Getting Something out of Nothing: Analyzing Patterns of Null Responses to Improve Data Collection Methods in sub-Saharan Africa

Abstract

Careful development and adaptation of assessments is imperative for cultural psychological research. However, despite the best efforts, the use of assessments in new contexts can reveal atypical and/or unexpected patterns of performance. We found this to be the case in the testing of assessments to be used for a larger investigation of Specific Reading Disabilities in Zambia. In a sample of 207 children (100 female) from grades 2 to 7, we illustrated that assessment characteristics (i.e., stimulus type, answer choice, and response type) differentially impact patterns of responsiveness. The number of missing values was highest for assessments that (1) used written stimuli, (2) had an open-ended answer choice, and (3) required an action response. Age and socio-economic status explained some of the variance in responsiveness in selected, but not all assessments. Consideration of the impact of stimulus and response types when adapting assessments cross-linguistically and cross-culturally is essential.

Keywords: Zambia; assessment adaptation; assessment design; assessment translation; missing data; sub-Saharan Africa.

Figure 1

Sample items for expressive vocabulary and three subtests of the Zambian Achievement Test (ZAT).