, contraceptive use, health, and women's employment at different stages of life." (SUMMARY IN KOR)

To what extent does breastfeeding explain birth-interval effects on early childhood mortality?

This article shows that in Nepal breastfeeding almost completely explains the effects of following birth interval on childhood mortality during the first 18 months of age and partially explains the effect of following birth interval on childhood mortality between 18 and 60 months of age. Breastfeeding does not explain the effect of preceding birth interval on childhood mortality. The analysis is based on application of hazard models to data from the 1976 Nepal Fertility Survey.

Ethnic differentials in early childhood mortality in Nepal.

This paper investigates the association of early childhood mortality (between birth and second birthday) with ethnicity in Nepal, based on data from the 1976 Nepal Fertility Survey, which was part of the World Fertility Survey. The approach is through a series of hazard models, which incorporate ethnicity, year of birth, mother's illiteracy, father's illiteracy, rural-urban residence, region, sex, maternal age, survival of previous birth, previous birth interval, and breast-feeding as covariates. Ethnic differentials in early childhood mortality are not explained by the other socioeconomic and demographic covariates, except for a modest effect of illiteracy, but the remaining covariates explain a great deal of variability in early childhood mortality itself. Analysis using an improved specification of breast-feeding as an age-varying covariate indicates, on average, that breast-feeding, relative to not breast-feeding, reduces age-specific mortality risks during the first 2 years of life by 76%, a very large effect.

PIP: Hazard models were fitted to data from the 1976 Nepal Fertility Survey (NFS) to investigate ethnic differentials in early childhood mortality between birth and 2nd birthday. The NFS included 5940 ever-married women aged 15-49. In addition to data on early childhood mortality, the survey collected considerable information on demographic and socioeconomic factors relevant to childhood mortality. Biomedical and environmental data were not collected. The births in the analysis were restricted to last and next-to last births of order 2 or higher occurring during 1962-76. A table shows the results of fitting 5 hazard models. Model I incorporates only ethnicity and year of birth as covariates. The full model II includes ethnicity and year of birth and also socioeconomic and demographic covariates. Relative to model II, model III deletes the ethnicity covariates, model IV deletes the socioeconomic covariates, and model V deletes the demographic covariates. Comparison of the 5 models indicates how the effects of ethnicity of early childhood mortality are affected by each block of covariates and by all covariates simultaneously. Model I shows that ethnicity, with only year of birth controlled, has a considerable mortality effect, as measured by relative risk. Model II, the complete model, shows that the introduction of socioeconomic and demographic covariates makes little difference in the relative risks by ethnicity. Comparison of model II with model I shows that relative risks increase only slightly for 3 ethnic groups and decrease only slightly for 4, and that levels of statistical significance of ethnicity coefficients are reduced only marginally. Thus, the socioeconomic and demographic covariates contribute little to explaining ethnic differentials in early childhood mortality. With Model III, except for the now significant effects of illiteracy of the mother and, to a lesser extent of the father, which become somewhat stronger, relative risks for the remaining covariates are changed hardly at all by this deletion. Relative risks for the various ethnic groups increase slightly with Model IV, indicating that socioeconomic factors explain some of the effects of ethnicity on early childhood mortality. In model V, deleting demographic covariates from model II, relative risks for the ethnic groups tend to decrease slightly. The average effect of breastfeeding is huge. The early childhood mortality risk for breastfed children is only 24% of that for children who are not breastfed.

Contraceptive method choice in the Philippines, 1973-83.

Contraceptive use and method mix were analysed using Philippines national survey data of 1973, 1978, and 1983. The analyses suggest that the reported decline in contraceptive prevalence between 1978 and 1983 was due to under-reporting of use in 1973 and 1983. The shifts in contraceptive method mix were also partly from under-reporting of rhythm and other methods in 1973 and 1983. Nevertheless, the determinants of method choice were similar in all three sets of data. Filipino couples were making rational choices in terms of their contraceptive goals, access, evaluation, and competence. Modifications in the directions and magnitude of the relationships in determining method choice also occurred, partly reflecting the increased use of sterilization by older, higher parity women.

The fertility of Korean minority women in China: 1950-1985.

PIP: Korean minority women in China have registered dramatic declines in fertility since 1950. The annual growth rate of the Chinese population averaged 2.1% in the 1964-82 period and that for the Han majority was 2.0%. The growth rates for all other ethnic groups were higher than the national average, except for the Korean minority (1.5%). Between the mid-1960s and mid-1970s the fertility decline was particularly marked among Yanbian Korean women; the total fertility rate dropped from 4.8 in 1965 to 1.9 in 1975. With small fluctuations, fertility has remained at or slightly below replacement level since 1975. The observation that Yanbian Koreans began and completed the fertility transition earlier than the rest of China is puzzling given the fact that government family planning programs have less stringent requirements for minorities. Since 1975, the childbearing of Korean women in China has been virtually limited to ages 20-29 years and there are relatively small differences in fertility for women with different levels of education. It is speculated that the migrant Koreans are a selected group of people who place less value on traditional norms and thus are more receptive to the small family size ideal. The fact that this minority has its own language schools and mass communication systems has meant that new norms adopted by the elite are spread to the rest of the Korean population at a pace faster than that in typical Chinese communities. When China introduced its ambitious national family planning program in the 1970s, the Koreans who had already begun the fertility transition were given the means to complete it at a very rapid pace. Overall, the fertility transition of Yanbian Korean women shows that rapid economic progress is not essential to a rapid decline in fertility if other conditions are favorable.^ieng

Infant mortality trends and differentials in Nepal.

Infant mortality trends and differentials are estimated from the 1981 Nepal Contraceptive Prevalence Survey (NCPS) and compared with similar estimates from the 1976 Nepal Fertility Survey (NFS) and the 1981 Census of Nepal. The analysis indicates that infant mortality rates derived directly from the NFS maternity histories are the most accurate. Infant mortality rates derived directly from the NCPS maternity histories are severely underestimated and yield a strongly biased trend that is the reverse of the true downward trend. Indirect estimates of infant mortality trends derived from child survivorship data do not result in a consistent pattern. Infant mortality differentials, when expressed in relative rather than absolute terms, are generally consistent with findings from earlier studies. Possible reasons for data quality differences among the three data sources are discussed.

PIP: Infant mortality trends and differentials are estimated from the 1981 Nepal Contraceptive Prevalence Survey (NCPS) and compared with similar estimates from the 1976 Nepal Fertility Survey (NFS) and the 1981 Census of Nepal. The analysis indicates that infant mortality rates derived directly from the NFS maternity histories are the most accurate. Infant mortality rates derived directly from the NCPS maternity histories are severely underestimated and yield a strongly biased trend that is the reverse of the true downward trend. Indirect estimates of infant mortality trends derived from child survivorship data do not result in a consistent pattern. Infant mortality differentials, when expressed in relative rather than absolute terms, are generally consistent with findings from ealier studies. More resources were devoted to the NFS than to the NCPS. These additional resources devoted to quality control spelled the difference between usable and largely unusable estimates of infant mortality. In Nepal, where survey-taking is such a difficult enterprise, substantial investiments in quality control in demographic surveys are needed. The analysis also indicates the importance of using as many estimation methods applied to as many data sources as possible, in order to maximize the number of comparisons and consistency checks.

Sex differentials in infant and child mortality in Korea.

PIP: This paper reports on a study of infant and child mortality in the Republic of Korea, a country known for a strong son preference, using the 1974 World Fertility Survey data. When the age-specific probabilities of dying for ages 0 to 5 are compared for male and female children, an unusual pattern of relatively high female mortality is observed. The higher female mortality is more pronounced during childhood than during infancy. Multivariate analysis of life tables, using a hazard model, shows that covariates influencing the mortality at young ages differ for male and female children and suggests that male and female children receive unequal care by their parents. The analysis also reveals different patterns of interaction between infant and child mortality and mother's fertility control behavior depending on the sex of the child.^ieng

Evaluation of contraceptive history data in the Republic of Korea.

The consistency of retrospective and current status data on contraceptive use from a series of national fertility surveys carried out during the 1970s in Korea is investigated. Aggregate consistency is examined among random samples from the same cohort or cohorts of women interviewed in each survey. The results indicate that estimates of trends in contraceptive use from a retrospective history in one survey, or from cross-sectional estimates in a series of surveys, can each yield misleading findings. Data from the 1974 Korean National Fertility Survey (KNFS) appear to be more reliable than those from other surveys, possibly because an interval-by-interval contraceptive history was used, explicit definitions of contraceptive methods were given prior to taking the contraceptive history, and the KNFS involved longer interviewer training and, perhaps, less time pressure during interviews.

PIP: This report investigates the consistency of retrospective and current status data on contraceptive use from a series of national fertility surveys carried out during the 1970s in Korea. Aggregate consistency is examined among random samples from the same cohort or cohorts of women interviewed in each survey. The results indicate that estimates of trends in contraceptive use from a retrospective history in 1 survey, or from cross-sectional estimates in a series of surveys, can each yield misleading findings. Data from the 1974 Korean Fertility Survey (KNFS) appear more reliable than those from other surveys. This is possibly the result of 1) the use of an interval by interval contraceptive history; 2) presentation of explicit definitions of contraceptive methods prior to taking the contraceptive history; or 3) the KNFS involving longer interviewer training and perhaps less time pressure during interviews. The results clearly show that there are dangers inherent in estimating trends in contraceptive use. On the other hand, evidence from the 1974 survey indicates that contraceptive use data of reasonable quality can be collected in a retrospective history, given adequate questionnaire design and interviewer training and effort. The fact that the results indicate large intersurvey differences in the completeness of reporting of use stresses the importance of evaluating contraceptives use data before they are analyzed.