Validierung eines Messinstrumentes zur Bestimmung von Informationsverhalten : die Miller Behavioral Style Scale

Abstract

This diploma thesis examines the psychometric properties ofthe Miller Behavioral Style Scale (MBSS), a psychological questionnaire.The test isdesigned to identify the tendency to seek or avoidinformation concerning a threatening and uncontrollable event.Voluminous information induces stress to some people while itreduces it for others.Therefore, it is often used in medical settings.The questionnaire consists of four stress-evoking scenarios (dentist,terrorist, dismissal and airplane), each followed by eight copingoptions of which four are measuring information seeking ("monitoring") and fourrepresenting distraction ("blunting").Three scores can be derived: a total monitoring score, a total blunting scoreand a total difference score.The coping measure is widely used.Despite the common usage of the scale, the psychometricproperties are often found deficient. The internal consistency,especially of the blunting sub-scale, isproblematic and the construct and criterium validity is seldomconfirmed.Conceptional and methodological shortcomings of the constructinformation seeking under threat can be a reason. Section2 summarizes these problems: Thus, the dimensionality of the constructis not clear. Furthermore, it is not clarified if informationseeking is situation-specific or not.In the subsequentchapter the reliability of thequestionnaire is tested in an internet setting. The findingsconfirm the poor results especially for the blunting sub-scale.An item analysis identifies problematic items supposablyresponsible for these outcomes.The predictive validity is tested by presenting one half of the participantsof Study 1 with a threatening and the other half with a non-threateningscenario. Here, they could obtain additional information about the scenario.No connection could be made between the score on the questionnaireand the amount of information obtained. Additionally, the degree of the threatdid not influence the information behaviour.Whether internet usage and education is an influential factor forinformation seeking in the questionnaire is tested in a non-internetsetting. This could not be confirmed. Moreover, not only the internal consistencyfor the blunting sub-scale was found problematic but also for the monitoring sub-scale.According to the present findings further usage of thequestionnaire is not recommended.