Running head: The seduction of easiness
The seduction of easiness: How science depictions influence laypeople’s reliance on their own evaluation of scientific information[1]
Lisa Scharrera*, Rainer Brommea, M. Anne Brittb & Marc Stadtlera
a Department of Psychology, University of Muenster, Fliednerstrasse 21, 48149 Muenster, Germany, Email: ; ;
b Psychology Department, Northern Illinois University, 363 Psychology-Math Building, Dekalb, IL60115, USA, Email:
* Corresponding author: Westfälische Wilhelms-University, Department of Psychology, Fliednerstrasse 21, 48149 Muenster, Germany. Email: ; Tel.: +49 251 8339379; Fax: +49 251 8339105
Acknowledgments
This research was supported by the Deutsche Forschungsgemeinschaft (DFG), grant BR1126/6-1.
Abstract
The present research investigated whether laypeople are inclined to rely on their own evaluations of the acceptability of scientific claims despite their knowledge limitations. Specifically, we tested whether laypeople are more prone to discount their actual dependence on expert knowledge when they are presentedwith simplified science texts. In two experiments, participants read scientific arguments that varied in comprehensibility and type of argument support and therefore in apparent easiness. We assessed participants’ inclination to rely on their own evaluation rather than deferring to expert advice when judging argument persuasiveness. The results showed that laypeoplewere more strongly persuaded by apparently easy arguments than by difficult ones. Furthermore, they were more confident in their own evaluation of the information and less inclined to turn to an expert for decision-making support after reading easy compared to difficult arguments.
Keywords: knowledge evaluation; expertise; argument comprehensibility; causal explanations; evidence
- Introduction
Laypeople frequently encounter situations where they must make decisions about science-related issues. Whether making up one’s mind about undergoing a particular medical treatment or deciding whether certain behaviors are detrimental to the environment, laypeople must judge scientific claims and the arguments provided to support such claims. While the Internet has reduced problems of access toscientific information, a major challenge remains in the evaluation of this information, regarding its acceptability, usefulness and sufficiency for solving a problem at hand (Blair & Johnson, 1987; Brand-Gruwel & Stadtler, 2011;Bromme, Kienhues & Porsch, 2010;Goldman, 2011; Kienhues, Stadtler & Bromme, 2011; Mason, Ariasi Boldrin, 2011; Wiley et al., 2009). One reason why evaluation presents such a challenge lies in the specific nature of science knowledge. Advances in science and technology have led to enormous growth of scientific knowledge, accompanied by specialization and differentiation. To manage this complexity, scientific knowledge is organized into different disciplines that are represented by specialized experts (Keil, Stein, Webb, Billings & Rozenblit, 2008), constituting a‘division of cognitive labor’ (Keil et al., 2008, Putnam, 1975; Stehr, 1994).
Due to such division of cognitive labor in modern societies, we remain laypeople throughout our whole lifetime regarding most topics and domains of science. Laypeople generally lack the high level of background knowledge and specialized training required to come to a well-founded decision about the veracity and thus the appropriate persuasiveness of scientific claims. This is what sets laypeople apart from ‘experts’,who are specifically trained and experienced in a domain, and from ‘novices’,who are not yet experts, but strive to become experts through relevant training and will already have acquired basic domain knowledge and a certain discipline-based perspective (Bromme, Rambow & Nückles, 2001).
Hence, when laypeople are in a situation which requires them to decide about a complex science-related issue, they are confronted witha paradox: On the one hand they need to come to a decision but on the other hand they lack the expert knowledge necessary for making well-founded decisions. This problem has been discussed predominately in health and medical research focusing on patient decision-making. In this context, the ideal of ‘informed decision-making’ has been proposed, according to which individuals make their own judgment (for example about a certain treatment option) only if they have acquired sufficient knowledge and information about the issue at hand to do so (Bekker et al., 1999; Charles, Gafni & Whelan, 1997). Such information might be delivered by the expert (the doctor in charge) or the layperson might find it elsewhere, nowadays mostly on the Internet (Fox, 2005).Even if this ideal of ‘informed decision-making’ is realized, laypeople must judge if their understanding of the issue is sufficient, or if it is still necessary to defer to an expert. Judging which of these two strategies is the appropriate course of action logically requires that the individual can correctly assess the sufficiency of their own knowledge for the decision task at hand. In other words, they have to successfully engage in metacognitive calibration by forming an accurate awareness of what they know and what they do not know (Glenberg & Epstein, 1985; Pieschl, 2009).
Unfortunately, as Keil (Keil et al., 2010) and his group have shown, laypersons tend to overestimate their understanding of complex science- and technology-based issues.If laypeople fail to recognize their own knowledge limitationsthey might falsely assume that their understanding of the gathered information related to a certain knowledge claim enables them to evaluate the persuasiveness of the claim. As a consequence, they might underestimate their actual need to defer to an expert to reach a truly informed decision. Conversely, only if individuals feel they lack sufficient understanding of the matter should they refrain from relying on their own evaluations. Ascertaining the conditionsunder which laypeople become aware of their own epistemic limitations when they process science-based text information remains an open empirical question. In this investigation, we are interested in thecharacteristics of science textsthat might affectthe development of such awareness.
Judging scientific claims is no simple and straightforward act.Persuasiveness itself is a graded concept. Persuasiveness of arguments can be indicated by intensity of claim agreement and evaluation of argument strength, two measures that feature prominently in persuasion research (O’Keefe, 2002). Both can be expected to be closely related but are nevertheless not equivalent. While lay recipients may perceive the evaluation of argument strength as a more abstract, comparably ‘safe’ task without any major consequences, expressing agreement with a claim must be approached more cautiously, since claim agreement might be regarded as closely connected to actual behavioral consequences.
To summarize, laypeople’s reliance on their own evaluations should manifest itself in their inclination to be easily persuaded by provided information. In contrast, if recipients do not feel competent to decide about information quality, they should be more hesitant in their judgments and refrain from indicating a strong opinion. Furthermore, an inclination to rely on their own evaluations should be reflected even more directly by laypersons’ confidence in their own decision about a claim. Strong confidence in a decision should manifest itself in high levels of trust in theown judgment and conversely in a weak desire to turn to an expert for decision support. This pattern is consistent with the concept of informed decision-making (Bekker et al., 1999), which implies that only if recipients feel that they are sufficiently informed and competent to form an opinion about the subject matter should they let themselves be persuaded by a given argument.
1.1 Characteristics of science information
It may be particularly difficult for laypeople to recognize that they should not rely on their own judgment when they encounter scientific information written in a way that makes the subject matter appear less complex and easier to evaluate than it actually is. Such is the case when laypeople read scientific information especially prepared for their consumption, i.e. presented in a simplified way in order to make it superficially comprehensible for the lay public(Kajanne & Pirttilä-Backman, 1999; Wagner, Elejabarrieta & Lahnsteiner, 1995; Zimmerman, Bisanz, Bisanz, Klein & Klein, 2001). When laypeople encounter simplified texts, their sense of understanding may mislead them to consider the subject matter easy and uncomplicated and to judge their mental representations of the described phenomena as more complete and accurate than they actually are (cf.Goldman & Bisanz, 2002). Such an impression may manifest itself in the conviction that their knowledge and skillsdo not differ meaningfully from that of an expert in the field. In this case, laypeople would fail to differentiate between the background knowledge required to comprehend presented information and the background knowledge required to evaluate the acceptability of the information for justifying a given claim. Comprehending simplified information might thus lead laypeople to regard the described phenomena and mechanisms to be so transparent and self-evident that they themselves are able to evaluate the information (cf. Keehner & Fischer, 2011) and that deferring to an expert is an unnecessary waste of time and energy. Therefore, if presented with a simple account, laypeople should be inclined to rely on their own content evaluation. If, however, the information is presented in a way that does not conceal the inherent difficulty of the subject matter, then laypeople should refrain from relying solely on their own judgement about information persuasiveness (cf. Reber & Schwarz, 1999; Schwartz, 2004).
1.2 Influences on perceived easiness of scientific contents
The assumption, that easy informationcauses recipients to rely more strongly on their own evaluation of scientific claims,raises the question as to what characteristics make scientific information appear easy to laypeople. We presume that perceived easiness is influenced by at least two message characteristics: (1) comprehensibility of the depicted information and (2) type of argument support which is presented to back a claim. Our assumption is based on theoretical considerations from previous literature and empirical findings which show both characteristics to have an impact on the persuasiveness of arguments (e.g. Brem & Rips, 2000; Eagly, 1974; Slusher & Anderson, 1996).
1.2.1 Information comprehensibility
Previous findingshave shown that people agree more strongly with claims presented as part of comprehensible arguments than with those that are difficult or impossible to comprehend (e.g. Bradley & Meeds, 2004; Murphy, Long, Hollerana & Esterly, 2003). However, research on this issue has mainly dealt with arguments supporting moral claimsand arguments advertising the usefulness of consumer products. Due to this focus on moral claims and advertisement, it remains unclear whether the observed comprehensibility effect also applies to scientific arguments. In fact, the question of whether it generalizes is not trivial. In the case of moral arguments or arguments intended to advertise everyday products,recipients may perceive their own values and personal experiences as a sufficient basis for argument evaluation. However, in the case of scientific informationlaypeople might be aware that its evaluation is usually beyond their capabilities—regardless of whether they feel that they understand a piece of information.If this is the case, laypeople should not be inclined to agree more strongly with a scientific claim supported by comprehensible than by incomprehensible information.
Two studies did in fact investigate the persuasive effect of comprehensibility for scientific arguments. While Miller et al. (1976) detected no difference in persuasiveness between comprehensibility conditions, Eagly (1974) found incomprehensible arguments to cause weaker claimagreement than comprehensible arguments.However, neither study adequately answers the question of whether laypeople are more persuaded by arguments they perceive as easy. In Miller et al.’s (1976) study, it remains unclear whether the manipulation of comprehensibility was indeed successful, as no manipulation check was conducted. Furthermore, Eagly’s (1974) study investigated a situation of expert–novice communication rather than of expert–layperson communication. Participants (psychology students) were required to evaluate contents from their own field of study and, due to their prior training, may have perceived themselves asprincipally able to evaluate the given message.
In sum, previous findings on moral issues and advertisement support the notion that comprehensibility has an influence on argument persuasiveness. However, it is not clearwhether comprehensibility also affects the persuasiveness of scientific arguments and whether any persuasiveness effects of comprehensibility extend to laypeople’s confidence in their own judgments of the provided information.
1.2.2 Type of argument support
Previous research has differentiated between two types of argument support that can back a causal claim:causal support and empirical support. Causal support explains the mechanism underlying the claimed causal connection (e.g. ‘Beta blockers decrease high blood pressure, because they inhibit the activity of hormones which cause blood pressure-raising stress reactions in the body.’). In contrast, empirical support backs the claim by referring to relevant statistical data(e.g. ‘Beta blockers decrease high blood pressure, because blood pressure was decreased by 20-28% among patients who regularly took beta blockers.’) (for comparable distinctions see Brem & Rips, 2000; Koslowski, 1996; Sandoval & Cam, 2011; Slusher & Anderson, 1996).In other words, whereas causal support answers the question ‘Why are two concepts causally related?’, empirical support answers the question ‘How do we know that two concepts are related?’ (Glassner, Weinstock & Neuman, 2005). Kuhn (1991) considers only evidence (i.e. empirical support) as ‘genuine evidence’ while discounting explanations (i.e. causal support) which are not complemented by additional data as ‘pseudoevidence’. According to Kuhn, explanations serve to establish the plausibility of a claim, which is neither necessary nor sufficient to supporting its believability. Rather, it is only evidence that provides direct support for the correctness of a causal claim. This notion is also prevalent in the empirical sciences, where themost highly regarded justification for claims is empirical evidence.
However, despite the prominent role that evidence plays in empirical science, previous literature regards explanations as a particularly persuasive form of argument support (Gopnik, 2000; Lombrozo, 2006; Trout, 2002) and suggests that laypeople prefer causal support as claim justifications, possibly because its evaluation is perceived as easier. According to Keil (2006, 2010), individuals have a sophisticated sense for causal relations and structure and seek out explanations. These activities formthe essence of individuals’ folk science. Thus, laypeople may consider causal support as easier to evaluate than empirical support,since causal support more closely reflects the kinds of claim justifications they consider in everyday life. This assumption is also in line with Brem and Rips’(2000) reasoning thatexplanations provide a method for vetting claims by allowing individuals to judge the described mechanisms for internal coherence and consistency with prior knowledge. Due to allowing this kind of vetting, laypeople may perceive explanations as adequate subjects for their own evaluation. Laypeople should therefore be more easily and confidently persuaded by arguments containing causal support than by arguments containing empirical support.
The persuasive effect of causal support compared to empirical support has been investigated in several studies.However, while findings by Slusher and Anderson (1996) indeedconfirmthat arguments containing causal supportcause greater claim agreement than arguments containing empirical support,Brem and Rips (2000)conversely found recipients toprefer evidence over explanations.Similarly, in astudy by Sandoval and Cam (2011),8 to 10 year old children were shown to slightly prefer empirical over causal support. Overall, findings on the persuasive impact of scientific causal and empirical support indicate that there is indeed a difference in persuasiveness between the two support types. Although theoretical considerations suggest a persuasive advantage of causal over empirical support from a layperson’s point of view, results are mixed as to the direction of the persuasive difference.A possible explanation for the inconsistency of findings is that type of argument supportand comprehensibility might have been confounded in at least some studies. In cases where empirical supporthad been more comprehensible than its causal counterpart, the perceived easiness of comprehensible support might have outweighed the easiness ascribed to causal support. Moreover, and similar to the state of affairs regarding comprehensibility, the effect of support type on recipients’ confidence in their decision about information persuasiveness has not been investigated directly. Thus there is a need to assess whether and how support type influences laypeople’s readiness to rely on their own evaluations.
1.3 The present research
In twoexperimentswe investigated how features of science depictions influence lay recipients’ inclination to rely on their own evaluations of scientific information. Therefore, we investigated how scientific argumentative texts with different levels of comprehensibility and argument support influence argument persuasiveness and laypeople’s confidence of their claim agreement decision.
We expect that if scientific information is presented in a way that makes it difficult for laypeople to process, they will be more likely to realize that as non-experts they are in fact unable to decide whether the information poses a sound argument to support a given claim.
Specifically, we hypothesizedthat recipients would be more intensely persuaded by comprehensible than incomprehensible arguments. Therefore, reading comprehensible arguments should lead to greater agreement with the claim than reading incomprehensible arguments (hypothesis 1a) and laypeople should evaluate comprehensible arguments as stronger (i.e. more supportive of the claim) than incomprehensible arguments (hypothesis 1b). Analogously, based on the assumption that causal support is perceived as easier by laypeople, arguments containing causal support should be more persuasive than arguments containing empirical support. Therefore, we expect that reading causal supportwill lead to stronger claim agreement than empirical support (hypothesis 2a) and moreover, that laypeople will evaluate arguments containing causal support as stronger than those containing empirical support (hypothesis 2b).
We also expect that laypeople would be more confident in their own decision about the acceptability of provided scientific information when information processing is easy than when it is difficult. Therefore, comprehensible arguments should causelaypeople to have greater trust in their own decision about the claim (hypothesis 3a) and conversely a weaker desire to consult an expert for further decision-making support than they would have with incomprehensible arguments (hypothesis 3b). Similarly, receiving arguments with causal support should lead lay recipients to trust more strongly in their own decision about the claim (hypothesis 4a) and to have a weaker desire to consult an expert than when they receive arguments with empirical support (hypothesis 4b).
- Experiment I
To investigate our hypotheses, we conducted a first experiment in which participants read texts about scientific topics which provided support for a claim. The extent to which laypeople were persuadedby the provided information and their confidence in the claim agreement decision were assessed.