University of Twente Student Theses

Login

Using mixed-effects modeling to account for the acquiescence response style bias in HCI research

Schwabe, I. (2013) Using mixed-effects modeling to account for the acquiescence response style bias in HCI research.

[img] PDF
13MB
Abstract:Many Human Computer Interaction (HCI) studies use questionnaires with multiple Likert-scales as measurement tool. However, it is well known in the statistical literature that such measures are often biased by the influence of responses styles. One of these response styles is the acquiescence response style (ARS) – defined as the ‘disproportionate use of positive response options’ (Weijters, Geuen, & Schillewaert, 2010, p.1). The impact of this response style has to be taken seriously. For example, means become non-interpretable and correlations can be found that do not reflect reality. The purpose of this thesis was twofold: 1) Show that the influence of the ARS is a threat to the validity of HCI research results and 2) Show that mixed-effects modelling can be a reliable tool to account for its impact. By replicating the study by Hassenzahl and Monk (2010), it was shown that a mixed-effects model can indeed correct for the impact of the ARS. However, contrary to our expectations, the influence of the ARS was so small that it can be neglected. Furthermore, the results of an analysis on the psychometric level suggest that the scales “beauty” and “hedonic quality” are indistinguishable and therefore measure the same underlying latent variable. In the discussion section, possible explanations and directions for further research are provided.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:77 psychology
Programme:Psychology MSc (66604)
Link to this item:https://purl.utwente.nl/essays/63333
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page