• Home
• Site Map
• Foreword
  • Foreword
  • Contributors & Copyright
• Introduction
  • Introduction
  • Methods
  • Sites
    • Why Pando and East Kalimantan?
    • Municipal Governments and Decentralisation in Bolivia
    • Municipal Governments and Decentralisation in Indonesia
• Part I. Concepts
  • Wellbeing and Poverty
    • Poverty is More Than Low Income
    • Poverty is a Lack of Wellbeing
    • Poverty and Wellbeing Have Many Dimensions
    • Further Reading
  • Local Governments and Poverty
    • How can Local Governments Reduce Poverty?
    • Four Steps to Improve Local Governmental Action
    • Further Reading
  • Forests and Poverty
    • How do Forests Influence Poverty?
    • What is Poverty in Forest Communities?
    • Are Forests Safety Nets, Poverty Traps or Escape Routes?
    • Further Reading
• Part II. Tools
  • Monitoring
    • What is monitoring?
    • Types of Monitoring
    • Indicators
    • When to monitor?
    • Who conducts monitoring?
    • Linking monitoring to planning
  • Participatory Planning
    • What is participatory planning?
    • Why is participatory planning useful?
    • When and where is participatory planning done?
    • Who should participate?
  • Tool 1: Interactive Mapping
    • What is interactive mapping?
    • Interactive mapping step-by-step
    • Use interactive mapping for monitoring
  • Tool 2: Wellbeing Monitoring
    • What is a local wellbeing monitoring system?
    • Wellbeing monitoring step-by-step
    • Step 1: Define purpose, develop plan and budget
    • Step 2: Develop local poverty indicators and formulate survey questions
    • Step 3: Design the sampling strategy
    • Step 4: Train staff and perform the survey
    • Step 5: Analyse and present the data
    • Step 6: Present findings
  • Tool 3: Community Evaluation
    • Evaluating local government programmes through community focus groups
    • Steps 1-3
    • Steps 4-6
    • Steps 7-8
    • Community evaluation step-by-step
  • Tool 4: Scenario-based Planning
    • What is scenario-based planning?
    • Scenario-based planning step-by-step
    • Visioning
    • Pathways
• Final Remarks
• Acknowledgements
• Additional Resources
• PDF Version

Step 3. Design the sampling strategy

The sampling strategy depends on both the goals of the monitoring and the resources available. A detailed sampling at the household level provides the most accurate data and the most information, but requires sufficient funds, time and qualified staff. If a less detailed picture is sufficient, the local government might decide to examine poverty at the community level instead of at the household level.

A. Define the sampling unit

Decide if the sampling unit will be the household or the community. If selecting the household, perform household surveys with standardised questionnaires.

If the sampling unit is at the community level, use focus group discussions. Keep in mind that the two techniques require different skills and experience. If using focus groups, consider requesting assistance from NGOs with trained facilitators. (An example for using focus group discussions is given in Tool 3.)

B. Select villages

Try to include all communities in the poverty monitoring programme to avoid the risk of producing a skewed picture. If this is not possible, select communities that adequately represent the variety of the total population.

When selecting the communities, consider whether ethnicity, community size, accessibility, dependence on forest resources or other factors could influence wellbeing and poverty in the sample. This method, called purposive sampling, can be effective if it coincides with the objective of the survey. However, the patterns that emerge from sampling are less exact and never provide the same quality of data as a survey that includes all communities. For instance, if a health problem is detected in one community, the method would lead to the conclusion that all communities in the same group have a similar problem, which might not be the case.

If using a purposive sample, schedule discussions with the monitoring team to review the results and to judge the quality and accuracy of data.

C. Select respondents for household interviews

If there are 20 or fewer households in a community, survey all households. If there are 21–60 households in a community and if the population is homogeneous with no special patterns in distribution of households, then select 20 houses randomly (e.g. by writing all the household numbers down on small pieces of paper and blindly selecting 20). If there are more than 60 households in a community, survey one-third of the houses to have a representative sample.

There are many ways of doing random sampling. One possibility is to map and number all the houses of the community and then to pick on the first of every three households, i.e. house numbers 1, 4, 7, 10, etc. (see Box 22). If possible, seek professional advice from the government statistical service regarding the most suitable sampling strategy.

© 2007 Center for International Forestry Research (CIFOR)
Contributors and copyright information