A Manual for Planning and Implementing the LSMS Survey

Chapter 1: Introduction

The following is from:
Grosh, M. & Muñoz, J. 1996. A Manual for Planning and Implementing the Living Standards Measurement Study Surveys. LSMS Working Paper #126, The World Bank.

A. What This Manual Covers

This manual explains the planning process, technical procedures, and standards used in Living Standard Measurement Study (LSMS) household surveys, including what these procedures entail, why they are used, and how they can be implemented. The "what" is the factual description of procedures and standards. The explanation of the "why" will help the reader to understand the importance of the different procedures. Moreover, if some aspect of the procedures is to be changed or eliminated in a particular country, knowing what they were designed to achieve may aid the survey planner in finding an alternate strategy to accomplish the same objective. The "how" comprises explicit instructions, along with examples of ways the procedures have been adapted in different countries that have implemented LSMS surveys. Although the lessons presented here are derived from LSMS surveys, many of them are applicable to surveys generally, and especially to those that are complex or especially concerned with quality control.

This manual is part of a multi-pronged effort to document, evaluate, and improve LSMS surveys. As such it is not designed to stand alone, but to fill part of the gap in available materials. The planner of any new LSMS survey will need to consult many other documents as well as the one in hand. The basics are listed in Box 1.1 and a more extensive annotated bibliography is provided in Annex II.

Topics covered in this manual include the technical aspects of questionnaire formatting and testing, ways to implement a sample design, and what fieldwork and data management procedures have been successful. Ideas about directions to pursue in analyzing the data are sketched. A brief section describes how to assess local statistical capacity. Generic work plans and budgets are presented to give ballpark estimates of how long each process will take and what must be included in a budget.

The manual does not attempt to cover institutional factors in developing the scope and design of the survey project, the content of the questionnaires, or analysis of the data. As this manual is being written, the LSMS division of the World Bank (footnote 1) is just beginning a major research effort which will culminate in a separate volume covering these themes. The final product of that effort is expected in about 1998, but draft papers should be available beginning in 1996. Moreover, this manual provides only brief summaries of some technical topics for which extensive information is already available, for example, sampling and anthropometric measurement. Suggestions for further reading are provided in Annex II.

B. Who Should Read This Manual

This manual will be useful to a broad spectrum of those who collaborate on an LSMS survey, including the staffs of the statistical agency, planning agency, or university that will design, implement and analyze the survey, and of the international agencies that finance the survey, as well as technical assistants who are not familiar with LSMS survey practice. The authors have tried to write so that persons who are not specialists can read all parts of the manual. BOX1.2

The following aids are provided to make it easier for the reader to find the parts of this document most relevant to his or her particular purposes:

• brief suggestions are provided in Box 1.2;
• key messages are summarized in bullet form at the beginning of each chapter;
• the chapter's structure and potential usefulness to different audiences is outlined at the beginning of each chapter; and
• information that should be read by all audiences is presented first in the chapter, with information of interest primarily to the different specialized members of the technical teams presented in the latter parts.

C. Some Assumptions Implicit in the Manual

Many strategic decisions must be made when designing a survey project, more than can be discussed in this manual. These issues are treated in other materials already available or that are scheduled to be made available soon. However, these choices have repercussions for the parts of planning a survey that are described in this manual, and therefore this subsection briefly mentions the issues and choices that are implicit in the rest of the document. These can be thought of as the "base case" for implementation of an LSMS survey. Packages can be tailored by adding or subtracting elements from the base case. The assumptions made here about these strategic decisions are as follows:

One Year vs Multi-year Program. This manual describes a single year of an LSMS survey. When surveys are repeated once a year or once every two years, most of the same steps are required for each round. Some of them may be accomplished more easily, with less technical assistance, and with less need for new equipment. Their content, however, remains the same so the manual is still fully applicable to multiyear projects.

How Much Data Analysis to Include in the Project. This manual focuses on the production of data, although projects often include a good deal of analysis as well. Thus the manual is a guide to what may be one component of a larger project or may be a first project to be complemented by other projects that focus on analysis of data (footnote 2).

Amount of Capacity Building. This manual again focuses on the narrowest likely definition of a project. Some training will take place in the scenario used here. It includes complete training for field staff, extensive training for the data manager, and some on-the-job training for the survey manager and field manager as they interact with the technical assistants. Projects that emphasize capacity building would arrange additional training for staff involved in questionnaire design and formatting, sampling, and data management and analysis.

Source of Financing. The term "survey project" is used throughout the manual as though a special source of funds were to be sought. This has usually been the case for LSMS surveys, although of course countries could finance them from their normal national budgets. The source of funds is largely immaterial to the information provided in this manual.

Implementing Agency. This manual assumes that the survey will be carried out by the government's central statistical office, though in some countries a university or private research firm may be used instead. In the great majority of LSMS surveys, the agency chosen has been the government statistical agency.

Permanent vs Temporary Staff. Lastly there is the division of labor how much should come from the permanent staff of the statistical agency and how much from people hired on short-term contracts. The first choice may be better for institution building. The second choice may be speedier and, depending on the wages that can be paid, may make it easier to ensure high-quality staff. This manual discusses the full staff necessary to carry out a survey without differentiating whether they come from inside or outside the statistical agency.

Revised 08/06/97