A Comparative Evaluation of Different Single-Item Selection Procedures for Construct Measurement

Abstract

While the application of single items (SI) for construct measurement was generally accepted in the 1960s and 1970s, seminal works by Churchill (1979), Peter (1979), and Jacoby (1978) marked the start of a rethinking process by pointing out their deficits in terms of validity and reliability and by offering comprehensive guidelines for developing and validating multi-item (MI) measures of theoretical constructs. Afterwards, the use of MI scales has become an implicite condition for succefully publishing in academic literature. In the last decade, however, an increasing number of articles has appeared that revive the question whether theoretical constructs can be adequately captured with SI measures. Considering the practical advantages of Sis — such as brevity, ease of use, and lower survey costs — they are still an appealing alternative to long MI scales. However, if researchers — after carefully weighing up pros and cons — decide to opt for SI measurement, one fundamental question remains: what should this item be? Unfortunately, extant literature is not very informative on this issue.