Household surveys are an important source of socio-economic data. Important indicators to inform and monitor development policies are often derived from such surveys. In developing countries, they have become a dominant form of data collection, supplementing or sometimes even replacing other data collection programmes and civil registration systems.
The present publication presents the “state of the art” on several important aspects of conducting household surveys in developing and transition countries, including sample design, survey implementation, non-sampling errors, survey costs, and analysis of survey data. The main objective of this handbook is to assist national survey statisticians to design household surveys in an efficient and reliable manner, and to allow users to make greater use of survey generated data.
The publication's 25 chapters have been authored by leading experts in survey research methodology around the world. Most of them have practical experience in assisting national statistical authorities in developing and transition countries. Some of the unique features of this publication include:
- Special focus on the needs of developing and transition countries;
- Emphasis on standards and operating characteristics that can applied to different countries and different surveys;
- Coverage of survey costs, including empirical examples of budgeting for surveys, and analyses of survey costs disaggregated into detailed components;
- Extensive coverage of non-sampling errors;
- Coverage of both basic and advanced techniques of analysis of household survey data, including a detailed empirical comparison of the latest computer software packages available for the analysis of complex survey data;
- Presentation of examples of design, implementation and analysis of data from some household surveys conducted in developing and transition countries;
- Presentation of several case studies of actual large-scale surveys conducted in developing and transition countries that may be used as examples to be followed in designing similar surveys.
Brief history of the publication
The project that led to this publication went through many
interrelated phases and involved many players. The first step
was to develop a proposal that clearly stated the goals of
the publication, and provided a provisional table of contents
with proposed chapter titles and brief descriptions of the
topics to be covered, and the manner in which they were expected
to be covered. Second, authors and reviewers were recruited
from among well-known experts in the relevant fields. In the
recruitment process, due consideration was given not just
to technical expertise, but also relevant experience in developing
and transition countries.
A total of 29 chapters were contributed by various authors.
Each chapter was reviewed by two referees. The revised
chapters were then assembled to produce the first draft of
the publication, which was critically reviewed at an expert
group meeting (EGM) organized by the United Nations Statistics
Division in New York, in October 2002. At the end of the EGM,
an editorial board was established to review the whole volume
and make final recommendations about its structure and contents.
This phase of the review process led to a restructuring and
streamlining of the whole volume to make it more coherent,
more complete, and more internally consistent. A total of
25 chapters were selected for the final version of the publication.
Each of these chapters then went through a third round of
review by two referees before a final decision was taken on
whether or not to select it for inclusion in the publication.
A team of editors then did a final review of the whole volume,
ensuring that the material presented was technically sound,
internally consistent, and faithful to the primary goals of
the publication. A detailed description of the contents of
the publication is provided in the following overview.
Overview of the publication
The publication is divided into two parts. Part one consists of 21 chapters dealing with survey design, implementation, and analysis. It is divided into five sections, labelled A through E. The following is a summary of the contents of each section of part one.
Section A: Survey design and implementation. This section contains three chapters. Chapter II presents an overview of various issues pertinent to the design of household surveys in the context of developing and transition countries. Chapters III and IV, discuss issues pertaining to questionnaire design and issues pertaining to survey implementation, respectively, in developing and transition countries.
Section B: Sample design. This section contains an introductory note and three chapters dealing with the specifics of sample design. Chapter V deals with the design of master samples and master frames. The use of design effects in sample design and analysis is discussed in chapter VI and chapter VII provides an empirical analysis of design effects for surveys conducted in several developing countries.
Section C: Non-sampling errors. This section contains an introductory note and four chapters dealing with various aspects of non-sampling error measurement, evaluation, and control in developing and transition countries. Chapter VIII deals with non-observation error (non-response and non-coverage). Measurement errors are considered in chapter IX. Chapter X presents quality assurance guidelines and procedures with application to the World Health Surveys, a programme of surveys conducted in developing countries and sponsored by the World Health Organization (WHO). Chapter XI describes a case study of measurement, evaluation, and compensation for non-sampling errors of household surveys conducted in Brazil.
Section D: Survey costs. This section contains an introductory note and three chapters. Chapter XII provides a general framework for analysing survey costs in the context of surveys conducted in developing and transition countries. Using empirical data, chapter XIII describes a cost model for an income and expenditure survey conducted in a developing country. Chapter XIV discusses issues pertinent to the development of a budget for the myriad phases and functions in a household survey and includes a number of examples and case studies that are used to draw comparisons and to illustrate the important budgeting issues discussed in the chapter.
Section E: Analysis of survey data . This section contains an introductory note and seven chapters devoted to the analysis of survey data. Chapter XV provides detailed guidelines for the management of household survey data. Chapter XVI discusses basic tabular analysis of survey data, including several concrete examples. Chapter XVII discusses the use of multi-topic household surveys as a tool for poverty reduction in developing countries. Chapter XVIII discusses the use of multivariate statistical methods for the construction of indices from household survey data. Chapter XIX deals with statistical analysis of survey data, focusing on the basic techniques of model-based analysis, namely, multiple linear regression, logistic regression and multilevel methods. Chapter XX presents more advanced approaches to the analysis of survey data that take account of the effects of the complexity of the design on the analysis. Finally, chapter XXI discusses the various methods used in the estimation of sampling errors for survey data and also describes practical data analysis techniques, comparing several computer software packages used to analyse complex survey data. The strong relationship between sample design and data analysis is also emphasized. Further details on the comparison of software packages, including computer output from the various software packages, are contained in the CD-ROM that accompanies this publication.
Part two of the publication, containing four chapters preceded by an introductory note, is devoted to case studies providing concrete examples of surveys conducted in developing and transition countries. These chapters provide a detailed and systematic treatment of both user-paid surveys sponsored by international agencies and country-budgeted surveys conducted as part of the regular survey programmes of national statistical systems. The Demographic and Health Surveys (DHS) programme is described in chapter XXII; the Living Standards Measurement Study (LSMS) surveys programme is described in chapter XXIII. The discussion of both survey series includes the computation of design effects of the estimates of a number of key characteristics. Chapter XXIV discusses the design and implementation of household budget surveys, using a survey conducted in the Lao People’s Democratic Republic for illustration. Chapter XXV discusses general features of the design and implementation of surveys conducted in transition countries, and includes several cases studies.