Habitual trust entails the motivation to affiliate with others (e.g. Evans & Revelle, 2008; Slepian et al., 2012) and the cognitive tendency to think in congruent terms (Kleiman et al., 2015; Mayo, 2015; Mayo et al., 2014; Posten & Mussweiler, 2013). Both a motivation to affiliate (Steinmetz & Posten, 2017) and a congruent thinking style should foster answering affirmatively to questions irrespective of the question content. Across six studies, we demonstrate that, indeed, people who are habitually trusting answer more affirmatively. Using four different measures of habitual trust, six studies consistently demonstrated that people who are habitually trusting show an acquiescent response bias. Habitual trust predicted people's acquiescent responding to questions addressing various aspects of life (Study 1), response scales as used in classic psychological questionnaires (Studies 2a and 2b) and when evaluating others’ work determining their monetary outcomes (Study 3). Over and above response scales, individual levels of habitual trust also predicted the choice of default options (Study 4) and agreement in a forced-choice setting, in which participants could only agree or disagree (Study 5). The relation of habitual trust and acquiescence held across responses to ingroup or outgroup members (Study 5). Notably, the relation even showed in contexts that should be unaffected by social desirability concerns (Study 2b).
12.1 Alternative explanations and limitations
Our approach of using a variety of acquiescence measures might have limitations. Specifically, the correlations of the habitual trust and acquiescence measures differ between our Studies 1–2b that use more traditional survey questions to assess acquiescence (rs between .248 and .314) and our Studies 3–5 that use more behavioural manifestations of acquiescence (rs between .127 and .170). Thus, the relation between habitual trust and acquiescence might be more pronounced (and thus potentially more problematic for researchers) who use survey batteries, whereas more behavioural measures might be less affected. The smaller correlations between habitual trust and such behavioural measures might stem from the fact that participants might have been aware of the consequences of their responses for others (Study 3) and for the self (Study 4). This awareness might have increased the motivation to respond accurately, which might reduce but not fully eliminate acquiescence. Such awareness of consequences might be reduced when people respond to survey questions. Future research could test whether instructions that highlight the importance of surveys for public policy or for research would reduce the relation between habitual trust and acquiescence.
One might speculate whether the relation between habitual trust and acquiescence is driven by an underlying relation between habitual trust and the content of the questions that we used to assess acquiescence. Indeed, there might be some ‘true’ relation between habitual trust and the content of some questions that is not attributable to acquiescence, for example, on the interdependence subscale of the acquiescence measure in Studies 2a and 2b (Zeffane, 2017). Whereas there might be a ‘true’ relation between habitual trust and the content of some items in acquiescence measures, such a relation cannot explain our findings as a whole. For one, we measured acquiescence in a variety of ways across our studies. Specifically, in Studies 1–2b, we measured acquiescence as the responses to traditional survey questions (i.e. the WVS and the Singelis scale). It is unlikely that habitual trust relates to all the various concepts tested in the WVS or relates to interdependence and independence at the same time. For another, we used more behavioural measures of acquiescence in several studies that had no apparent overlap with habitual trust. For example, in Study 4, participants indicated their binary preference to stay with a default ad, without measuring any trust-related personality or values questions. In Study 5, we measured acquiescence as a binary choice between judging factual statements as true versus false. The measures in Studies 4 and 5 capture not a particular psychological content but a simple choice or judgement.
An alternative explanation for the observed correlation between habitual trust and acquiescence could be that this correlation is driven by trait agreeableness (Costa & McCrae, 2008). However, whereas agreeableness might well be related to acquiescence (Hibbing et al., 2019), the trust sub-facet of agreeableness could be one of the drivers of such a relation. With the exception of compliance, the other subscales of agreeableness (i.e. altruism, straightforwardness, modesty and tender-mindedness) seem less likely to correlate with the measures of acquiescence that we used. For example, someone high in modesty might show less acquiescence on the scale in Studies 2a and 2b that measures positive traits and behaviour, for fear of being immodest. Compliance has indeed been shown to correlate with acquiescence (measured as deference; Schuman & Presser, 1996). Because of this finding and our findings that habitual trust correlates with acquiescence, we would expect these two sub-facets of agreeableness to be the main drivers of an observed correlation between the entire agreeableness scale and an acquiescence measure. Thus, we focus on the sub-facet habitual trust instead of the entire agreeableness scale.
We base our hypothesis on motivational and cognitive mechanisms accompanying habitual trust that foster acquiescence. However, there might be more parsimonious explanations for our findings. First, perceiving a person as trustworthy often means also to perceive the person as similar to oneself (Farmer et al., 2014; Posten & Mussweiler, 2019). Similarity in turn leads to assimilation (Mussweiler, 2003), and thus, potentially, to affirmative answering. Second, trusting a person could induce halo effects (Nisbett & Wilson, 1977) by attributing generally positive characteristics to the counterpart, for instance being knowledgeable in a quiz (such as in Study 5). However, these alternative accounts would not predict a relation between habitual trust and acquiescence in low-trust contexts (e.g. outgroup interactions in Study 5), in which the question asker is neither similar to the self nor (presumably) perceived particularly positively. As we find that habitual trust correlates with acquiescence also when interacting with less trustworthy others, these alternative accounts are unlikely to explain our findings.
One limitation of our findings is that we did not directly test the motivational and cognitive mechanisms we propose by which habitual trust influences acquiescence. However, on the basis of previous research, we expect that the motivational and cognitive consequences of habitual trust (e.g. Mayo, 2015) are also operant in the case of acquiescence. Future research should nevertheless determine under which circumstances motivational versus cognitive factors play a larger role in the relation between habitual trust and acquiescence.
12.2 Implications and future directions
Much research has been done to identify, quantify and reduce acquiescent response biases (Krautz & Hoffmann, 2018). Many of these efforts are time-costly to administer (Baumgartner & Steenkamp, 2001; Couch & Keniston, 1960). Other approaches require creating additional items, which incur potential problems caused by common method variance (Podsakoff, 2003), or are limited to knowledge-based items (Krautz & Hoffmann, 2018). Within these approaches, most research has focused on response biases that originate from cultural differences (Krautz & Hoffmann, 2018; Marin et al., 1992).
Unlike existing approaches, the present research focuses on the detection of individuals who are more likely to acquiesce. We show that habitual trust correlates with acquiescence within a given culture (Studies 2–5: US-based MTurk workers and German students) and controlling for country differences (Study 1). Thus, we investigate individual differences that relate to acquiescence over and beyond cultural differences. Notably, habitual trust can be easily and reliably assessed by using short scales that consist of fewer than ten items (e.g. Costa & McCrae, 2008; Evans & Revelle, 2008; Rotter, 1967). With such scales, researchers could identify individuals or populations that score high on habitual trust. By assessing an individual's or a population's habitual trust, one may then have an indication of when the use of––however imperfect––existing corrective measures against acquiescent responding might be especially necessary because the likelihood of acquiescent responding could be increased.
One may further speculate whether researchers should avoid signals of explicit trustworthiness in their studies and experiments. The participant recruitment process often capitalizes on trustworthiness cues, such as university logos and researcher titles, to increase participation rates. Although our studies do not test such an effect, one may wonder whether trustworthiness cues might, however, foster the researcher's perceived trustworthiness. Our results in Study 5 suggest that the more a researcher is perceived as trustworthy, the more participants might agree with their statements (and questions). Thus, future research could test whether recruitment materials that use trustworthy cues to attract participants increase acquiescence by increasing the perceived trustworthiness of the person asking the questions. If this is indeed the case, one implication would be to use trustworthiness cues in the recruitment flyers to increase participation rates, but to ensure that the actual questionnaire (e.g. web page or paper questionnaire) is designed in a more neutral manner, without explicit trustworthiness cues, as a means to reduce acquiescence.
Posts
