Selecting a Data Collection Design for Linking in Educational Measurement: Taking Differential Motivation into Account

Abstract

In educational measurement, multiple test forms are often constructed to measure the same construct. Linking procedures can be used to disentangle differences in test form difficulty and differences in the proficiency of examinees so that scores for different test forms can be used interchangeably. Multiple data collection designs can be used for collecting data to be used for linking. Differential motivation refers to the difference in test-taking motivation that exists between high-stakes and low-stakes administration conditions. In a high-stakes administration condition, an examinee is expected to work harder and strive for maximum performance, whereas a low-stakes administration condition elicits typical, rather than maximum, performance. Differential motivation can be considered a confounding variable when choosing a data collection design. We discuss the suitability of different data collection designs and the way they are typically implemented in practice with respect to the effect of differential motivation. An example using data from the Eindtoets Basisonderwijs (End of Primary School Test) highlights the need to consider differential motivation.