Demographic and Health Survey (standard) - DHS III
The 1996 Nepal Family Health Survey 1996 is the fith in a series of national-level population and health surveys conducted in Nepal. It is the first nationally representative comprehensive survey conducted as part of the global Demographic and Health Survey (DHS) program.
The 1996 Nepal Family Health Survey (NFHS) is a nationally representative survey of 8,429 ever- married women age 15-49. The survey is the fifth in a series of demographic and health surveys conducted in Nepal since 1976. The main purpose of the NFHS was to provide detailed information on fertility, family planning, infant and child mortality, and matemal and child health and nutrition. In addition, the NFHS included a series of questions on knowledge of AIDS.
The primary objective of the Nepal Family Health Survey (NFHS) is to provide national level estimates of fertility and child mortality. The survey also provides information on nuptiality, contraceptive knowledge and behaviour, the potential demand for contraception, other proximate determinants of fertility, family size preferences, utilization of antenatal services, breastfeeding and food supplementation practices, child nutrition and health, immunizations, and knowledge about Acquired Immune Deficiency Syndrome (AIDS). This information will assist policy-makers, administrators and researchers to assess and evaluate population and health programmes and strategies. The NFHS is comparable to Demographic and Health Surveys (DHS) conducted in other developing countries.
Survey results indicate that fertility in Nepal has declined steadily from over 6 births per woman in the mid-1970s to 4.6 births per woman during the period of 1994-1996. Differentials in fertility by place of residence are marked, with the total fertility rate (TFR) for urban Nepal (2.9 births per woman) about two children less than for rural Nepal (4.8 births per woman). The TFR in the Mountains (5.6 births per woman) is about one child higher than the TFR in the Hills and Terai (4.5 and 4.6 births per woman, respectively). By development region, the highest TFR is observed in the Mid-western region (5.5 births per woman) and the lowest TFR in the Eastern region (4.1 births per woman).
Fertility decline in Nepal has been influenced in part by a steady increase in age at marriage over the past 25 years. The median age at first marriage has risen from 15.5 years among women age 45-49 to 17.1 years among women age 20-24. This trend towards later marriage is supported by the fact that the proportion of women married by age 15 has declined from 41 percent among women age 45-49 to 14 percent among women age 15-19. There is a strong relationship between female education and age at marriage. The median age at first marriage for women with no formal education is 16 years, compared with 19.8 years for women with some secondary education.
Despite the trend towards later age at marriage, childbearing begins early for many Nepalese women. One in four women age 15-19 is already a mother or pregnant with her first child, with teenage childbearing more common among rural women (24 percent) than urban women (20 percent). Nearly one in three adolescent women residing in the Terai has begun childbearing, compared with one in five living in the Mountains and 17 percent living in the Hills. Regionally, the highest level of adolescent childbearing is observed in the Central development region while the lowest is found in the Western region.
Short birth intervals are also common in Nepal, with one in four births occurring within 24 months of a previous birth. This is partly due to the relatively short period of insusceptibility, which averages 14 months, during which women are not exposed to the risk of pregnancy either because they are amenorrhoeic or abstaining. By 12-13 months after a birth, mothers of the majority of births (57 percent) are susceptible to the risk of pregnancy. Early childbearing and short birth intervals remain a challenge to policy-makers. NFHS data show that children born to young mothers and those born after short birth intervals suffer higher rates of morbidity and mortality.
Despite the decline in fertility, Nepalese women continue to have more children than they consider ideal. At current fertility levels, the average woman in Nepal is having almost 60 percent more births than she wants--the total wanted fertility rate is 2.9 births per woman, compared with the actual total fertility rate of 4.6 births per woman. Unplanned and unwanted births are often associated with increased mortality risks. More than half(56 percent) of all births in the five-year period before the survey had an increased risk of dying because
the mother was too young (under 18 years) or too old (more than 34 years), or the birth was of order 3 or higher, or the birth occurred within 24 months of a previous birth.
Nevertheless, the percentage of women who want to stop childbearing in Nepal has increased substantially, from 40 percent in 1981 to 52 percent in 1991 and to 59 percent in 1996. According to the NFHS, 41 percent of currently married women age 15-49 say they do not want any more children, and an additional 18 percent have been sterilized. Furthermore, 21 percent of married women want to wait at least two years for their next child and only 13 percent want to have a child soon, that is, within two years.
Knowledge of family planning is virtually universal in Nepal, with 98 percent of currently married women having heard of at least one method of family planning. This is a five-fold increase over the last two decades (1976-1996). Much of this knowledge comes from media exposure. Fifty-three percent of ever-married women had been exposed to family planning messages on the radio and/or the television and 23 percent have been exposed to messages through the print media. In addition, about one in four women has heard at least one of three specific family planning programmes on the radio.
There has been a steady increase in the level of ever use of modern contraceptive method over the past 20 years, from 4 percent of currently married women in 1976, to 27 percent in 1991 and 35 percent in 1996. Among ever-users, female sterilization and male sterilization are the most popular methods (37 percent), indicating that contraceptive methods have been used more for limiting than for spacing births.
The contraceptive prevalence rate among currently married women is 29 percent, with the majority of women using modern methods (26 percent). Again, the most widely used method is sterilization (18 percent, male and female combined), followed by injectables (5 percent). Although current use of modern contraceptive methods has risen steadily over the last two decades, the pace of change has been slowest in the most recent years (1991-1996). Current use among currently married non-pregnant women increased from 3 percent in 1976 to 15 percent in 1986 to 24 percent in 1991 and to 29 percent in 1996. While female sterilization increased by only 3 percent from 45 percent of modern methods in 1986 to 46 percent in 1996, male sterilization declined by almost 50 percent from 41 percent to 21 percent over the same period.
The level of current use is nearly twice as high in the urban areas (50 percent) as in rural areas (27 percent). Only 18 percent of currently married women residing in the Mountains are currently using contraception, compared with 30 percent and 29 percent living in the Hills and Terai regions, respectively. There is a notable difference in current contraceptive use between the Far-western region (21 percent) and all the other regions, especially the Central and Eastern regions (31 percent each). Educational differences in current use are large, with 26 percent of women with no education currently using contraception, compared with 52 percent of women who have completed their School Leaving Certificate (SLC). In general, as women's level of education rises, they are more likely to use modem spacing methods.
The public sector figures prominently as a source of modem contraceptives. Seventy-nine percent of modem method users obtained their methods from a public source, especially hospitals and district clinics (32 percent) and mobile camps (28 percent). The public sector is the predominant source of sterilizations, 1UDs, injectables, and Norplant, and both the public and private sectors are equally important sources of the pill and condoms. Nevertheless, the public sector's share of the market has fallen over the last five years from 93 percent of current users in 1991 to 79 percent in 1996.
There is considerable potential for increased family planning use in Nepal. Overall, one in three women has an unmet need for family planning--14 percent for spacing and 17 percent for limiting. The total demand for family planning, including those women who are currently using contraception, is 60 percent. Currently, the family planning needs of only one in two women is being met. While the increase in unmet need between 1991 (28 percent) and 1996 (31 percent) was small, there was a 14 percent increase in the percentage of women using any method of family planning and, over the same period, a corresponding increase of 18 percent in the demand for family planning.
MATERNAL AND CHILD HEALTH
At current mortality levels, one of every 8 children born in Nepal will die before the fifth birthday, with two of three deaths occurring during the first year of life. Nevertheless, NFHS data show that mortality levels have been declining rapidly in Nepal since the eighties. Under-five mortality in the period 0-4 years before the survey is 40 percent lower than it was 10-14 years before the survey, with child mortality declining faster (45 percent) than infant mortality (38 percent).
Mortality is consistently lower in urban than in rural areas, with children in the Mountains faring much worse than children living in the Hills and Terai. Mortality is also far worse in the Far-western and Mid-western development regions than in the other regions. Maternal education is strongly related to mortality, and children of highly educated mothers are least likely to die young. For example, infant mortality is nearly twice as high among children of mothers with no education as among children of mothers with some secondary education.
Neonatal mortality is expectedly higher among males than females. However, child mortality is 24 percent higher among females than males. First births, higher order births and births spaced less than 24 months apart also experience higher mortality.
Perinatal mortality in Nepal has declined by 17 percent over the last 15 years from 63 deaths per 1,000 stillbirths and live births in the period 10-14 years before the survey to 52 deaths in the period 0-4 years before the survey. First pregnancies, high order pregnancies, pregnancies to women residing in rural areas, in the Terai, and in the Eastern region, are all at an increased risk of loss.
One possible reason for the declining mortality is improvements in childhood vaccination coverage. The NFHS results show that about 76 percent of children age 12-23 months have been vaccinated against tuberculosis, DPT and polio. However, coverage declines after the first dose for DPT and polio, with one-third of children who start the series not completing it. Fifty-seven percent of children age 12-23 months were vaccinated against measles. Overall, two of five children had all the recommended vaccinations, and 36 percent were fully vaccinated before their first birthday.
Diarrhoeal and respiratory illnesses are common causes of child deaths in Nepal. In the two weeks before the survey, 28 percent of children suffered from diarrhoea and 34 percent were ill with acute respiratory infections (AR1). However, use of health facilities is low in Nepal: only 14 percent of children with diarrhoea and 18 percent of children with ARI were taken to a health facility. Solution prepared from oral rehydration salts (ORS) was given to 26 percent of children with diarrhoea, and 4 percent received recommended home fluids (RHF). In addition, 35 percent of children with diarrhoea were given more to drink than before the diarrhoea. Over one-third of children with diarrhoea received no treatment at all.
The care that a woman receives daring pregnancy and childbirth reduces the risk of illness and death for both mother and child. The NFHS data show that mothers received antenatal care from a doctor for only 13 percent of births in the three years preceding the survey, and from a nurse/midwife for I 1 percent of births. One of two women who received some antenatal care had fewer than 3 visits. For the majority of births (56 percent), mothers did not receive any antenatal care.
An important component of antenatal care is protection against tetanus. Two or more doses of tetanus toxoid vaccines were received by mothers for about one-third of births, while 13 percent received one dose. For well over half of births, mothers did not receive a single dose.
The majority of Nepalese children are born at home without assistance from trained medical personnel. Overall, only 8 percent of births are delivered in a health facility and 9 percent are delivered under the supervision of a doctor or nurse/midwife.
Breastfeeding is nearly universal in Nepal and the average length of time that children are breastfed is relatively long (28 months). However, a significant minority of children (40 percent) are not breastfed within one day of birth. Bottle feeding is relatively rare in Nepal and less than 3 percent of children under three years of age were fed with a bottle using a nipple. Even though exclusive breastfeeding is recommended until 4-6 months of age, one-third of children age 4-5 months receive complementary foods.
There is considerable malnutrition among children in Nepal. Forty-eight percent of children under age three are stunted, 11 percent are wasted, and 47 percent are underweight. Variation by place of residence is marked, with rural children, children living in the Mountains and in the Far-western regions of Nepal more likely to be malnourished than other children.
Maternal nutritional status was also assessed from the NFHS data. Fit~een percent of Nepalese mothers of children born during the three years before the survey were less than 145 centimetres tall, the height below which a woman is considered to be at nutritional risk. Furthermore, more than one of four women fell below the cutoff of 18.5 for body mass index, which measures thinness, indicating that the level of chronic energy deficiency in Nepal is relatively high. Maternal deaths are high relative to developed countries. According to the NFHS data, maternal deaths accounted for 27 percent of all deaths to women age 15-49, with a maternal mortality ratio of 5 deaths per 1,000 live births.
Kind of Data
Sample survey data
Unit of Analysis
- Women age 15-49
- Children under Five
The Nepal Family Health Survey 1996 covers the following topics:
- Full Pregnancy History
- HIV Knowledge–Questions assess knowledge/sources of knowledge/ways to avoid HIV
- Iodine salt test
- Maternal Mortality
Regional : 5 developmental regions : Eastem, Central, Western, Mid-western and Far-western (Due to their small size, the mountain areas of the Westem, Mid-western and Far-western regions were combined).
The population covered by the 2008 DHS is defined as the universe of all women ever-married women and men in the reproductive age groups of 15-49 and 15-59
Producers and sponsors
Ministry of Health/New ERA
Macro International inc.
United States Agency for International Development
The sample for the Nepal Family Health Survey (NFHS) was designed to provide estimates, with acceptable precision, of population and health indicators including fertility and mortality rates for the country as a whole, and for urban and rural areas separately. In addition the sample was designed to provide estimates of most key variables, with the exception of fertility and mortality estimates, for the three ecological regions (Mountains, Hills and Terai or plains), the five developmental regions (Eastern, Central, Western, Mid-western, and Far-western) and the 13 domains obtained by cross classifying the three ecological regions with the five developmental regions. Due to their small size, the mountain areas of the Westem, Mid-western and Far-western regions were combined.
An initial sample size of 7,500 completed individual interviews was chosen, taking into consideration budgetary constraints and the needs of data users. This sample size was based on the need to provide estimates of several health indicators including contraceptive rates for the 13 domains, which required a minimum target sample of 350 completed interviews with eligible women in each of the domains. This allocation by domain was considered the most efficient, based on the experience of previous DHS surveys in similar countries. The sample also had to take into consideration a 10 percent overall non-response rate. A separate estimate of mortality for the entire urban area required a minimum sample size of 1,000 completed interviews with eligible women. Furthermore, because Nepal is predominantly rural (90 percent), it was also necessary to oversample each selected urban censal enumeration area by 50 percent to obtain a minimum urban sample for calculating mortality rates. Based on these considerations, a final targeted sample size of 8,252 households was deemed to yield adequate numbers of completed interviews with eligible women. The number of households actually visited was 8,500 (because in some cases there was more than one household present in the location of a selected household at the time of interview) and interviews were successfully completed for 8,429 eligible respondents.
The 1991 Population Census served as the sampling frame for the NFHS. Administratively, Nepal is divided into 75 districts. Each district is subdivided into Village Development Committees (VDCs), and each VDC into wards. The primary sampling unit (PSU) for the NFHS is a ward or group of wards in rural areas, and subwards in the urban areas. Each rural PSU is expected to yield about 100 households, according to the 1991 Population Census. The average size of 100 households per PSU (also called a standard segment) was established as a convenient measure for a complete updating of all structures and their corresponding households. In practice, a rural PSU can have between 100 and 500 households. In the rural areas, the ward is small enough in size for a complete household listing but in urban areas the ward size is large. It was therefore necessary to subdivide each urban ward into subwards. Information for the subdivision of urban wards was obtained from the Living Standards Measurement Survey, a World Bank-funded project. In total, 253 PSUs were selected--34 in the urban areas and 219 in the rural areas.
The sample for the NFHS is a two-stage stratified sample consisting of the 253 wards (or subwards) selected. The sample was allocated to each district by urban and rural areas and the number of PSUs was calculated based on an average sample take (which is the number of ultimate sample units or households in a cluster) of 25 completed interviews per PSU. In each urban or rural area of a district, the first stage of sampling was done by selecting wards (or subwards) systematically with probability proportional to size (in terms of the number of households in each ward according to the 1991 Population Census).
Unlike most other DHS surveys, households in Nepal were selected contiguously, beginning with a randomly selected start number, from the household listing for each ward (or subward). This selection process was used to minimise the difficulty encountered in moving from one selected household to another, given the mountainous terrain in most parts of Nepal and the scattered nature of households. Even though this was not the case in the urban areas of Nepal, a similar selection process was adopted in order to maintain procedural consistency. The intraclass correlation is smaller in the rural areas than in the urban areas and selecting a compact group of households will yield the same effect as selecting them systematically. Even though intraclass correlation in the urban areas may be higher, the selection of subwards (which are smaller in area and hence tend to be more homogenous) reduces the possibility of any serious bias.
A total of 8,500 households were selected for the NFHS, of which 8,111 were located by the field teams. Of the total 8,111 households that were occupied, 8,082 were successfully interviewed, yielding a response rate of 99.6 percent. The household response rate was almost the same in urban and rural areas.
A total of 8,580 women were identified as eligible for the individual interview, indicating a ratio of 1.06 eligible women per household. Interviews were completed for 8,429 women, yielding a 98 percent overall individual response rate. The individual response rate was slightly higher in rural areas (98 percent) than in urban areas (97 percent).
Dates of Data Collection
Data Collection Mode
Data Collection Notes
Prior to the main survey, 26 household listing teams, each consisting of a lister and a mapper, were recruited and trained for about a week in August 1995 at the New ERA office in Kathmandu. Household listing began soon after and was completed by the end of September 1995. Spot checks were conducted while the listers were in the field to ensure that the work was being done correctly and completely. In some cases, listers were sent back to relist areas where households had been missed or mis-listed.
Recruitment for the main survey was also carried out at the New ERA office in Kathmandu. However, keeping in mind the local language requirements, interviewers who spoke Maithili and Bhojpuri were recruited from the Central and Eastern Terai regions in order to administer these local language questionnaires. In general interviewers were recruited for their language skills, academic qualifications and previous work experience.
Training for the main survey was conducted in Kathmandu from December 17, 1995 to January 15, 1996. A total of 81 interviewers participated in the training. Because of the large number, interviewers were split up into two groups and were trained simultaneously in two separate classrooms in the same location. The training was conducted by senior project staff of New ERA, the Ministry of Health, and two representatives from Macro International.
The four-week training course consisted of instruction in general interviewing techniques and field procedures for the survey, a detailed review of the questionnaires, practice in weighing and measuring children, mock interviews between participants in the classroom, and practice interviews in the field. In addition, five special lectures were arranged---one each on the health delivery system in Nepal, family planning, maternal health, child health and AIDS. The female trainees whose participation was satisfactory were selected as female interviewers and field editors; male trainees whose participation was satisfactory were selected as male interviewers. Based on the performance of the trainees, field supervisors were also selected.
In order to maintain uniform survey procedures, four manuals relating to different aspects of the survey were prepared. The Interviewer's Manual discussed the objectives of the NFHS, interviewing techniques, field procedures, general procedures for completing the questionnaires, and included a detailed discussion of the Household and Individual Questionnaires. The Supervisor's and Editor's Manual contained instructions on organizing and supervising fieldwork, maintaining and monitoring control sheets, and general rules for editing the completed questionnaires. The manual also contained information on height and weight charts, assignment sheets, and the interviewer's progress sheet. Trainers were given the Training Guidelines for DHS Surveys Manual, which described the administrative and logistical aspects of training and data quality checks. The Household Listing Manual described the mapping and household listing procedures used in DHS surveys.
The NFHS fieldwork was carried out by 12 teams, each comprised of three female interviewers, one male interviewer, a female field editor, and a field supervisor who was either male or female (see Appendix D for a complete list of the persons involved in the NFHS). The male interviewer administered the Household Questionnaire and the female interviewers administered the Individual Questionnaire. Unlike most DHS surveys in which the Household Questionnaire and the Individual Questionnaire are administered by the same interviewer, in the NFHS, the Household Questionnaire and the Individual Questionnaire were administered by two different interviewers. This procedure was adopted to prevent possible age shifting by interviewers (to lighten their workload) when collecting information in the Individual Questionnaire on children born since the cutoffdate, which is Baisakh 2049 in the Nepalese calendar. The fieldwork started in mid-January and ended in mid-June 1996. Assignment of sample points to the teams and various logistical decisions were made by the NFHS staffofNew ERA in Kathmandu. Each team was allowed a fixed period of time to complete fieldwork in a primary sampling unit (PSU) before moving to the next PSU. In order to maintain close supervision of all the teams during the initial two weeks of field work, all teams started their fieldwork in the Terai region (10 teams in the Eastern and Central Terai regions and two teams in the Western Terai region) and were accessible within a few hours of driving. After two weeks of fieldwork in these regions, the teams went to their assigned district. All teams began their fieldwork from the Terai region and gradually moved to the north. During the initial stage, each interviewer was instructed not to conduct more than two individual interviews of women per day. As interviewers became more familiar with the questionnaire they were conducting up to a maximum of four interviews per day. A minimum of three call-backs were made by the interviewers to ensure that eligible women identified in the sampled households were successfully interviewed.
The main duty of the field editor was to examine the completed questionnaires in the field and ensure that they were correctly filled out. An additional duty was to periodically observe ongoing interviews and verify the accuracy of the method of asking questions, recording answers, following skip instructions, and identifying eligible respondents. Throughout the survey, the senior staffofNew ERA maintained close contact with all the teams through direct communication and spot checking. Data collection work was also supervised by staffof the Family Health Division and Macro International. The objective was to provide support and advice to maintain good data quality. Data quality was also ensured by providing feedback to individual teams on the results of the field check tables. These tables were produced by computers at regular intervals from data obtained in the completed questionnaires. These results were discussed with the teams to improve their performance.
Two types of questionnaires were used in the NFHS: the Household Questionnaire and the Individual Questionnaire. The contents and design of the questionnaires were based on the DHS Model B Questionnaire, which is designed for use in countries with low contraceptive prevalence. The questionnaires were adapted to local conditions and a number of questions pertaining to on-going health and family planning programmes were added in consultation with various ministries and organizations. These questionnaires, which were developed in English, were translated into Nepali, the national language, and two other local languages, Maithali and Bhojpuri.
a) The Household Questionnaire listed all usual residents of each sampled household and all visitors who had slept in the household the night before the interview. For every person listed, some basic information such as their relationship to the head of the household, sex, age, education, and marital status was collected. The main purpose of this section of the Household Questionnaire was to identify women who were eligible for the individual interviews, that is, ever-married women age 15 to 49 years. In addition, the Household Questionnaire also obtained information on the source of water, type of toilet facilities, rooms used for sleeping, main materials of the floor, ownership of various consumer durable goods, and characteristics of household heads such as religion and ethnicity. In addition, the salt used in each household was tested for its iodine content.
b) The Individual Questionnaire was administered to all ever-married women age 15-49 who were usual residents or who were present in the household the night before the survey. It obtained information on the following topics:
- Background characteristics including age, education, religion;
- Reproductive history;
- Knowledge and use of family planning methods;
- Fertility preferences and attitudes about family planning;
- Antenatal and delivery care;
- Breastfeeding and weaning practices;
- Vaccinations and health of children;
- Husband's background characteristics and women's work;
- Knowledge of AIDS;
- Maternal mortality;
- Height and weight of children and their mothers.
Information on vaccinations and health of children, and height and weight measurements were obtained for all children born since Baisakh 2049 of the Nepalese calendar, which roughly corresponds to April 14, 1992 in the Western calendar.
All the completed questionnaires were brought to the New ERA office for data entry and processing. The data entry operation consisted of office editing, coding, data entry, and machine editing. Although all completed questionnaires were throughly edited in the field, codes for ethnicity and occupation were entered in the office. In addition, the line numbers of eligible women and the birth order of all pregnancies were rechecked. Appropriate codes for "other" responses were also assigned. One supervisor and five data entry operators were responsible for the data entry and computer editing operations. The data were entered and edited using five microcomputers and the ISSA (Integrated System for Survey Analysis) software, which was developed for DHS surveys. Data entry was also 100 percent verified in order to minimize errors. The data entry was done directly from the questionnaires and was initiated within two weeks of the first receipt of the completed questionnaires. All data entry and editing operations were completed within two weeks of completion of the fieldwork, that is, by June 1996. Computer based checks were done to rectify inconsistencies.
A preliminary report highlighting the key findings of the NFHS was released in September 1996. The purpose of this report was to disseminate the basic findings of the NFHS to policy-makers, programme planners, and administrators. The report contained 19 tables and findings on fertility, awareness and practice of family planning, fertility preferences, utilization of antenatal services, unmet need, immunizations, health of children, infant and child mortality, and knowledge about AIDS.
The NFHS followed the DHS tabulation plan in order to maintain data comparability with other countries where DHS has been implemented. Final tables were generated at Macro International.
Estimates of Sampling Error
Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the NFHS is only one of many samples that could have been selected from the same population, using the same design and expected size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability between all possible samples. Although the degree of variability is not known exactly, it can be estimated from the survey results.
A sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95 percent of all possible samples of identical size and design.
If the sample of respondents had been selected as a simple random sample, it would have been possible to use straightforward formulae for calculating sampling errors. However, the NFHS sample is the result of a two-stage stratified design and, consequently, it was necessary to use more complex formulae. The computer software used to calculate sampling errors for the NFHS is the ISSA Sampling Error Module (ISSAS). This module used the Taylor linearization method of variance estimation for survey estimates that are means or proportions. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.
The Jackknife repeated replication method derives estimates of complex rates from each of several replications of the parent sample, and calculates standard errors for these estimates using simple formulae. Each replication considers all but one cluster in the calculation of the estimates. Pseudo-independent replications are thus created. In the NFHS, there were 253 non-empty clusters. Hence, 253 replications were created.
In addition to the standard error, ISSAS computes the design effect (DEFT) for each estimate, which is defined as the ratio between the standard error using the given sample design and the standard error that would result if a simple random sample had been used. A DEFT value of 1.0 indicates that the sample design is as efficient as a simple random sample, while a value greater than 1.0 indicates the increase in the sampling error due to the use of a more complex and less statistically efficient design. ISSAS also computes the relative standard error and 95 percent confidence limits for the estimates.
Sampling errors for the NFHS are calculated for selected variables considered to be of primary interest. The results are presented in an appendix to the Final Report for the country as a whole, for urban and rural areas, for each of the five development regions: Eastern, Central, Western, Mid-western, and Far-western; and for each of three ecological regions: Mountain, Hill and Terai. For each variable, the type of statistic (mean, proportion, or rate) and the base population are given in Table B.1 of the Final report. Tables B.2 to B.12 present the value of the statistic (R), its standard error (SE), the number of unweighted (N) and weighted (WN) cases, the design effect (DEFT), the relative standard error (SE/R), and the 95 percent confidence limits (R+2SE), for each variable. The DEFT is considered undefined when the standard error considering simple random sample is zero (when the estimate is close to 0 or 1).
In general, the relative standard error for most estimates for the country as a whole is small, except for estimates of very small proportions. There are some differentials in the relative standard error for the estimates of sub-populations. For example, for the variable of Using contraception among currently married women age 15-49, the relative standard errors as a percent of the estimated mean for the whole country, for urban areas, and for rural areas are 4.1 percent, 5.5 percent, and 4.7 percent, respectively.
The confidence interval for the contraceptive prevalence rate (e.g., Using contraception among currently married women age 15-49) can be interpreted as follows: the overall national sample proportion is 0.285 and its standard error is 0.012. Therefore, to obtain the 95 percent confidence limits, one adds and subtracts twice the standard error to the sample estimate, ie. 0.2854-2x0.012. There is a high probability (95 percent) that the true value of the contraceptive prevalence rate among currently married women age 15 to 49 is between 0.261 and 0.308.
The appendix C of the Final Report presents an assessment of the quality of the data collected in the Nepal Family Health Survey (NFHS). Unlike Appendix B which discusses the effect of sampling errors on the survey results, the discussion in this appendix focuses on the magnitude ofnonsampling errors and its potential effects on interpreting the findings of the survey. Nonsampling errors can take several forms: digit preference; rounding or heaping on certain ages or dates; omission of certain events in the past; deliberate distortion of information by some interviewers who want to lighten their workload; lack of cooperation by respondents in giving information about themselves or their children; respondents not agreeing to be weighed and measured; respondents not allowing their children to be weighed or measured.
Data and Data Related Resources
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.