Relevance & Research Question: Online questionnaires offer a way to collect paradata, such as response time or mouse movements. As they require participants' behavior, personality is expected to have an influence on how the questions are answered. The present study investigates how impulsiveness influences response times in an online questionnaire. A hypothesis that found mixed evidence in the past (Moltó, Segarra & Avila, 1993; Malle & Neubauer, 1991). Further we were interested in the effect of response speed on the reliability of questionnaires. Montag & Reuter (2008) found no such link. We extend their study by using more precise client-side response times.

Methods & Data: A convenience sample of N=572 participants was recruited, mainly under-graduate students at the universities Bonn and Ulm. Participants studying psychology could receive course credit for participation. Participants completed an online questionnaire that included demographic data and two personality questionnaires (NEO-FFI, 60 items, and Barrett Impulsiveness Scale, 30 items, each in German translation). Personality items were presented in randomized order for each participant. This enabled us to measure reaction time for each item independently.
Bayesian regression analysis (Rouder & Morey, 2012; Morey & Rouder, 2015) was used to test the relationship between impulsiveness and overall completion time against a model including only age and education (Yan & Tourangeau, 2008). Reliability was measured in terms of internal consistency using Cronbach's Alpha (Cronbach, 1951).

Results:Bayesian regression analysis yielded substantial evidence for an effect of age and education on completion time against no effect (BF_10 = 7.11). Testing our hypothesized model including impulsiveness against the model including only age and educational level yielded some little evidence against our model (BF_10 = 0.18).No notable differences in internal consistency were observed in any of the scales for neither very fast nor very slow subjects. All alpha scores ranged between 0.66 and 0.84 showing acceptable to good reliability.

Added Value: Our results show that (a) impulsiveness has no impact on the time a subject needs to complete a questionnaire and that (b) the speed of completion does not impact the internal consistency of self-reports. Further studies should strengthen this evidence.