MEASURE Evaluation - Assessment of Hand Washing Behavior and Diarrheal Prevalence
Assessment of Hand Washing Behavior and Diarrheal Prevalence
Abstract
Name: Assessment of Hand Washing Behavior and Diarrheal Prevalence
Purpose: The Assessment of Handwashing Behavior and Diarrheal Prevalence tool is a
market research household study developed for the purposes of evaluating handwashing
attitudes and behavior in Central America. It provides quantitative information for
designing a communication strategy to promote handwashing among children and
mothers.
Type of Design: Cross-sectional household study of households with children under ten
years of age.
Language: Spanish
Sample Size: In Central America, a sample of 1,500 respondents was taken. A 1,000
respondent sample allows for the analysis of results within a 95% confidence interval
(p=q=50%), accepting a maximum sample error of +/- 3.1% while the 1,500 respondent
sample allows a maximum error of +/- 2.5 % .
Where Used: Guatemala, Costa Rica, Honduras and El Salvador
Name: Assessment of Handwashing Behavior and Diarrheal Prevalence
Origin: Camille Saadé
Source: BASICS
Basic Description: The tool is a quantitative household study to assess the handwashing behavior
of children below ten years of age and mothers of children under ten. In addition, it measures the
incidence of diarrhea in children below age five. It has been utilized in a pre-test/post-test design
on a nationally representative sample of households with children under five years of age. The
main objective of the research is to understand the motivation of mothers and children to practice
correct handwashing and to assess the obstacles (i.e., personal or environmental, such as access
to clean water) to proper handwashing. Proper handwashing includes duration of washing,
washing during critical moments of contamination and technique.
Country Applications: Guatemala, Costa Rica, Honduras and El Salvador
Technical Scope: Child survival social marketing
Purpose: The Assessment of Handwashing Behavior and Diarrheal Prevalence tool is a market
research household study developed for the purposes of evaluating handwashing attitudes and
behavior in Central America. It provides quantitative information for designing a
communication strategy to promote handwashing among children and mothers.
Method: Quantitative
Type of Design: Cross-sectional household study of households with children under ten years of
age.
Frequency of administration: Before and after intervention; particular to each country application
Key Users of Information
Donors, MOH, project staff, pharmaceutical companies or other private sector institutions and
advertising agencies.
Objectives and Scope of the Tool
Define the target consumer profile for a "handwashing with soap" campaign
Assess the impact of handwashing campaign on the attitudes and behavior of the target
consumer group
Key Monitoring Needs and Evaluation Questions Tool Seeks to Address
What motivates mothers to practice correct handwashing?
How do children under ten learn about handwashing? What do they practice?
Is the incidence of diarrhea in children under five associated with the practice of
handwashing?
What are the sources of clean water in the intervention countries?
What are mothers' attitudes toward use of soap?
Observation of mothers and children under ten washing hands
Key Indicators
Timing, technique and occasion of handwashing by mothers of children under five years
Timing, technique and occasion of handwashing by children under ten years
Source of clean and/or potable water in households
Incidence of diarrhea among children under five in previous two weeks
Knowledge of handwashing among mothers and children under ten
Research Design
Standard Protocol:
A stratified random sample is taken from among the two lower socioeconomic income levels.
The universe of the sample is all households in urban and rural areas throughout the country that
correspond to the lower two socioeconomic levels (D and E) with children less than ten years of
age. In Central America, a sample of 1,500 respondents was taken. A 1,000 respondent sample
allows for the analysis of results within a 95% confidence interval (p=q=50%), accepting a
maximum sample error of +/- 3.1% while a 1,500 respondent sample allows a maximum error of
+/- 2.5 % . Both sample sizes will be adequate to represent rural and urban areas of Central
American countries but should be recalculated to fit the setting. A smaller sample size would not
be as effective, and could be applied to only rural/urban areas separately. The sample is selected
using a stratified random process, using area (urban vs. rural) and region as stratifying variables
to ensure that it is representative.
Lessons from experience:
Market research firms, such as the one that implemented this study in Central America,
may require additional training in public health and epidemiological data analysis. Many
market research agencies do not rigorously analyze data for relationships and causality
between variables. Careful attention should be paid to the design of the study and the
selection of the sample to ensure consistency and adequate rigor. Many firms are
accustomed to promoting efficiency in sampling (i.e., selecting a convenience sample
and/or taking shortcuts in database structuring and analysis).
In order to ensure data quality, there must be one technical coordinator who can oversee
the work, from design to data analysis.
Training and Implementation
Standard protocol: Classroom training of interviewers is followed by field-based training and pretesting of
questionnaires. In Central America, the questionnaire was pretested three times to ensure pan-regional standardization. The interview is conducted with the principal woman of the household.
One supervisor is assigned to four to five interviewers. The supervisor validates at least 35% of
the interviews, using random selection. In addition, data is validated through direct observation
as well as post-interview visits with interviewed households.
Lessons from experience:
Training should be supported closely by an external technical coordinator to ensure
standardization and quality.
Implementation should also be closely monitored to ensure proper selection of
respondents and avoid "shortcuts" in data collection.
Analysis
Standard protocol: Questionnaires are edited, entered and validated into databases. The data is analyzed by
sociodemographic variables and by technique and reported frequency of handwashing.
Lessons from experience:
In the case of Central America, data was re-cleaned and re-analyzed in Washington. This
process can be avoided if data collection and entry is carefully monitored by an external
party to ensure standardization of coding and avoid reporting bias.
Research objectives and hypotheses should be clearly defined and reviewed with the
implementing agency prior to data collection, perhaps including development of dummy
tables, to avoid missed opportunities for analysis.
Reporting
Standard protocol: The narrative describes design, sampling and implementation. Results of key indicators are
displayed in tabular form where appropriate. Data is presented in graphical form
Dissemination of Results
Results should be distributed to all stakeholders, including donors and project managers.
Contact Person
Camille Saade
BASICS Project
1600 Wilson Blvd.
Arlington, VA 22209
703-312-6800
