Unobtrusive Measurements

Abstract

Usually, participants know that they are measured, which implies that they can react to the measurement situation, for example, by satisficing . Unobtrusive measurements are measurements where participants cannot react to the situation. Four measurement modes are distinguished, and for each of these modes examples of unobtrusive measurement procedures are given. First, the self-report mode where participants answer questions themselves (e.g., tests). An unobtrusive self-report disguises the construct that is measured. For example, a writing test to measure aggression instead of writing skill by counting the number of aggressive words in a student’s essay. Second, the other-report mode where an other person answers questions on a participant. For example, a teacher who answers questions on a student’s aggressive behavior. Third, the somatic indicators mode where a participant’s somatic reactions are observed. For example, the observation of a participant’s blushing. Fourth, the physical traces mode where physical traces of a participant are used. For example, traffic offences to assess a participant’s attitude on traffic safety. Random errors decrease the precision of unobtrusive measurements , and systematic errors bias these measurements. The main systematic error is that the unobtrusive measure does not measure the construct that the researcher wants to measure. Unobtrusive measures that involve human judgment (e.g., ratings) are prone to systematic errors , such as, the halo-effect . Unobtrusive measures should not replace reactive measures, but they can complement these measures.