Abstract

The 2003 Agriculture Sample Census was designed to meet the data needs of a wide range of users down to district level including policy makers at local, regional and national levels, rural development agencies, funding institutions, researchers, NGOs, farmer organisations, etc. As a result the dataset is both more numerous in its sample and detailed in its scope compared to previous censuses and surveys. To date this is the most detailed Agricultural Census carried out in Africa.

The census was carried out in order to: · Identify structural changes if any, in the size of farm household holdings, crop and livestock production, farm input and implement use. It also seeks to determine if there are any improvements in rural infrastructure and in the level of agriculture household living conditions; · Provide benchmark data on productivity, production and agricultural practices in relation to policies and interventions promoted by the Ministry of Agriculture and Food Security and other stake holders. · Establish baseline data for the measurement of the impact of high level objectives of the Agriculture Sector Development Programme (ASDP), National Strategy for Growth and Reduction of Poverty (NSGRP) and other rural development programs and projects. · Obtain benchmark data that will be used to address specific issues such as: food security, rural poverty, gender, agro-processing, marketing, service delivery, etc.

Geographic coverage

Tanzania Mainland and Zanzibar

Analysis unit

• Households
• Individuals

Universe

Large scale, small scale and community farms.

Kind of data

Census/enumeration data [cen]

Sampling procedure

The Mainland sample consisted of 3,221 villages. These villages were drawn from the National Master Sample (NMS) developed by the National Bureau of Statistics (NBS) to serve as a national framework for the conduct of household based surveys in the country. The National Master Sample was developed from the 2002 Population and Housing Census. The total Mainland sample was 48,315 agricultural households. In Zanzibar a total of 317 enumeration areas (EAs) were selected and 4,755 agriculture households were covered. Nationwide, all regions and districts were sampled with the exception of three urban districts (two from Mainland and one from Zanzibar).

In both Mainland and Zanzibar, a stratified two stage sample was used. The number of villages/EAs selected for the first stage was based on a probability proportional to the number of villages in each district. In the second stage, 15 households were selected from a list of farming households in each selected Village/EA, using systematic random sampling, with the village chairpersons assisting to locate the selected households.

Mode of data collection

Face-to-face [f2f]

Research instrument

The census covered agriculture in detail as well as many other aspects of rural development and was conducted using three different questionnaires: • Small scale questionnaire • Community level questionnaire • Large scale farm questionnaire

The small scale farm questionnaire was the main census instrument and it includes questions related to crop and livestock production and practices; population demographics; access to services, resources and infrastructure; and issues on poverty, gender and subsistence versus profit making production unit.

The community level questionnaire was designed to collect village level data such as access and use of common resources, community tree plantation and seasonal farm gate prices.

The large scale farm questionnaire was administered to large farms either privately or corporately managed.

Questionnaire Design The questionnaires were designed following user meetings to ensure that the questions asked were in line with users data needs. Several features were incorporated into the design of the questionnaires to increase the accuracy of the data: • Where feasible all variables were extensively coded to reduce post enumeration coding error. • The definitions for each section were printed on the opposite page so that the enumerator could easily refer to the instructions whilst interviewing the farmer. • The responses to all questions were placed in boxes printed on the questionnaire, with one box per character. This feature made it possible to use scanning and Intelligent Character Recognition (ICR) technologies for data entry. • Skip patterns were used to reduce unnecessary and incorrect coding of sections which do not apply to the respondent. • Each section was clearly numbered, which facilitated the use of skip patterns and provided a reference for data type coding for the programming of CSPro, SPSS and the dissemination applications.

Cleaning operations

Data processing consisted of the following processes: · Data entry · Data structure formatting · Batch validation · Tabulation

Data Entry Scanning and ICR data capture technology for the small holder questionnaire were used on the Mainland. This not only increased the speed of data entry, it also increased the accuracy due to the reduction of keystroke errors. Interactive validation routines were incorporated into the ICR software to track errors during the verification process. The scanning operation was so successful that it is highly recommended for adoption in future censuses/surveys. In Zanzibar all data was entered manually using CSPro.

Prior to scanning, all questionnaires underwent a manual cleaning exercise. This involved checking that the questionnaire had a full set of pages, correct identification and good handwriting. A score was given to each questionnaire based on the legibility and the completeness of enumeration. This score will be used to assess the quality of enumeration and supervision in order to select the best field staff for future censuses/surveys.

CSPro was used for data entry of all Large Scale Farm and community based questionnaires due to the relatively small number of questionnaires. It was also used to enter data from the 2,880 small holder questionnaires that were rejected by the ICR extraction application.

Data Structure Formatting A program was developed in visual basic to automatically alter the structure of the output from the scanning/extraction process in order to harmonise it with the manually entered data. The program automatically checked and changed the number of digits for each variable, the record type code, the number of questionnaires in the village, the consistency of the Village ID Code and saved the data of one village in a file named after the village code.

Batch Validation A batch validation program was developed in order to identify inconsistencies within a questionnaire. This is in addition to the interactive validation during the ICR extraction process. The procedures varied from simple range checking within each variable to the more complex checking between variables. It took six months to screen, edit and validate the data from the smallholder questionnaires. After the long process of data cleaning, tabulations were prepared based on a pre-designed tabulation plan.

Tabulations Statistical Package for Social Sciences (SPSS) was used to produce the Census tabulations and Microsoft Excel was used to organize the tables and compute additional indicators. Excel was also used to produce charts while ArcView and Freehand were used for the maps.

Analysis and Report Preparation The analysis in this report focuses on regional comparisons, time series and national production estimates. Microsoft Excel was used to produce charts; ArcView and Freehand were used for maps, whereas Microsoft Word was used to compile the report.

Data Quality A great deal of emphasis was placed on data quality throughout the whole exercise from planning, questionnaire design, training, supervision, data entry, validation and cleaning/editing. As a result of this, it is believed that the census is highly accurate and representative of what was experienced at field level during the Census year. With very few exceptions, the variables in the questionnaire are within the norms for Tanzania and they follow expected time series trends when compared to historical data. Standard Errors and Coefficients of Variation for the main variables are presented in the Technical Report (Volume I).

Sampling error estimates

The Sampling Error found on page (21) up to page (22) in the Technical Report for Agriculture Sample Census Survey 2002-2003

Abstract

The primary objective of the 2018 ZDHS was to provide up-to-date estimates of basic demographic and health indicators. Specifically, the ZDHS collected information on: - Fertility levels and preferences; contraceptive use; maternal and child health; infant, child, and neonatal mortality levels; maternal mortality; and gender, nutrition, and awareness regarding HIV/AIDS and other health issues relevant to the achievement of the Sustainable Development Goals (SDGs) - Ownership and use of mosquito nets as part of the national malaria eradication programmes - Health-related matters such as breastfeeding, maternal and childcare (antenatal, delivery, and postnatal), children’s immunisations, and childhood diseases - Anaemia prevalence among women age 15-49 and children age 6-59 months - Nutritional status of children under age 5 (via weight and height measurements) - HIV prevalence among men age 15-59 and women age 15-49 and behavioural risk factors related to HIV - Assessment of situation regarding violence against women

Geographic coverage

National coverage

Analysis unit

• Household
• Individual
• Children age 0-5
• Woman age 15-49
• Man age 15-59

Universe

The survey covered all de jure household members (usual residents), all women age 15-49, all men age 15-59, and all children age 0-5 years who are usual members of the selected households or who spent the night before the survey in the selected households.

Kind of data

Sample survey data [ssd]

Sampling procedure

The sampling frame used for the 2018 ZDHS is the Census of Population and Housing (CPH) of the Republic of Zambia, conducted in 2010 by ZamStats. Zambia is divided into 10 provinces. Each province is subdivided into districts, each district into constituencies, and each constituency into wards. In addition to these administrative units, during the 2010 CPH each ward was divided into convenient areas called census supervisory areas (CSAs), and in turn each CSA was divided into enumeration areas (EAs). An enumeration area is a geographical area assigned to an enumerator for the purpose of conducting a census count; according to the Zambian census frame, each EA consists of an average of 110 households.

The current version of the EA frame for the 2010 CPH was updated to accommodate some changes in districts and constituencies that occurred between 2010 and 2017. The list of EAs incorporates census information on households and population counts. Each EA has a cartographic map delineating its boundaries, with identification information and a measure of size, which is the number of residential households enumerated in the 2010 CPH. This list of EAs was used as the sampling frame for the 2018 ZDHS.

The 2018 ZDHS followed a stratified two-stage sample design. The first stage involved selecting sample points (clusters) consisting of EAs. EAs were selected with a probability proportional to their size within each sampling stratum. A total of 545 clusters were selected.

The second stage involved systematic sampling of households. A household listing operation was undertaken in all of the selected clusters. During the listing, an average of 133 households were found in each cluster, from which a fixed number of 25 households were selected through an equal probability systematic selection process, to obtain a total sample size of 13,625 households. Results from this sample are representative at the national, urban and rural, and provincial levels.

For further details on sample selection, see Appendix A of the final report.

Mode of data collection

Face-to-face [f2f]

Research instrument

Four questionnaires were used in the 2018 ZDHS: the Household Questionnaire, the Woman’s Questionnaire, the Man’s Questionnaire, and the Biomarker Questionnaire. The questionnaires, based on The DHS Program’s Model Questionnaires, were adapted to reflect the population and health issues relevant to Zambia. Input on questionnaire content was solicited from various stakeholders representing government ministries and agencies, nongovernmental organisations, and international cooperating partners. After all questionnaires were finalised in English, they were translated into seven local languages: Bemba, Kaonde, Lozi, Lunda, Luvale, Nyanja, and Tonga. In addition, information about the fieldworkers for the survey was collected through a self-administered Fieldworker Questionnaire.

Cleaning operations

All electronic data files were transferred via a secure internet file streaming system to the ZamStats central office in Lusaka, where they were stored on a password-protected computer. The data processing operation included secondary editing, which required resolution of computer-identified inconsistencies and coding of open-ended questions. The data were processed by two IT specialists and one secondary editor who took part in the main fieldwork training; they were supervised remotely by staff from The DHS Program. Data editing was accomplished using CSPro software. During the fieldwork, field-check tables were generated to check various data quality parameters, and specific feedback was given to the teams to improve performance. Secondary editing and data processing were initiated in July 2018 and completed in March 2019.

Response rate

Of the 13,595 households in the sample, 12,943 were occupied. Of these occupied households, 12,831 were successfully interviewed, yielding a response rate of 99%.

In the interviewed households, 14,189 women age 15-49 were identified as eligible for individual interviews; 13,683 women were interviewed, yielding a response rate of 96% (the same rate achieved in the 2013-14 survey). A total of 13,251 men were eligible for individual interviews; 12,132 of these men were interviewed, producing a response rate of 92% (a 1 percentage point increase from the previous survey).

Of the households successfully interviewed, 12,505 were interviewed in 2018 and 326 in 2019. As the large majority of households were interviewed in 2018 and the year for reference indicators is 2018.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results 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 2018 Zambia Demographic and Health Survey (ZDHS) to minimise this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2018 ZDHS 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 among all possible samples. Although the degree of variability 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 will fall within a range of plus or minus two times the standard error of that statistic in 95% 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 formulas for calculating sampling errors. However, the 2018 ZDHS sample is the result of a multi-stage stratified design, and, consequently, it was necessary to use more complex formulas. Sampling errors are computed in SAS, using programs developed by ICF. These programs use the Taylor linearisation method to estimate variances for survey estimates that are means, proportions, or ratios. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.

Note: A more detailed description of estimates of sampling errors are presented in APPENDIX B of the survey report.

Data appraisal

Data Quality Tables - Household age distribution - Age distribution of eligible and interviewed women - Age distribution of eligible and interviewed men - Completeness of reporting - Births by calendar years - Reporting of age at death in days - Reporting of age at death in months - Completeness of information on siblings - Sibship size and sex ratio of siblings - Height and weight data completeness and quality for children - Number of enumeration areas completed by month, according to province, Zambia DHS 2018

Note: Data quality tables are presented in APPENDIX C of the report.

Abstract

The primary objective of the 2018 NDHS is to provide up-to-date estimates of basic demographic and health indicators. Specifically, the NDHS collected information on fertility, awareness and use of family planning methods, breastfeeding practices, nutritional status of women and children, maternal and child health, adult and childhood mortality, women’s empowerment, domestic violence, female genital cutting, prevalence of malaria, awareness and behaviour regarding HIV/AIDS and other sexually transmitted infections (STIs), disability, and other health-related issues such as smoking.

The information collected through the 2018 NDHS is intended to assist policymakers and programme managers in evaluating and designing programmes and strategies for improving the health of the country’s population. The 2018 NDHS also provides indicators relevant to the Sustainable Development Goals (SDGs) for Nigeria.

Geographic coverage

National coverage

Analysis unit

• Household
• Individual
• Children age 0-5
• Woman age 15-49
• Man age 15-49

Universe

The survey covered all de jure household members (usual residents), all women aged 15-49 years resident in the household, and all children aged 0-5 years resident in the household.

Kind of data

Sample survey data [ssd]

Sampling procedure

The sampling frame used for the 2018 NDHS is the Population and Housing Census of the Federal Republic of Nigeria (NPHC), which was conducted in 2006 by the National Population Commission. Administratively, Nigeria is divided into states. Each state is subdivided into local government areas (LGAs), and each LGA is divided into wards. In addition to these administrative units, during the 2006 NPHC each locality was subdivided into convenient areas called census enumeration areas (EAs). The primary sampling unit (PSU), referred to as a cluster for the 2018 NDHS, is defined on the basis of EAs from the 2006 EA census frame. Although the 2006 NPHC did not provide the number of households and population for each EA, population estimates were published for 774 LGAs. A combination of information from cartographic material demarcating each EA and the LGA population estimates from the census was used to identify the list of EAs, estimate the number of households, and distinguish EAs as urban or rural for the survey sample frame. Before sample selection, all localities were classified separately into urban and rural areas based on predetermined minimum sizes of urban areas (cut-off points); consistent with the official definition in 2017, any locality with more than a minimum population size of 20,000 was classified as urban.

The sample for the 2018 NDHS was a stratified sample selected in two stages. Stratification was achieved by separating each of the 36 states and the Federal Capital Territory into urban and rural areas. In total, 74 sampling strata were identified. Samples were selected independently in every stratum via a two-stage selection. Implicit stratifications were achieved at each of the lower administrative levels by sorting the sampling frame before sample selection according to administrative order and by using a probability proportional to size selection during the first sampling stage.

For further details on sample selection, see Appendix A of the final report.

Mode of data collection

Computer Assisted Personal Interview [capi]

Research instrument

Four questionnaires were used for the 2018 NDHS: the Household Questionnaire, the Woman’s Questionnaire, the Man’s Questionnaire, and the Biomarker Questionnaire. The questionnaires, based on The DHS Program’s standard Demographic and Health Survey (DHS-7) questionnaires, were adapted to reflect the population and health issues relevant to Nigeria. Comments were solicited from various stakeholders representing government ministries and agencies, nongovernmental organisations, and international donors. In addition, information about the fieldworkers for the survey was collected through a self-administered Fieldworker Questionnaire.

Cleaning operations

The processing of the 2018 NDHS data began almost immediately after the fieldwork started. As data collection was completed in each cluster, all electronic data files were transferred via the IFSS to the NPC central office in Abuja. These data files were registered and checked for inconsistencies, incompleteness, and outliers. The field teams were alerted to any inconsistencies and errors. Secondary editing, carried out in the central office, involved resolving inconsistencies and coding the open-ended questions. The NPC data processor coordinated the exercise at the central office. The biomarker paper questionnaires were compared with electronic data files to check for any inconsistencies in data entry. Data entry and editing were carried out using the CSPro software package. The concurrent processing of the data offered a distinct advantage because it maximised the likelihood of the data being error-free and accurate. Timely generation of field check tables allowed for effective monitoring. The secondary editing of the data was completed in the second week of April 2019.

Response rate

A total of 41,668 households were selected for the sample, of which 40,666 were occupied. Of the occupied households, 40,427 were successfully interviewed, yielding a response rate of 99%. In the households interviewed, 42,121 women age 15-49 were identified for individual interviews; interviews were completed with 41,821 women, yielding a response rate of 99%. In the subsample of households selected for the male survey, 13,422 men age 15-59 were identified and 13,311 were successfully interviewed, yielding a response rate of 99%.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results 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 2018 Nigeria Demographic and Health Survey (NDHS) to minimise this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2018 NDHS 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 among all possible samples. Although the degree of variability 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 will fall within a range of plus or minus two times the standard error of that statistic in 95% 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 formulas for calculating sampling errors. However, the 2018 NDHS sample is the result of a multistage stratified design, and, consequently, it was necessary to use more complex formulas. Sampling errors are computed in SAS, using programs developed by ICF. These programs use the Taylor linearisation method to estimate variances for survey estimates that are means, proportions, or ratios. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.

Note: A more detailed description of estimates of sampling errors are presented in APPENDIX B of the survey report.

Data appraisal

Data Quality Tables - Household age distribution - Age distribution of eligible and interviewed women - Age distribution of eligible and interviewed men - Completeness of reporting - Births by calendar years - Reporting of age at death in days - Reporting of age at death in months - Standardisation exercise results from anthropometry training - Height and weight data completeness and quality for children - Height measurements from random subsample of measured children - Sibship size and sex ratio of siblings - Pregnancy-related mortality trends - Data collection period - Malaria prevalence according to rapid diagnostic test (RDT)

Note: See detailed data quality tables in APPENDIX C of the report.

Abstract

The Pakistan Demographic and Health Survey PDHS 2017-18 was the fourth of its kind in Pakistan, following the 1990-91, 2006-07, and 2012-13 PDHS surveys.

The primary objective of the 2017-18 PDHS is to provide up-to-date estimates of basic demographic and health indicators. The PDHS provides a comprehensive overview of population, maternal, and child health issues in Pakistan. Specifically, the 2017-18 PDHS collected information on:

• Key demographic indicators, particularly fertility and under-5 mortality rates, at the national level, for urban and rural areas, and within the country’s eight regions
• Direct and indirect factors that determine levels and trends of fertility and child mortality
• Contraceptive knowledge and practice
• Maternal health and care including antenatal, perinatal, and postnatal care
• Child feeding practices, including breastfeeding, and anthropometric measures to assess the nutritional status of children under age 5 and women age 15-49
• Key aspects of family health, including vaccination coverage and prevalence of diseases among infants and children under age 5
• Knowledge and attitudes of women and men about sexually transmitted infections (STIs), including HIV/AIDS, and potential exposure to risk
• Women's empowerment and its relationship to reproductive health and family planning
• Disability level
• Extent of gender-based violence
• Migration patterns

The information collected through the 2017-18 PDHS is intended to assist policymakers and program managers at the federal and provincial government levels, in the private sector, and at international organisations in evaluating and designing programs and strategies for improving the health of the country’s population. The data also provides information on indicators relevant to the Sustainable Development Goals.

Geographic coverage

National coverage

Analysis unit

• Household
• Individual
• Children age 0-5
• Woman age 15-49
• Man age 15-49

Universe

The survey covered all de jure household members (usual residents), children age 0-5 years, women age 15-49 years and men age 15-49 years resident in the household.

Kind of data

Sample survey data [ssd]

Sampling procedure

The sampling frame used for the 2017-18 PDHS is a complete list of enumeration blocks (EBs) created for the Pakistan Population and Housing Census 2017, which was conducted from March to May 2017. The Pakistan Bureau of Statistics (PBS) supported the sample design of the survey and worked in close coordination with NIPS. The 2017-18 PDHS represents the population of Pakistan including Azad Jammu and Kashmir (AJK) and the former Federally Administrated Tribal Areas (FATA), which were not included in the 2012-13 PDHS. The results of the 2017-18 PDHS are representative at the national level and for the urban and rural areas separately. The survey estimates are also representative for the four provinces of Punjab, Sindh, Khyber Pakhtunkhwa, and Balochistan; for two regions including AJK and Gilgit Baltistan (GB); for Islamabad Capital Territory (ICT); and for FATA. In total, there are 13 secondlevel survey domains.

The 2017-18 PDHS followed a stratified two-stage sample design. The stratification was achieved by separating each of the eight regions into urban and rural areas. In total, 16 sampling strata were created. Samples were selected independently in every stratum through a two-stage selection process. Implicit stratification and proportional allocation were achieved at each of the lower administrative levels by sorting the sampling frame within each sampling stratum before sample selection, according to administrative units at different levels, and by using a probability-proportional-to-size selection at the first stage of sampling.

The first stage involved selecting sample points (clusters) consisting of EBs. EBs were drawn with a probability proportional to their size, which is the number of households residing in the EB at the time of the census. A total of 580 clusters were selected.

The second stage involved systematic sampling of households. A household listing operation was undertaken in all of the selected clusters, and a fixed number of 28 households per cluster was selected with an equal probability systematic selection process, for a total sample size of approximately 16,240 households. The household selection was carried out centrally at the NIPS data processing office. The survey teams only interviewed the pre-selected households. To prevent bias, no replacements and no changes to the pre-selected households were allowed at the implementing stages.

For further details on sample design, see Appendix A of the final report.

Mode of data collection

Face-to-face [f2f]

Research instrument

Six questionnaires were used in the 2017-18 PDHS: Household Questionnaire, Woman’s Questionnaire, Man’s Questionnaire, Biomarker Questionnaire, Fieldworker Questionnaire, and the Community Questionnaire. The first five questionnaires, based on The DHS Program’s standard Demographic and Health Survey (DHS-7) questionnaires, were adapted to reflect the population and health issues relevant to Pakistan. The Community Questionnaire was based on the instrument used in the previous rounds of the Pakistan DHS. Comments were solicited from various stakeholders representing government ministries and agencies, nongovernmental organisations, and international donors. The survey protocol was reviewed and approved by the National Bioethics Committee, Pakistan Health Research Council, and ICF Institutional Review Board. After the questionnaires were finalised in English, they were translated into Urdu and Sindhi. The 2017-18 PDHS used paper-based questionnaires for data collection, while computerassisted field editing (CAFE) was used to edit the questionnaires in the field.

Cleaning operations

The processing of the 2017-18 PDHS data began simultaneously with the fieldwork. As soon as data collection was completed in each cluster, all electronic data files were transferred via IFSS to the NIPS central office in Islamabad. These data files were registered and checked for inconsistencies, incompleteness, and outliers. The field teams were alerted to any inconsistencies and errors. Secondary editing was carried out in the central office, which involved resolving inconsistencies and coding the openended questions. The NIPS data processing manager coordinated the exercise at the central office. The PDHS core team members assisted with the secondary editing. Data entry and editing were carried out using the CSPro software package. The concurrent processing of the data offered a distinct advantage as it maximised the likelihood of the data being error-free and accurate. The secondary editing of the data was completed in the first week of May 2018. The final cleaning of the data set was carried out by The DHS Program data processing specialist and completed on 25 May 2018.

Response rate

A total of 15,671 households were selected for the survey, of which 15,051 were occupied. The response rates are presented separately for Pakistan, Azad Jammu and Kashmir, and Gilgit Baltistan. Of the 12,338 occupied households in Pakistan, 11,869 households were successfully interviewed, yielding a response rate of 96%. Similarly, the household response rates were 98% in Azad Jammu and Kashmir and 99% in Gilgit Baltistan.

In the interviewed households, 94% of ever-married women age 15-49 in Pakistan, 97% in Azad Jammu and Kashmir, and 94% in Gilgit Baltistan were interviewed. In the subsample of households selected for the male survey, 87% of ever-married men age 15-49 in Pakistan, 94% in Azad Jammu and Kashmir, and 84% in Gilgit Baltistan were successfully interviewed.

Overall, the response rates were lower in urban than in rural areas. The difference is slightly less pronounced for Azad Jammu and Kashmir and Gilgit Baltistan. The response rates for men are lower than those for women, as men are often away from their households for work.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results 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 2017-18 Pakistan Demographic and Health Survey (2017-18 PDHS) to minimise this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2017-18 PDHS 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 among all possible samples. Although the degree of variability 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