Demographic and Health Survey 1990-1991
Pakistan, 1990 - 1991
Get MicrodataIdentification
PAK_1990_DHS_v01_M
Demographic and Health Survey 1990-1991
| Name | Country code |
|---|---|
| Pakistan | PAK |
Demographic and Health Survey (standard) - DHS II
The Pakistan Demographic and Health Survey 1990/1991 is the first survey of its kind to be conducted in Pakistan. The PDHS was intended to serve as a source of demographic data for comparison with earlier surveys, particularly the 1975 Pakistan Fertility Survey (PFS) and the 1984-85 Pakistan Contraceptive Prevalence Survey (PCPS).
The primary objective of the Pakistan Demographic and Health Survey (PDHS) was to provide national- and provincial-level data on population and health in Pakistan. The primary emphasis was on the following topics: fertility, nuptiality, family size preferences, knowledge and use of family planning, the potential demand for contraception, the level of unwanted fertility, infant and child mortality, breastfeeding and food supplementation practices, maternal care, child nutrition and health, immunisations and child morbidity. This information is intended to assist policy makers, administrators and researchers in assessing and evaluating population and health programmes and strategies. The PDHS is further intended to serve as a source of demographic data for comparison with earlier surveys, particularly the 1975 Pakistan Fertility Survey (PFS) and the 1984-85 Pakistan Contraceptive Prevalence Survey (PCPS).
MAIN RESULTS
Until recently, fertility rates had remained high with little evidence of any sustained fertility decline. In recent years, however, fertility has begun to decline due to a rapid increase in the age at marriage and to a modest rise in the prevalence of contraceptive use. The lotal fertility rate is estimated to have fallen from a level of approximately 6.4 children in the early 1980s to 6.0 children in the mid-1980s, to 5.4 children in the late 1980s. The exact magnitude of the change is in dispute and will be the subject of further research. Important differentials of fertility include the degree ofurbanisation and the level of women's education. The total fertility rate is estimated to be nearly one child lower in major cities (4.7) than in rural areas (5.6). Women with at least some secondary schooling have a rate of 3.6, compared to a rate of 5.7 children for women with no formal education.
There is a wide disparity between women's knowledge and use of contraceptives in Pakistan. While 78 percent of currently married women report knowing at least one method of contraception, only 21 percent have ever used a method, and only 12 percent are currently doing so. Three-fourths of current users are using a modem method and one-fourth a traditional method. The two most commonly used methods are female sterilisation (4 percent) and the condom (3 percent). Despite the relatively low level of contraceptive use, the gain over time has been significant. Among married non-pregnant women, contraceptive use has almost tripled in 15 years, from 5 percent in 1975 to 14 percent in 1990-91. The contraceptive prevalence among women with secondary education is 38 percent, and among women with no schooling it is only 8 percent. Nearly one-third of women in major cities arc current users of contraception, but contraceptive use is still rare in rural areas (6 percent).
The Government of Pakistan plays a major role in providing family planning services. Eighty-five percent of sterilised women and 81 percent of IUD users obtained services from the public sector. Condoms, however, were supplied primarily through the social marketing programme.
The use of contraceptives depends on many factors, including the degree of acceptability of the concept of family planning. Among currently married women who know of a contraceptive method, 62 percent approve of family planning. There appears to be a considerable amount of consensus between husbands and wives about family planning use: one-third of female respondents reported that both they and their husbands approve of family planning, while slightly more than one-fifth said they both disapprove. The latter couples constitute a group for which family planning acceptance will require concerted motivational efforts.
The educational levels attained by Pakistani women remain low: 79 percent of women have had no formal education, 14 percent have studied at the primary or middle school level, and only 7 percent have attended at least some secondary schooling. The traditional social structure of Pakistan supports a natural fertility pattern in which the majority of women do not use any means of fertility regulation. In such populations, the proximate determinants of fertility (other than contraception) are crucial in determining fertility levels. These include age at marriage, breastfeeding, and the duration of postpartum amenorrhoea and abstinence.
The mean age at marriage has risen sharply over the past few decades, from under 17 years in the 1950s to 21.7 years in 1991. Despite this rise, marriage remains virtually universal: among women over the age of 35, only 2 percent have never married. Marriage patterns in Pakistan are characterised by an unusually high degree of consangninity. Half of all women are married to their first cousin and an additional 11 percent are married to their second cousin.
Breasffeeding is important because of the natural immune protection it provides to babies, and the protection against pregnancy it gives to mothers. Women in Pakistan breastfeed their children for an average of20months. Themeandurationofpostpartumamenorrhoeais slightly more than 9 months. After tbebirth of a child, women abstain from sexual relations for an average of 5 months. As a result, the mean duration of postpartum insusceptibility (the period immediately following a birth during which the mother is protected from the risk of pregnancy) is 11 months, and the median is 8 months. Because of differentials in the duration of breastfeeding and abstinence, the median duration of insusceptibility varies widely: from 4 months for women with at least some secondary education to 9 months for women with no schooling; and from 5 months for women residing in major cities to 9 months for women in rural areas.
In the PDHS, women were asked about their desire for additional sons and daughters. Overall, 40 percent of currently married women do not want to have any more children. This figure increases rapidly depending on the number of children a woman has: from 17 percent for women with two living children, to 52 percent for women with four children, to 71 percent for women with six children. The desire to stop childbearing varies widely across cultural groupings. For example, among women with four living children, the percentage who want no more varies from 47 percent for women with no education to 84 percent for those with at least some secondary education.
Gender preference continues to be widespread in Pakistan. Among currently married non-pregnant women who want another child, 49 percent would prefer to have a boy and only 5 percent would prefer a girl, while 46 percent say it would make no difference.
The need for family planning services, as measured in the PDHS, takes into account women's statements concerning recent and future intended childbearing and their use of contraceptives. It is estimated that 25 percent of currently married women have a need for family planning to stop childbearing and an additional 12 percent are in need of family planning for spacing children. Thus, the total need for family planning equals 37 percent, while only 12 percent of women are currently using contraception. The result is an unmet need for family planning services consisting of 25 percent of currently married women. This gap presents both an opportunity and a challenge to the Population Welfare Programme.
Nearly one-tenth of children in Pakistan die before reaching their first birthday. The infant mortality rate during the six years preceding the survey is estimaled to be 91 per thousand live births; the under-five mortality rate is 117 per thousand. The under-five mortality rates vary from 92 per thousand for major cities to 132 for rural areas; and from 50 per thousand for women with at least some secondary education to 128 for those with no education.
The level of infant mortality is influenced by biological factors such as mother's age at birth, birth order and, most importantly, the length of the preceding birth interval. Children born less than two years after their next oldest sibling are subject to an infant mortality rate of 133 per thousand, compared to 65 for those spaced two to three years apart, and 30 for those born at least four years after their older brother or sister.
One of the priorities of the Government of Pakistan is to provide medical care during pregnancy and at the time of delivery, both of which are essential for infant and child survival and safe motherhood. Looking at children born in the five years preceding the survey, antenatal care was received during pregnancy for only 30 percent of these births. In rural areas, only 17 percent of births benefited from antenatal care, compared to 71 percent in major cities. Educational differentials in antenatal care are also striking: 22 percent of births of mothers with no education received antenatal care, compared to 85 percent of births of mothers with at least some secondary education.
Tetanus, a major cause of neonatal death in Pakistan, can be prevented by immunisation of the mother during pregnancy. For 30 percent of all births in the five years prior to the survey, the mother received a tetanus toxoid vaccination. The differentials are about the same as those for antenatal care generally.
Eighty-five percent of the births occurring during the five years preceding the survey were delivered at home. Sixty-nine percent of all births were attended by traditional or trained birth attendants, while 19 percent were assisted by a doctor or nurse.
The Expanded Programme on Immunisatlon in Pakistan has met with considerable success. Among children 12 to 23 months of age, 70 percent had received a BCG vaccination, 50 percent a measles vacci- nation, and 43 percent had received all three doses of DPT and polio vaccine. Only 35 percent, however, had received all of the recommended vaccinations, while 28 percent had received none at all. Thirty-nine percent of boys were fully protected, compared to 31 percent of girls.
Sixteen percent of children under the age of five had been ill with a cough accompanied by rapid breathing during the two weeks preceding the survey. Children 6-11 months old were most prone to acute respiratory infections (23 percent). Two-thirds (66 percent) of children who were sick were taken to a health facility or provider. All but 15 percent of the sick children received some kind of treatment.
About the same proportion of children (15 percent) had suffered from diarrhoea in the two weeks preceding the survey, with the highest incidence among children under two years of age. Nearly half (48 percent) were taken to a health facility or provider. About two of five (39 percent) children with diarrhoea were treated with oral rehydration solution prepared from ORS packets. Knowledge of oral rehydration therapy is widespread: 90 percent of mothers recognise ORS packets. Nearly two-thirds (63 percent) of mothers have used ORS packets at some time, and among these, three-quarters had mixed the solution correctly the last time they prepared it.
Thirty percent of children had suffered from fever in the two weeks preceding the survey. Those most prone to illness were age 6 to 11 months. Two-thirds of children with fever were taken to a health facility or provider.
Inadequate nutrition continues to be a serious problem in Pakistan. Fifty percent of children under five years of age suffer from stunting (an indicator of chronic undemutrition), as measured by height for age. The prevalence of stunting increases with age, from 16 percent for children under 6 months to 63 percent of four-year olds. The lowest prevalence is found in Punjab (44 percent), and the highest in Balochistan (71 percent). The mother's level of education is an important factor; the prevalence of stunting varies from 18 percent for mothers with some secondary education to 56 percent for mothers with no education.
Acute undemutrition, low weight for height, is less of a problem in Pakistan than chronic undemutrition. Nine percent of children suffer from acute undemutrition (wasting). The prevalence of wasting does not vary substantially between geographic groupings. The largest differential is for mother's education: 4 percent of children of mothers with some secondary school or higher education are wasted, compared to 10 percent of children of mothers with no schooling.
A systematic subsample of households in the women's survey was selected to obtain information from the husbands of currently married women. The focus was on obtaining information about attitudes, behaviour, and the role of husbands regarding family planning. Husbands' responses concerning knowledge and use of contraception were remarkably similar to women's responses: about four-fifths knew of at least one method, two-thirds knew of a source of supply, one-fourth reported that they and their spouses had used contraception sometime in the past, and about one-seventh were current users.
Although a majority of husbands (56 percent) approve of family planning, wives are more likely to favour family planning than their husbands. Since husbands usually have a predominant role in family decision making, the family planning programme should increase efforts to educate and motivate husbands.
Sample survey data
- Household
- Women age 15-49
Scope
The Pakistan Demographic and Health Survey 1990/1991 covers the following topics:
- Breastfeeding and food supplementation practices
- Child nutrition and health
- Contraception
- Family size preferences
- Fertility
- Immunisations and child morbidity
- Infant and child mortality
- Maternal care
- Nuptiality
- Knowledge and use of family planning
Coverage
The sample design adopted for the Pakistan Demographic and Health Survey is a stratified, clustered and systematic sample of households. The universe consists of urban and rural areas of the four provinces of Pakistan as defined in the 1981 Population Census, excluding the Federally Administered Tribal Areas (FATA), military restricted areas, the districts of Kohistan, Chitral and Malakand, and protected areas of North West Frontier Province (NWFP). The population of excluded areas constitutes about 4 percent of the total population.
All ever-married women age 15-49 years who were either usual residents of the households in the sample or visitors present in the household on the night before the survey were eligible to be interviewed in the survey.
Producers and sponsors
| Name |
|---|
| National Institute of Population Studies (NIPS) |
| Name | Role |
|---|---|
| IRD/Macro International Inc. (IRD) | Technical assistance |
| Federal Bureau of Statistics, Statistics Division | Collaborator |
| Name | Role |
|---|---|
| United States Agency for International Development | Funding |
Sampling
SAMPLE DESIGN
The sample design adopted for the Pakistan Demographic and Health Survey is a stratified, clustered and systematic sample of households. The universe consists of urban and rural areas of the four provinces of Pakistan as defined in the 1981 Population Census, excluding the Federally Administered Tribal Areas (FATA), military restricted areas, the districts of Kohistan, Chitral and Malakand, and protected areas of North West Frontier Province (NWFP). The population of excluded areas constitutes about 4 percent of the total population.
For the urban sample, the sampling frame used was the master sample prepared by the Federal Bureau of Statistics. This frame was developed by dividing each city/t0wn into enumeration blocks of approximately 200-250 households with detailed and clearly recognizable boundary particulars and maps. The updating of the frame was done on the basis of the information obtained from the 1988 Census of Establishmants. For the rural sample, the sampling frame used was the village list published by the 1980 Housing Census. The primary sampling units in the urban domain were enumeration blocks; in the rural domain they were mouzas/dehs/villages.
SAMPLE SIZE AND ALLOCATION
The PDHS sample is a subsample of the Federal Bureau of Statistics master sample, which includes 7,420 primary sampling units (PSUs). Consideration in the selection of the PDHS sample was given to the population parameters and geographic levels for which estimates were required, the resources available, and the expected rate of nonresponse. A sample of 8,019 households (secondary sampling units) was selected for coverage from 408 sample areas (PSUs). The distribution of primary sampling units, secondary sampling units (SSUs), eligible women and eligible husbands and their actual coverage in the four provinces is given in Tables 2.1, 2.2 and 2.3.
STRATIFICATION PLAN
Cities having a population of 500,000 and above (Faisalabad, Gujranwala, Hyderabad, Karachi, Lahore, Multan, Peshawar, and Rawalpindi) were included to form the domain for the major cities. Quetta, which had a population of less than 500,000 but is the capital of Balochistan, was also included as a major city. For the selection of the sample, each of these cities constituted a separate stratum which was further stratified into low, middle, and high income areas, based on information collected in each enumeration block at the time the urban sampling frame was updated. For the remaining urban cities/towns, divisions of NWFP, Sindh, Punjab and Balochistan were grouped together to form a stratum. For the rural domain, each district in each province was considered a stratum, except in Balochistan where each division constituted a stratum.
A two-stage stratified sample design was adopted for the survey. The sample PSUs from each urban stratum were selected with probability proportional to the number of households. The sample PSUs from each rural stratum were selected with probability proportional to the population enumerated in the 1981 census.
Households within each sample PSU were considered secondary sampling units (SSUs). A fixed number of SSUs were selected systematically with equal probability using a random start and a sampling interval: 18 SSUs from each PSU in the urban domain in the four provinces and in the rural domain of Punjab Province and 25 SSUs from each PSU in the rural domain of the remaining three provinces of Sindh, NWFP and Baiochistan. Unlike previous surveys in Pakistan, the PDHS did not allow the substitution of households in the case of nonresponse.
From the selected sample of SSUs, a systematic subsample of one in three households was chosen for inclusion in the husbands' sample. The husbands of eligible women in these households were eligible to be interviewed, provided that they slept in the household the night before the interview.
The sample was designed to produce reliable estimates of population and health indicators separately for Karachi and for urban and rural areas of Penjab, Sindh, NWFP and Balochistan. This objective required an oversampling of all urban areas as well as the provinces of NWFP, Balochistan and Sindh.
The target was to interview 8,019 ever-married women age 15-49. The size of the target sample was based on an assumption of 1.1 eligible women per household and a nonresponse rate of 10 percent.
In general, the sample was adequate in size and sufficiently representative of the population to provide reliable estimates for the country as a whole, for urban areas, for rural areas, and for each province. However, for smaller groups, the sampling errors are generally higher.
A total of 8,019 households were selected for the women's sample. About 90 percent of the selected households were successfully contacted and interviewed. The shortfall was primarily due to dwellings that were vacant or households which were absent when they were visited by interviewers. Of the 7,404 households found to be occupied (including listed dwellings that could not be found), 97 percent were successfully interviewed. In other words, once a household was contacted, it was almost certain to complete the household interview. The highest response rate for the household interview was recorded for NWFP (99 percent); the lowest was recorded for Baiochistan (92 percent). In more than 15 percent of the cases in Balochistan, either the dwellings were vacant or the households were absent due to the temporary migration of households because of severe cold weather in that region.
In the interviewed households, 6,910 women were identified as eligible for the individual interview. Interviews were successfully completed for 96 percent of the eligible women. The difference between the number of women targeted for interviewing and actual contacts was mainly due to the fact that the actual number of eligible women per household was lower than assumed in the sample design. The principal reason for nonresponse among eligible women was the failure to find them at home, despite repeated visits to the household. The refusal rate was low (only 1.2 percent).
A sample of 1,757 husbands of eligible women was identified as being eligible for the husbands' interview. However, only 77 percent of eligible husbands could be comacted and have interviews completed. The response rate was particularly low in Sindh where almost one-thinl of eligible husbands were not at home and in major cities where one-quarter of husbands were not at home. The major reason for the high level of nonresponse among husbands was their absence from the households and the fact that male interviewers could not contact them even after several visits.
Because of the nature of the PDHS sample, a separate weighting factor was required for every PSU. The weighting procedure has two major components: the design component and the response differential component, with the design component being the major one. The weights were standardized so that the weighted number of completed cases at the national level is equal to the unweighted total. After data entry, weights were applied to the households and individuals in each PSU, to insure that the weighted sample would properly represent the actual geographic distribution of the population of Pakistan. Weights for husbands followed the same methodology as weights for women, except that the husbands' nonresponse rates were used in the calculations.
Survey instrument
Three types of questionnaires were used in the PDHS: the Household Questionnaire, the Woman's Questionnaire and the Husband's Questionnaire. The contents of the questionnaires were based on the DHS Model B Questionnaire, which is designed for use in countries with low contraceptive prevalence. Additions and modifications to the model questionnaire were made after extensive consultatious with related ministries and interested organisations and with members of the PDHS Technical Advisory Committee. The questionnaires were translated from the original English version into the national language (Urdu) and three regional languages (Panjabi, Sindhi and Pushto).
a) The Household Questionnaire listed all usual residents of a sampledhousehold, plus all visitors who slept in the household the night before the interview. Some basic information was collected on the charac- teristics of each person listed, including their age, sex, marital status, education and relationship to the head of the household. The main purpose of this section oftbe Household Questionnaire was to identify women and men who were eligible for the Women's Questiotmaire and the Husband's Questionnaire. In addition, the Household Questionnaire collected information on the household itself, such as the source of water, type oftullet facilities, materials used in the construction of the house, and ownership of various durable consumer goods.
b) The Woman's Questionnaire was used to collect information from eligible wometr--that is, all ever-marriedwomenage 15-49 who slept in the house hold the night before the household interview. Eligible women were asked questions about the following topics:
- Background characteristics
- Reproductive history
- Knowledge and use of contraception
- Pregnancy and breasffeeding
- Vaccinations and the health of children
- Marriage Family size preferences
- Husband's background
In addition, interviewing teams measured the height, weight and arm circumference of all respond- ents' children under age five. The PDHS was the first national survey that collected demographic, health and anthropometric data simultaneously. The questionnaire was designed to be completed in an average interview time of about 60 minutes. The actual mean time for the individual interview was 53 minutes. The interview time ranged from 47 minutes for women with no children born since January 1986 to 60 minutes for women who had three or more children during that period.
Interviews were also conducted with a subsample of husbands of eligible women who were married at the time of the survey.
c) The Husband's Questionnaire consists of a subset of the questions on the Woman's Questionnaire, with particular emphasis on family planning, marriage, and family size preferences.
Data collection
| Start | End |
|---|---|
| 1990-12 | 1991-05 |
| Name |
|---|
| Federal Bureau of Statistics |
| National Institute of Population Studies |
The main duty of the field editors was to examine the completed questionnaires in the field and ensure that all necessary corrections were made. An additional duty was to examine the on-going interviews and verify the accuracy of information collected on the eligibility of respondents. Throughout the survey, PDHS staff maintained close contact with all 15 teams through direct communication and spot-checking. The objective was to provide support in the field and advice to enhance data quality and the efficiency of interviewers. This objective was accomplished by communicating data problems and possible solutions to the interviewing teams, reminding interviewers about proper probing techniques, and examining the fieldwork of the supervisors. Each team supervisor was provided by FBS with the original household listing and the household sample selected by computer for each designated PSU. In case of any error in the sample information, the supervisors contacted FBS headquarters to resolve the problem.
Field Problems : In some instances, the work of certain supervisors was found to be weak: they were not moving to new PSUs as planned; they lacked coordination among team members; they did not dispatch the questionnaires from completed PSUs on time; they gave unauthorized leave to interviewers; they sent in an incomplete set of questionnaires; and at times they did not help female interviewers to locate sample households.
The selection of field teams was done at the regional level in order to insure that interviewers were accustomed to local dialects and cultural norms and were acquainted with localities in adjacent areas. The majority of field interviewers had received either a bachelor's or a master's degree.
In September-October 1990, prior to the main survey, a pretest of the questionnaires and field procedures was carried out. A two-week training session for interviewers and supervisors was conducted at Punjab University, Lahore. The training session was followed by two weeks of fieldwork. A total of 309 pretest interviews were completed in urban and rural areas of all four provinces in Pakistan (Punjab, Sindh, North West Frontier Province, and Balochistan).
Training for the main survey took place in November-December 1990. Training was held simul-taneonsly at the Regional Training institutes of the Ministry of Population Welfare in three cities--Karachi, Lahore and Peshawar. Staff members from the National Institute of Population Studies, the Federal Bureau of Statistics, the Regional Training Institutes and IRD/Macro International conducted the training sessions.
Participants in the training course included 16 statistical officers from the Federal Bureau of Statistics (FBS) and more than 80 female and male interviewers. The four-week training course consisted of instruction in general interviewing techniques and field procedures, a detailed review of the questionnaires, practice in weighing and measuring children, and practice interviews in the field. Trainees who performed satisfactorily in the training programme were selected as interviewers for the main survey. The female interviewers whose performance was rated as superior were selected as field editors.
The fieldwork for the PDHS was carried out by 15 interviewing teams. Each team consisted of one field supervisor from FBS, one field editor, three female interviewers, one male interviewer and one driver (see Appendix A for a complete list of survey staff). The fieldwork started in December 1990 and was completed by May 1991. Transportation for the field teams was provided by FBS, provincial Population Welfare Departments, and NIPS. Assignment of PSUs to the teams and various logistic decisions were made by the PDHS staff. Each team was allowed a fixed period of time to complete fieldwork in a PSU before moving to the next PSU. All the teams started their fieldwork close to or adjacent to their headquarters.
Data processing
All completed questionnaires for the PDHS were sent to the National Institute of Population Studies for data entry and processing. The data entry operation consisted of office editing, coding, data entry and machine editing. Although field editors examined the completed questionnaires in the field, these were re-edited at the PDHS headquarters by specially trained office editors. This re-examination covered: checking all skip sequences, checking circled response codes, and checking the information recorded in the filter questions. Special attention was paid to the consistency of responses to age questions and the accurate completion of the birth history. A second stage of office editing comprised the assignment of appropriate occupational codes and the addition of commonly mentioned "other" responses to the coding scheme. One supervisor and five data entry operators were responsible for the data entry and computer editing operations. The data were processed using five microcomputers and the DHS data entry and editing programmes written in ISSA (the Integrated System for Survey Analysis). The data entry started in the first week of January 1991, within one week of the receipt of the first set of completed questionnaires. The data entry was done directly from the precoded questionnaires. All data entry and editing operations were completed by July 1991. A series of computer-based checks were done to clean the data and remove inconsistencies. Age imputation was also completed at this stage. As in all DHS surveys, age variables such as current age, age at first marriage, and the ages of all living or dead children were imputed for those cases in which information was missing or incorrect entries were detected.
The PDHS followed the DHS tabulation plan, in order to maintain comparability with other countries where DHS surveys have been conducted. Some additional tables were in chided to examine special topics included on the modified PDHS questionnaire.
Data appraisal
Estimates derived from a sample survey are affected by two types of errors: nonsampling error and sampling error. Nonsampling error is the result of mistakes made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made during the implementation of the PDHS to minimize these types of errors, nonsampling errors are impossible to avoid and difficult to evaluate statistically.
Sampling errors, on the other hand, can be evaluated statistically. The sample of women selected in the PDHS is only one of many samples that could have been selected from the same population, using the same design and expected sample size. Each one would have yielded results that differ somewhat from the actual sample selected. The sampling error is a measure of the variability between all possible samples. Although it is not known exactly, it can be estimated from the survey results.
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 as measured in 95 percent of all possible samples of identical size and design will fall within a range of plus or minus two times the standard error of that statistic.
If the sample of women had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the PDHS sample design was a two-stage stratified design, and, consequently, it was necessary to use more complex formulas. The computer package CLUSTERS, developed by the International Statistical Institute for the World Fertility Survey, was used to compute the sampling errors with the proper statistical methodology.
In addition to the standard errors, CLUSTERS 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 ifa 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. CLUSTERS also computes the relative error and confidence limits for the estimates.
Sampling errors for the PDHS are calculated by group of eligible women and by group of husbands for selected variables considered to be of primary interest. The results are presented in an appendix to the Final Report for the whole country, for major cities, other urban and rural areas, for the four provinces (Punjab, Sindh, NWFP, and Balochistan), and (for women only) for three major age groups. For each variable, the type of statistic (mean or proportion) and the base population are given in Table B.1 of the Final Report. Tables B.2 through B.13 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.
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 such as geographical areas. For example, for the variable living children, the relative standard error as a percent of the estimated mean for the whole country, for major cities and for Balochistan is 1.2 percent, 2.0 percent, and 4.5 percent, respectively.
The confidence interval has the following interpretation. For the contraceptive prevalence rate (the percentage of women currently using a method), the overall average from the national sample is .118 (that is, 11.8 percent) and its standard error is .005. Therefore, to obtain the 95 percent confidence limits, one adds and subtracts twice the standard error to the sample estimate, i.e.. 118:t:.010, which means that there is a high probability (95 percent) that the true prevalence rate is between .108 and .129 (that is, 10.8-12.9 percent).
FIELD PROBLEMS
Every survey is subject to a variety of field problems, which cannot be fullyanticipated. The major problems encountered in the PDHS are highlighted below, with a discussion of their possible effects.
- Transportation: Each field team was assigned a vehicle to visit dispersed PSUs and to move quickly from one sample area to the next. Unexpected heavy rains during the months of March, April, and May brought landslides and flooding in Punjab, NWFP and Balochistan, causing substantial delays. Tube-boats were, therefore, hired in some areas. Several attempts were made to reach engulfed PSUs. At times, travel on foot for several miles was necessary to reach the designare, xl PSUs.
- Security of Teams: The law and order situation in Sindh was at its worst from January through April, Teams in Sindh were advised to take full precautions before going to any disturbed rural PSU. In addition, local security officers had to accompany interviewing teams to several PSUs in Balochistan. One PSU in Balochistan could not be reached by a Sindh team because of the insecure situation and a lack of police protection for the PDHS team.
- Funds: Funds for the fieldwork were often not released by responsible agencies in a timely fashion. These delays caused frustration for interviewers as they had to rely on borrowed money rather than their own salaries. A loan of Rs. 2 million from the Ministry of Population Welfare provided timely relief and faci- litated the full execution of the fieldwork.
- Timing: The actual fieldwork was planned to be completed before the month of Ramadan which started in March 1991. Due to the unfortunate delays caused by heavy rains, almost all teams worked throughout the month of Ramadan without any break.
- Noncooperation: In a few areas in NWFP and Sindh, where the main cash crop is poppies or where dacoits reside, almost all households were apprehensive about talking to the interviewers, especially when questions were asked about household members and the ownership of durable consumer goods. PDHS team members were sometimes mistaken for members of the narcotics board or as television license examiners.
- Cultural Norms: In several PSUs in Balochistan and NWFP, respondents willingly completed the interview but refused to allow anthropometric measurements to be taken. Most women did not want any outsider to touch their children. Moreover, mothers did not want others to know the weight and height of their children to protect them from the evil eye.
Sample Selection: The sample for the PDHS was selected at FBS headquarters. Some errors were detected in the sampling interval for households in the overall sample and in the husbands' sample. These problems caused some delays and confusion in the field. Most of these problems resulted in short-term difficulties but did not deter the overall progress of the project.
Data Access
| Name | Affiliation | URL | |
|---|---|---|---|
| MEASURE DHS | ICF International | www.measuredhs.com | archive@measuredhs.com |
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.
Contacts
| Name | URL | |
|---|---|---|
| General Inquiries | info@measuredhs.com | www.measuredhs.com |
| Data and Data Related Resources | archive@measuredhs.com | www.measuredhs.com |
Metadata production
DDI_PAK_1990_DHS_v01_M
| Name | Role |
|---|---|
| World Bank, Development Economics Data Group | Generation of DDI documentation |