Better low-tech than sorry:
How technophile questionnaires may affect psychological representativeness

Frederik Funke & Ulf-Dietrich Reips

Poster presented at the 14th annual General Online Research Tagung (GOR) conference of the German Society for Online Research (D.G.O.F.)
March 5-7, 2012 in Mannheim (Germany)

Relevance & Research Question: Optimizing online questionnaires for certain browsers not only raises the risk of loosing respondents but also may bias the sample composition. In their GOR paper on the low-tech principle, Buchanan and Reips (2001) found that more educated users were more likely to turn JavaScript off. Furthermore, respondents using Mac OS scored higher on the personality trait Openness to Experience than users of Windows OS. This study focuses on the question if these differences still hold, ten years later and with a sample of experienced Internet users.
Methods & Data: The questionnaire was a Big Five personality inventory. Following the low-tech principle, participation was possible with any Web browser.
Results: Overall, 2.6% of all participants (N = 358) had JavaScript disabled. Male respondents were more likely to have JavaScript disabled than female respondents, chi^2(1, N = 344) = 11.64, p = .001, odds ratio = 15.6. Mac users scored higher on Openness (e.g., "I enjoy hearing new ideas") than Windows users, F(1, 294) = 12.14, p = .001, eta^2 = .040. Furthermore, respondents using Macs scored lower on Agreeableness (e.g., "I am interested in other people") than users of Windows, F(1, 294) = 9,02, p = .003, eta^2 = .030. Within respondents running Windows Extraversion (e.g., "I am the life of the party") was slightly higher for users of the Internet Explorer in comparison to users of Firefox, F(1, 277) = 4.13, p = .043, eta^2 = .015. Finally, there was a tendency that respondents with JavaScript deactivated scored lower on Openness than respondents with this technology activated, F(1, 312) = 3.05, p = .082, eta^2 = .010.
Added Value: The present study confirms and extends the results from Buchanan and Reips (2001). A questionnaire exclusively optimized for certain OSs or browsers can seriously bias the psychological and demographical sample composition. In the present study restricting participation to respondents with JavaScript enabled would have reduced the number of male participants. Overall, it is recommended either to refrain from using complex technologies or to implement alternative low-tech versions of questionnaires as fallback.