MECap Knowledge Product
A Commissioner’s Guide to Probability Sampling for Surveys at USAID
By Julie Uwimana and Jennifer Kuzara
This publication was produced at the request of the United States Agency for International Development (USAID). It was prepared under the
Expanding Monitoring and Evaluation Capacities (MECap) task order, contract number AID-OAA-M-14-00014, managed by Social Solutions
International. The views expressed in this publication are solely of the authors and do not reflect the views of USAID.
Document Purpose and Audience........................... 1
Surveys at USAID ..................................................... 2
Descriptive and Inferential Statistics ...................... 3
Sampling Basics ......................................................... 5
Probabilistic vs. Non-probabilistic ...................................... 5
Probability Sampling ............................................................... 6
How Do I Choose a Sample Frame?........................ 9
Sample Frames and Sampling Bias....................................... 9
Making Generalizations .......................................................10
How Do I Review Sample-Size Calculations? ....... 10
Common Parameters ..........................................................10
Design Effect.......................................................................... 11
Effect Size ...............................................................................12
Statistical Power ...................................................................12
Confidence Level, Margin of Error, and Confidence
Interval ....................................................................................12
Significance and Alpha.......................................................... 13
Non-Response....................................................................... 13
Attrition ..................................................................................13
Other Considerations ............................................. 14
Population Estimates............................................................ 14
Proportional Allocation of the Sample ............................ 15
Sensitivity Analysis................................................................ 16
What Does Sampling Look Like in Practice? ....... 17
Why is My Input in These Decisions Important? . 19
Annex A: Sampling Design Review Checklist....... 22
Annex B: Examples of Statistical Tests for Variable
Combinations .......................................................... 24
Annex C: Sampling Approaches ............................ 27
Annex D: Common Sample Size Parameters...... 28
Annex E: Timeframe............................................... 29
Annex F: Terms....................................................... 31
Resources ................................................................. 33
Document Purpose and Audience
Sampling is the process of studying a subset of a
population for the purposes of describing or
testing questions about a whole population. At
USAID, sampling may be used to collect data for
baselines, evaluations (both impact evaluations and
performance evaluations), assessments that require
survey data, and in some cases in indicator
reporting.
This document includes answers to common
questions that the authors have been asked by
USAID staff and is based on experience, best
practice, and existing agency guidance. Some
USAID bureaus have specific guidance for
population-based surveys, like the
Feed the
Future’s Population Based-Survey Sampling Guide,
which is guidance targeted at practitioners. This
guide reviews similar content at a simplified level
to support commissioners of surveys.
The purpose of this document is to provide a
foundational understanding of probability sampling
to USAID staff to equip them as well-informed
commissioners and consumers of surveys,
evaluations, and other products (hereafter
referred to as studies) that require probability
sampling. We hope that it will serve as a resource
for commissioners to make informed decisions
about surveys and to use monitoring, evaluation,
and learning (MEL) resources effectively. The main
audience for this document includes monitoring, evaluation, and learning specialists, Contracting Officer’s
Repr
esentative (CORs), and Agreement Officer’s Representative (AORs).
This d
ocument is intended to provide a general overview of sampling and r
elated concepts of representative
survey design. It is not official guidance. Rather, it represents good practices in the survey design field. It is
not specific to any USAID initiative or requirement, nor is it exhaustive. When pursuing a representative
April 2020 | 1