Professor Paul Sackett

Psychologists often have the need to reduce the length of a survey study to ensure data quality, meet practical constraints, or conserve participant resources. This study explores a survey technique called planned missingness (PM) as an approach to reducing study length. The researchers conducted a Monte Carlo simulation that directly compared using a PM design to the standard practice of using short forms of measures on their ability to reproduce true population intercorrelations. They manipulated a number of data characteristics including the number of constructs, missingness level, samples size, true intercorrelations, as well as the manner in which short forms are developed, and their impact on the short form–planned missingness comparison. Results show that the two approaches perform comparably across a large number of conditions. Short forms produce slightly more accurate estimates when empirically developed short forms are readily available for use. PM produce slightly more accurate estimate when short forms are developed not entirely empirically. However, discrepancies are small in general.