Lot Quality Assurance Sampling in Kenya

The desire within Kenya to guide program managers and policy makers toward informed decisions was the impetus for MEASURE Evaluation’s use of the lot quality assurance sampling (LQAS) methodology there. MEASURE Evaluation first used LQAS in Kenya in 2009 when it piloted the Child Survival Indicator Survey. The pilot study was undertaken in Kenya’s Western and Nyanza provinces in collaboration with the National Coordinating Agency for Population and Development (NCAPD) and the Division of Child Health.

“NCAPD served as the implementing agency, and they did a good job,” said Janine Barden-O’Fallon, who worked on the activity for MEASURE Evaluation. “They formed a technical working group, which met periodically to design the study, design and designate indicators and design the questionnaires.”

Additionally, the technical working group participated in a training of trainers and helped train interviewers. When interviewers came back from the field, everyone discussed their experiences and lessons learned. They also went through the process of taking raw data collected in the field and coming up with indicator calculations. “Then the NCAPD ran through the data, drafted the report and disseminated it,” Barden-O’Fallon said.

In 2010, MEASURE Evaluation built off the pilot and rolled the survey out to six more provinces – Central, Coast, Eastern, Nairobi, North Eastern and Rift Valley. The Kenya team used a traditional LQAS design, with five supervision areas selected from each province and 19 randomly selected sample points from each supervision area, for a total of 95 interviews from each province.  A team of four research assistants, a Regional Population Coordinator and a driver set out to each province to conduct fieldwork.   

As its name indicates, the Child Survival Indicator Survey includes questions on a variety of child survival indicators, and it targets children under the age of five and their mothers or caregivers. “The survey is fairly comprehensive,” Barden-O’Fallon explained. “We designed it so that districts can take out or add different things. For example, if there are indicators that aren’t that important in a district, those can be taken out, or if a district wants to add an indicator, that can be done too.”

Although the Kenya teams did confront a few obstacles while conducting the fieldwork – including trying to pin down nomadic communities in the northern part of the country and wading to a Lake Victoria island community during flooding – interviewers were met with very little resistance in the communities. “We interacted quite a bit with local authorities such as elders and elected officials,” Barden-O’Fallon recalled. "The elders would often help the teams break a district up into zones for our random selection, and sometimes the elder would join the teams so they had an entrée into the communities.”

MEASURE Evaluation has been asked to roll out the survey in additional districts in Kenya, which serves as a strong testament to the methodology’s usefulness. “We will go back to the same provinces, but different districts,” Barden-O’Fallon said. “LQAS is a good tool for local decision makers and project implementers because they can immediately address issues from the findings.”