《社会心理学》课程教学资源（文献资料）conformity——Speaking From Ignorance - Not Agreeing With Others We Believe Are Correct

INTERPERSONAL RELATIONS AND GROUP PROCESSES Speaking From Ignorance:Not Agreeing With Others We Believe Are Correct Bert H.Hodges Gordon College and University of Connecticut ERM PaietNao e en Oneht y(HodgesGeyer,2006)predicts that people will sometime n an asch (1956 s we fo nd evidence that eople in a po n of i h 27% of the y,rather th by awid nd moral co hi相 cand devekpmentaoeychology Keywords:conformity.divergence.pragmatics.truth.values distasteful or celebrities whose reputations we think are over gnorance.especially if they are politicians t d ve they are talking with people who knov more about th opic than they themse Suppose.for exampl you Iin what they see with you.If you were asked to speak,would you (a) gnorance. Humans must regularly act from ignoran For xam 1B something different?A tourist visits ible thing to do.it appear is to follow v more tha ve do heir heloful or ear of this ble n ng guided by th (e.g 218

INTERPERSONAL RELATIONS AND GROUP PROCESSES Speaking From Ignorance: Not Agreeing With Others We Believe Are Correct Bert H. Hodges Gordon College and University of Connecticut Benjamin R. Meagher University of Connecticut Daniel J. Norton Boston University Ryan McBain Harvard University Ariane Sroubek Yeshiva University Values-pragmatics theory (Hodges & Geyer, 2006) predicts that people will sometimes disagree with others they believe are correct, for reasons similar to those explaining agreement with incorrect answers in an Asch (1956) situation. In 3 experiments, we found evidence that people in a position of ignorance sometimes do not agree with the correct answers of others in positions of knowledge. Experiments 1a and 1b found this speaking-from-ignorance (SFI) effect occurred 27% of the time. Experiment 2 introduced experimental controls and self-report data indicating that the SFI effect (30%) was generated by realizing values (e.g., truth, social solidarity) and pragmatic constraints to act cooperatively, rather than by a wide array of alternatives (e.g., normative pressure, reactance). Experiment 3 experimentally manipulated concern for truthfulness, yielding 49% nonagreeing answers, even though there were monetary incentives to give correct, agreeing answers. The overall pattern suggests that people are not so much conformists or independents as they are cooperative truth tellers under social and moral constraints. Results, while surprising for social influence theories, illustrate the dynamics of divergence and convergence that appear across studies in cultural anthropology and developmental psychology, as well as in social psychology. Keywords: conformity, divergence, pragmatics, truth, values We often find ourselves laughing at people who speak from ignorance, especially if they are politicians whose views we find distasteful or celebrities whose reputations we think are over￾blown. In occasional moments of insight, though, we may realize that we too have talked about a topic about which we knew little or nothing as if we were well informed, all the while showing our ignorance. The question we address in this article is, “What do people do when they are invited to speak from ignorance and they believe they are talking with people who know more about the topic than they themselves?” Suppose, for example, you find yourself in a situation in which other people can see what is going on, while you cannot, but these more knowledgeable others share what they see with you. If you were asked to speak, would you (a) make up your own answer or (b) repeat what others have told you? Of course, the issue is much broader than just speaking from ignorance. Humans must regularly act from ignorance. For exam￾ple, an individual is new to farming and does not know what to plant: Should she plant what her more experienced neighbors plant or something different? A tourist visits another culture about which he knows very little: Should he follow the lead of others who are more culturally sophisticated or just wing it? The emi￾nently sensible thing to do, it appears, is to follow the lead of others who know more than we do (D. T. Campbell, 1990). When information is absent or ambiguous for an individual and others provide what is believed to be trustworthy information, the only sensible response seems to be conformity. Being guided by those with expertise is a fundamental tenet of cultural evolution (e.g., Bert H. Hodges, Department of Psychology, Gordon College, and De￾partment of Psychology, University of Connecticut; Benjamin R. Meagher, Department of Psychology, University of Connecticut; Daniel J. Norton, Department of Psychology, Boston University; Ryan McBain, School of Public Health, Harvard University; Ariane Sroubek, Ferkauf School of Graduate Psychology, Yeshiva University. Support from Gordon College (Initiative Grant) and the National Science Foundation (BCS-0843219) to Bert H. Hodges is gratefully acknowledged. Earlier versions of parts of this article were presented to the meeting of the Society of Experimental Social Psychology, Portland, Maine (2009, October). We thank Alyssa Barkley, Justina Cassidy, Timothy Couper, Ryan Daley, Jenni Damiani, Jennifer Davies, Alice Dennis, Dustin Foss, Michelle Halla, Nathan Josephs, Joyce Lacy, Jenna Meisenhelder, Aaron Noga, Katherine O’Donnell, Valerie Plichta, Rylee Rainwater, Landon Ranck, Zachary Reyn￾olds, Jocelyn Rioux, Aashley Thompson, Greg Walter, and Ann-Elise Wright, among others, for their crucial contributions in various phases of the research. We are grateful to Reuben Baron, Marilynn Brewer, Stephen Cowley, Erin Devers, Stephanie Dibble, Jonathan Gerber, John Levine, Kerry Marsh, Mi￾chael McCullough, Paul Nail, Radmila Prislin, Stephen Rizzo, and Anne Zell for their helpful comments on the studies or earlier versions of this article. Correspondence concerning this article should be addressed to Bert H. Hodges, Department of Psychology, Gordon College, Wenham, MA 01984. E-mail: bert.hodges@gordon.edu This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Journal of Personality and Social Psychology, 2014, Vol. 106, No. 2, 218 –234 © 2014 American Psychological Association 0022-3514/14/$12.00 DOI: 10.1037/a0034662 218

SPEAKING FROM IGNORANCE 219 Mesoudi..5).Nearly every theory in Accordng to ales-pragmaics thory. There are man ns to do so.Two of the relationship are con be th with others,and (b)people t to be accepted and ple,the field of safe driving is not a p of the driv C.g. 1979 rs tailor-made for conformity Hov and other objects.Thus,values stitute real on and a uld often cho e not to agre with the c of mo vary i ng so not ding to and rea g pa nd thers in a y,and trust.The experiments d a hypothesis explicitly ny oth r psychological activity,is neve out a single that has proved valuable for explaining Asch's (195 on in which he used a task would be ome intoler and thus inefficic d小wnan nt for ing (Hodges&B n.192).These multiple valu are hete ght be g when the hical.a term first used by Warre by all the valt vary in their ordering for the s of the In drivine for e may take pric Values-Pragmatics Theory place ut then re night lead to i Values tical approach to og ha ents to he and ability to identify bility to recognize and the goods (ie of clanf a lication to Asch' s the desi (19)e ation we use in this.According to the the ory.values are the n e that ion) sider.for e mple.the act of The Dilemma of speaking from knowledge field of safe travel (Gibson Crooks,1938),is a very rea vell;the participants questions fron a position other participants gave the same wrong answer.Although the others values defipe the field of action and are obligatory:Unles sch(1952. 12)himself offered safery,travel (i.e.,freedom of movement),and oth h in comm 86)ecological psychology

Mesoudi, 2009; Richerson & Boyd, 2005). Nearly every theory in social psychology, especially those about social influence, would agree. There are many compelling reasons to do so. Two of the most powerful are believed to be the following: (a) People want to be correct, and the most likely way to be correct in this situation is to agree with others, and (b) people want to be accepted and liked by others, and one of the most reliable ways to do this is to agree with other people’s views (e.g., J. D. Campbell & Fairey, 1989; Claidière & Whiten, 2012; Deutsch & Gerard, 1955). The speaking-from-ignorance (SFI) situation, described earlier, appears tailor-made for conformity. However, we argue that it is considerably more complicated than it first appears. More specif￾ically, we predicted that individuals asked to speak from ignorance would often choose not to agree with the correct answers of more knowledgeable others. Their reasons for doing so would not be stupidity, stubbornness, or independence. Rather, their failures to agree would be generated by pragmatic considerations for speak￾ing to others in a cooperative and pragmatically warranted way and by their seeking to acknowledge values, particularly truth, social solidarity, and trust. The experiments tested a hypothesis explicitly posed by values-pragmatics theory (Hodges & Geyer, 2006), an account that has proved valuable for explaining Asch’s (1951, 1956) well-known studies, in which he used a task that seems tailor-made for dissent. The overarching argument we offer is that a single set of complex dynamics can account for people some￾times agreeing when they might be expected to disagree (i.e., Asch situation) and sometimes disagreeing when they might be expected to agree (i.e., the SFI situation). We begin by introducing selected aspects of values-pragmatics theory, illustrating it, first, using driving as an example, then applying it to Asch’s (1956) speaking-from-knowledge situation. Then we turn to a more careful analysis of the SFI situation. Values-Pragmatics Theory Values-realizing theory is a general theoretical approach to perception, action, and cognition, which has been applied to perception-action tasks (e.g. Hodges, 2007b; Hodges & Lindhiem, 2006), social and developmental issues (e.g., Hodges & Baron, 1992; Hodges & Geyer, 2006), and language and cognition (e.g., Hodges, 2007a, 2009). Applied to social communicative contexts, values-realizing theory has sometimes been referred to as values￾pragmatics theory (Hodges & Geyer, 2006), and that is the desig￾nation we use in this article. According to the theory, values are the real goods that define fields of action within ecosystems; that is, they constrain what is possible to do that makes an activity worth doing. Consider, for example, the ecosystem of driving. The act of driving brings into existence a field (in the sense used in physics) that guides the driver’s actions. This field, which has been called the field of safe travel (Gibson & Crooks, 1938), is a very real physical field that opens up before the vehicle, is bounded by obstacles (e.g., buildings, trees, other vehicles), and is continu￾ously changing its size and shape. It is also a social field, since it depends crucially, often moment to moment, on the actions of others. Values define the field of action and are obligatory: Unless safety, travel (i.e., freedom of movement), and other values such as accuracy and tolerance (e.g., distance between vehicles) are real￾ized sufficiently, the field will literally cease to exist (Hodges, 2007b). According to values-pragmatics theory, values are not located in persons or in objects but are better thought of as real demands placed on relationships, so that action and perception are con￾strained by values, regardless of whether or not there is any explicit awareness or acknowledgment of those demands. For example, the field of safe driving is not a projection of the driver (e.g., it exists even if the driver is not looking), nor does it exist in cars, roads, or other objects. Rather, it exists as a dynamic field of possibilities “for good or ill” (i.e., affordances; Gibson, 1979/ 1986, p. 127) that obligate the driver’s ongoing actions relative to cars, roads, and other objects. Thus, values constitute real onto￾logical, epistemic, and ethical demands on possible actions, and as such, they cannot be equated with social norms, personal prefer￾ences, or biological needs. Individuals or groups may vary in how skilled they are in attending to and realizing particular values, but individuals and groups do not “possess” values. Rather, values define the goods that make their existence and action possible.1 According to values-pragmatics theory (Hodges, 2007b), any psychological activity is constrained by multiple goods. Driving, or any other psychological activity, is never about a single value, such as safety, or accuracy, or efficiency. If one cared only for safety, one would never drive; if one cared only for accuracy, driving would become intolerant and thus inefficient and unsafe. As a values-realizing activity, driving cannot be reduced to goal seeking (Hodges & Baron, 1992). These multiple values are heter￾archically organized. Heterarchical, a term first used by Warren McCulloch (1945), means that actions are mutually constrained by all the values, so that there is no fixed, hierarchical ordering of values; rather, across time and task, values vary in their ordering for the sake of the joint realization of all the values (i.e., the ecosystem as a whole). (Hodges, 2009, p. 632) In driving, for example, speed and accuracy may take priority over safety at one place or time but then reverse at another. These are not simple tradeoffs: Sometimes safety might lead to increased speed and increased accuracy (tighter tolerances). Good driving is a continuous balancing act, making ongoing adjustments to realize all the values. It is possible, of course, to engage in bad driving, but there are very real limits, and our very ability to identify such misbehavior depends on our implicit ability to recognize and enact the goods (i.e., values) of driving. To illustrate and clarify the worth of this way of characterizing values, we review its application to Asch’s (1951) social dilemma, in which he had people answer questions that others had answered incorrectly (see Hodges & Geyer, 2006, for an extended discus￾sion). The Dilemma of Speaking From Knowledge In Asch’s (1956) experimental paradigm, everyone could see the visual information about which they were being asked equally well; the participants answered questions from a position of knowledge. The only difficulty, of course, was when all of the other participants gave the same wrong answer. Although the 1 Asch (1952, chapter 12) himself offered an account of values that has much in common with values-pragmatics theory (Hodges, 2007b), al￾though he was influenced by the Gestalt tradition (e.g., Asch, 1968; Köhler, 1937/1966), while Hodges and Baron (1992) developed their account in relation to Gibson’s (1979/1986) ecological psychology. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. SPEAKING FROM IGNORANCE 219

220 HODGES.MEAGHER.NORTON.MCBAIN.AND SROUBEK studies are famous for the ants agreec ay be the u bigg n.To majority about wo thirds of the time (e.g.. epeat th others is to ackn wledee tha to dise casionall plain.Partic This hasi und ental problems for the rative intentions n the (Hodges Geyer.200 eing som with i 70% )and one's open to hat the can plau sibly acc es but suffe ant or It is because of the he n the say in an awk ho do not care what others think or if the people who of ght h to whor s add truthfully) subm ting to that mmon eround for commu ted C9. ied Asch's dile the gh to c ist that th ad values must he realized tha ing even i ing pra ing an m)seems to do oned the co xP sey.200 1000 that Aschng negat it is actually cial dile wrong but be The Dilemma of Speaking From Ignorance stake in the Asch ticipants do not have ton ava h views of self and others without deprecating eithe To sce more clearly hov ources.which w for of the vsiuatedurc re 12 critical trials which allow inants to yary their behavion mple Grice's (1975)well-knowr han being trapped by an alor-none dile ales).such as honesty and at stake in the Asch situatio in The pragmat nter,on own p hich a that for which you (Grice.1975).A em ns in the of tr can be captured by mere accuracy.Thus.the situation calls for sider the answers of others to be adequate evidence:the othe

studies are famous for the number of times participants agreed with wrong answers (e.g., Moscovici, 1985), the bigger surprise may be the much larger number of times they dissented from the unanimous majority, about two thirds of the time (e.g., Friend, Rafferty, & Bramel, 1990; Harris, 1985). The most common behavior was to dissent most of the time but to agree occasionally, captured in the median response of nine dissenting answers and three incorrect answers on critical trials. This basic result presents fundamental problems for the usual accounts offered to explain the results (Hodges & Geyer, 2006). Normative accounts simply do not address the massive amount of truth-telling dissent observed in Asch situations (over 70% of participants always or mostly dissent). Information-based ac￾counts, which assume that the dilemma is about epistemological uncertainty, can plausibly account for diverse responses but suffer from findings that there is a sharp reduction in agreeing answers when answers are not publicly shared with one’s peers (Asch, 1955, 1956). Results of this sort suggest that the dilemma is less a matter of epistemic confusion than it is a struggle to know what is appropriate to say in an awkward, tense situation. Asch (1955, 1956) thought he had designed a moral dilemma, an all-or-nothing choice between right (i.e., answering the experi￾menter’s question truthfully) and wrong (i.e., submitting to con￾sensus rather than maintaining one’s independence). Many psy￾chologists since Asch seem to have adopted a zero-tolerance standard (Friend et al., 1990; Krueger & Funder, 2004) regarding agreement with wrong answers: A single agreement with a wrong answer is enough to count one as a conformist, implying that the person is not dissenting even if he or she disagrees 11 times out of 12. Effectively, this all-or-nothing approach renders dissent as uninteresting and uninformative. Numerous researchers have questioned the conformity account of Asch’s (1956) results (e.g., Friend et al., 1990; Harris, 1985; Hodges & Geyer, 2006; Krueger & Massey, 2009; McCauley & Rozin, 2003). One of the most interesting of these is D. T. Campbell (1990), who argued that Asch was wrong to view consensus in such a negative way; it is actually a necessary good, especially when viewed from an evolutionary standpoint. Thus, Asch’s dilemma might better be regarded as a choice not between right and wrong but between multiple goods in tension. Following Campbell, Hodges and Geyer (2006) claimed that there are at least three salient values at stake in the Asch situation: truth (i.e., honestly expressing one’s own view), trust (i.e., acknowledging the value of others’ views), and social solidarity (i.e., integrating the views of self and others without deprecating either). Seeing the dilemma as a tension among values might naturally lead one to think of the situation in terms of tragic tradeoffs, but Hodges and Geyer proposed that there are pragmatic resources, which allow for a more creative, balanced response. One such resource is that there are 12 critical trials, which allow participants to vary their behavior rather than being trapped by an all-or-none dilemma. A values-pragmatic analysis suggests a complex array of rela￾tionships and obligations are at stake in the Asch situation: “How does one speak the truth in a complex, tense, and frustrating situation . . . in a way that simultaneously honors one’s peers, the experimenter, one’s own perception, and the situation in which all are embedded?” (Hodges, 2004, p. 344). The nature of truth, as it functions in the Asch situation, is larger and more complex than can be captured by mere accuracy. Thus, the situation calls for people to speak truthfully but in a way that reveals something of the awkwardness and tension of the situation and that is respectful of others’ views, even if one disagrees sharply with them. To repeat the wrong answers of others is to acknowledge that one hears and understands their point of view, even if one has other￾wise made one’s disagreement plain. Participants can realize mul￾tiple values in an inherently frustrating situation by varying pat￾terns of dissent and agreement to communicate larger scale truths and cooperative intentions. Agreeing some of the time with incorrect answers can function as a pragmatic signal of one’s commitment to taking others’ views seriously (i.e., social solidarity) and one’s openness to further conversation about the situation. If a person always dissented from a group’s expressed views, it would be easy for that person to be seen as arrogant or dismissive. It is because of the heterarchical relations among values that those participants who offer truthful dissent will increasingly be constrained to offer a sign of their trust and social solidarity with those from whom they are dissenting. Dissent, after all, cannot function if it is directed toward people who do not care what others think or if the people who offer it have no concern for those to whom the dissent is addressed. Dissent implicitly appeals to some sense of shared concern for truth and other goods that provide a common ground for communicative discourse and social interaction. The complexity of the physical, social, and moral dynamics embodied in Asch’s dilemma is not unlike the complexity that drivers face in negotiating the physical, social, and moral hazards of the road. Values must be realized continuously, in ways that require ongoing pragmatic judgments and adjustments. The dom￾inant pattern observed, mostly disagreeing but sometimes agreeing (e.g., the 9/3 pattern), seems to do just this, as does the diversity of responses in the experiment as a whole (Asch, 1956; Hodges & Geyer, 2006). From this perspective, it was wrong for Asch to assume that an individual speaking to others with whom he or she has a sharp disagreement should say exactly the same thing he or she would have said when alone with the experimenter. Acting in a social dilemma requires sensitivity and finesse, not simply inde￾pendence and honesty. The Dilemma of Speaking From Ignorance At first glance, the SFI situation seems quite unlike the Asch situation: Participants do not have visual information available that contradicts information provided by other witnesses. From this perspective, there should be no dilemma. To see more clearly how the SFI task could pose a dilemma for participants, however, consider the pragmatics of the situation. Pragmatics, which is the study of the contextual appropriateness of situated utterances, can itself be understood as grounded in values (Grice, 1991). For example, Grice’s (1975) well-known cooperativeness principle is an attempt to delineate key demands (i.e., values), such as honesty, economy, and coherence, that speakers and listeners must realize if they are to function effectively in conversations. The pragmatic cooperativeness that is entailed in ordinary conversations usually involves the following: saying neither what you believe to be false nor that for which you lack adequate evidence (Grice, 1975). An SFI situation pulls and twists these two aspects of cooperation inside out, creating a frustrating tension. One solution is to con￾sider the answers of others to be adequate evidence; the other This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 220 HODGES, MEAGHER, NORTON, MCBAIN, AND SROUBEK

SPEAKING FROM IGNORANCE 221 solutions is ntirely satisfactory Thus there is likely to he insta y.trust will tend to erate a ninetcmdcncicsbcreeoposiblcotetionsb ding nses to the dilemma ers unless the situati is one in which the ocial solidarit e also at play in the SFI dilemma Asch(1952. lons,it is onable to expect that most peopl e most of the tim 1955)thought it bvious that on should trust one's eyes in raint to speak truthfully and with epi t seems obvious that one should trust others and not oneself both m to th firs s possible re was te d ha Wha t thi d eak ba ed what they themselves can edo1 by a zero-tolerance standard.Any agreement Although it is ofen assumed that speaking truthfully Hodges.2004 Krueger Funder. 2004)The SFI ave argued tively inverts the Asch situ ne Asch situation.we think is one of the where the most lea ect information that they nto cons disagreement.given that one had good reasons to expect with thosce onform in an SFI situation,they would be expe d always to s s of the SFI to act cooperati ely and if might nudge the same criterion used to evaluate Asch's results. tion c other Overview of Experiments hand,the SFI situation haring o the The firs la and Iby ly to second id oo atics lead people to they can mak h h I people ha so ba d on se to stly one cial olidarity involvesmo than simple con mnts (e.g.reactance.refusal to imitate)that misht explain an ed with ma ing the tegrity of the SFI effect.In a third ffective group.Thi may requir answers.as predicted by value pragmatics theory. group I Experiments la and 1b d out.di Method Procedure.Participants volunteered for a study on "visua toward agreeme nt (the others are c some of the time ald be to"give the most correct answer that you can"to each

solution is to consider guessing at what one cannot see clearly, as an honest acknowledgment of one’s ignorance. Neither of these solutions is entirely satisfactory. Thus, there is likely to be insta￾bility, with fluctuating tendencies between possible solutions, both within and across individuals, leading to considerable variability in responses to the dilemma. The pragmatic analysis of the SFI situation suggests that the same three values at work in the Asch situation—truth, trust, and social solidarity—are also at play in the SFI dilemma. Asch (1952, 1955) thought it obvious that one should trust one’s own eyes in preference to others’ in his situation; however, in an SFI situation, it seems obvious that one should trust others and not oneself. Both predictions assume, though, that what is properly motivating par￾ticipants’ choices is the goal of answering as many questions correctly as possible. Perhaps participants’ motivations are more complex than this. What if, for example, participants are motivated to speak truthfully in an SFI situation? If participants think they are being asked to speak based only on what they themselves can see, then they are likely to feel constrained to indicate their ignorance, answering incorrectly. They will have spoken truthfully but not correctly. Although it is often assumed that speaking truthfully would entail answering a question correctly, we have argued there are cases where being truthful cannot be reduced to accuracy. The Asch situation, we think, is one of those where the most compre￾hensive accounting of what is true will take into consideration the differing views of one’s colleagues and the frustrating tension of the disagreement, given that one had good reasons to expect agreement. If the pragmatics of the SFI situation make it difficult to know how to act cooperatively and if truthfulness might nudge some people some of the time to choose to reveal their ignorance, what role is played by social solidarity and by trust? Regarding trust, we assume that participants in a position of ignorance will trust others in better positions to know unless they have evidence to the contrary. Social solidarity, though, is not as straightforward. On the one hand, the SFI situation invites a sharing of views; on the other hand, participants are likely to feel there is nothing for them to say except to repeat what others have said. Ordinary conversa￾tional pragmatics lead people to expect they can make a unique contribution to a conversation based on their own point of view and experience. It is not wrong to repeat what other, better in￾formed people have told you, but it is awkward to do so based only on blind trust. Thus, it also makes sense to offer honestly one’s own perspective, even if it is likely to be incorrect. Social solidarity involves more than simple consensus or uni￾formity. It is more concerned with maintaining the integrity of the relationship among group members and advancing the welfare and effectiveness of the group. This may require complementary ac￾tions, in which each member of the group fulfills differing roles in generating the overall action of the group. As numerous authors have pointed out, disagreements can be a reflection of interdepen￾dence and commitment to a group, rather than expressions of independence or disloyalty (e.g., Hornsey & Jetten, 2004; Packer, 2008). Generally, in an SFI situation, social solidarity would move one toward agreement (the others are correct after all), but since the agreement comes without pragmatic warrant, participants might feel constrained to acknowledge their ignorance at least some of the time. Overall, it is the pragmatic relation among the various values in a particular task that can have an effect on decisions to agree or disagree. Generally, trust will tend to generate agreeing answers, but truthfulness will constrain that tendency, yielding significant amounts of disagreement. Social solidarity will generally encour￾age agreeing answers unless the situation is one in which those answers would threaten the well-being of the group or strain the relationship among its members. Given this general set of rela￾tions, it is reasonable to expect that most people most of the time will agree with correct answers of others. However, the pragmatic constraint to speak truthfully and with epistemic warrant will lead to a significant number of answers that do not conform to the correct answers of better informed others. We refer to these non￾conforming, incorrect answers as the SFI effect. This effect, first predicted by Hodges and Geyer (2006), was tested in the experi￾ments reported here as a way of testing whether people would conform when it seems like the normatively correct thing to do. How large should such an SFI effect be, if it occurs? Perhaps, the best comparison is Asch’s situation, which has been interpreted by a zero-tolerance standard. Any agreement with wrong answers is considered as constituting conformity (Friend et al., 1990; Hodges, 2004; Krueger & Funder, 2004). The SFI situation effec￾tively inverts the Asch situation. When one has no information or entirely ambiguous information and others are believed to have clear, correct information that they share, there is every reason to believe that agreement with those answers will be unanimous. Thus, we assume a zero-tolerance standard as well. If people conform in an SFI situation, they would be expected always to give correct, agreeing answers. One could argue that the normative expectations (i.e., what seems socially appropriate) are even stron￾ger in an SFI situation than in an Asch situation. If that is the case, it seems all the more appropriate for the SFI effect to be evaluated by the same criterion used to evaluate Asch’s results. Overview of Experiments The first experiments we report (Experiments 1a and 1b) were exploratory: Is there evidence for an SFI effect? In the second experiment, participants gave answers both in a position of knowl￾edge and in a position of ignorance. This manipulation allowed for a stronger test of the SFI effect and, along with self-report mea￾sures on motivations and perceptions of the task, enabled us to evaluate whether there was evidence supporting a values￾pragmatics account (e.g., concern for truthfulness). In addition, it provided various means for testing whether there were alternative accounts (e.g., reactance, refusal to imitate) that might explain an SFI effect. In a third experiment, we manipulated participants’ concern for speaking truthfully to see if this increased nonagreeing answers, as predicted by values-pragmatics theory. Experiments 1a and 1b Method Procedure. Participants volunteered for a study on “visual processing from different angles and distances” and were told that they would be placed in different positions and asked to answer questions about information projected onto the screen. Their task would be to “give the most correct answer that you can” to each This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. SPEAKING FROM IGNORANCE 221

222 HODGES.MEAGHER.NORTON.MCBAIN.AND SROUBEK ng in f of the cren.Then they were plac d in Positions A.B.and C.as shown The 橱 in the order A.B.C on Trial B.A. n Tri Bon Trial 3:repeating for the rema d ment lb oon of with one small change Stimulus materials and questionnaire. On each trial.partic .These figures.called CAPTCHAs (Camegie Mellon Un ersity d on the screen so that selected A and B which C wa nthe figur To establish th ility of igure 2 on C to se al n r”Ans ept that They were never able to ight red ut th ns of th ew g position.ho ions they th ght each participant had ans cor whole was ow muc tal layout. In order to sure that Particip dent,an ds (e C-6 participants C answered first nd.s they eive the iment as one that her answers and those B-4 uration meet pistemological criteria The critica vords but no dentify the correct answer.I A-1 clearly.Thu able to correct ereasily if they are】 approximate length of words or individual letters were discemible

question, which would require “looking and listening carefully.” They were shown one sample slide, which demonstrated that the correct word could be easily identified when sitting in front of the screen. Then they were placed in Positions A, B, and C, as shown in Figure 1, and told that they would take turns giving their answers, by position, as called out by the experimenter. They were instructed to sit straight without leaning and reminded to give the most correct answer they could. The experimenter presented 18 slides. Answers were given in the order A, B, C on Trial 1; B, A, C, on Trial 2; and C, A, B on Trial 3; repeating for the remaining trials. After answering the 18 questions, participants completed a questionnaire about the task and were debriefed. Experiment 1b was a replication of Experiment 1a with one small change in procedure to be described later. Stimulus materials and questionnaire. On each trial, partic￾ipants were asked to identify a target word, which was superim￾posed on other words and/or visual patterns, as illustrated in Figure 2. These figures, called CAPTCHAs (Carnegie Mellon University, www.captcha.net), sometimes resized or reshaped, were posi￾tioned on the screen so that selected target words could be read from Positions A and B but not from Position C. Figures mostly appeared on the upper portion of the left side, the opposite side from which C was viewing the figures. To establish the inability of individuals at Position C to see clearly, a group of 10 people from the same pool as experimental participants followed the same procedure, except that they were alone. They were never able to give a correct answer to any of the slides. After the experimental task, participants answered questions about their impressions of their viewing position, how they felt when they could not see clearly and others answered first, how many questions they thought each participant had answered cor￾rectly, how much weight they gave to others’ answers relative to their own view in deciding what to say, whether they knew other participants and what their impression of them was (e.g., trusting, suspicious), and what their impression of the experiment as a whole was. Experimental layout. In order to assure that Participants A and B gave the same, correct answer to all questions, Participants A and B were trained confederates. Their answers were deliberate, confident, and accurate. All answers were single words (e.g., sponge, mouth). The true participant, C, was the last to answer on 12 trials: These were the critical trials in which participants, having heard two better positioned persons give the same answer, were confronted by their inability to see clearly. On the other six trials, participants at Position C answered first and, since they could not see clearly, were only able to guess. These trials reduced the likelihood that C would perceive the experiment as one that was focused on the relationship between his or her answers and those of A and B. Position A was located about 6.5 m from the projected image on the screen and orthogonal to it, B was about 5.1 m at a 45° angle, and C was about 3.8 m at a 3° angle. The screen was approxi￾mately 2.5 m high
3.0 m wide, and figures generally took up less than 0.4 m2 . These details matter less than whether their config￾uration meets the proper epistemological criteria. The critical participants at C must be able to see just enough that it is plausible that they are being asked to answer questions about the projected words but not so much that they can identify the correct answer. It must also be obvious to C that the other positions are much better for seeing clearly. Thus, participants in Positions A and B should be able to give correct answers easily if they are being attentive and careful. Pretesting established that Participant C was able to see that there were wordlike forms on the screen; occasionally, approximate length of words or individual letters were discernible, screen___________________________________ C- 6 5 B- 4 3 2 A- 1 Figure 1. Positions (not to scale) of participants in Experiments 1a and 1b (A, B, C) and Experiment 2 (1, 2, 3, 4, 5, 6). In Experiment 3, Positions 2, 4, and 6 were used. Figure 2. A sample slide used in experiments (reduced to black and white). The question asked was, “What is the bottom word in the lower right hand corner?” Answer: warm. (Sample question had indicated mean￾ing of spatial terms, such as bottom, to be two-dimensional.) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 222 HODGES, MEAGHER, NORTON, MCBAIN, AND SROUBEK

SPEAKING FROM IGNORANCE 223 Table f and Percentages per Participant try,and not enou to succeed in reading the targe for Experiments la and Ib Disagreements Experiment la Experiment Ib ipated in Experim 142 edropped for not follo ing 7% comments indicating their igne 283%) 3.7% y (cg What ar e there were slight variation 90 ve affected answers 3.13 give a word that ght was "th wer that you can give,"effee tively repeating th Results mploved for this analysis.Consistent with our hypothesis.the umber of Manipulation and suspicion checks. Only one p nt ir gave ss al We retain d he n analyses since her d )137()1000.3 nts.Exp imen 2).When Discussion asked to dese indicate Parti are that es- ne EXp nd Ib The siz hat it fa It like to b 70 riment 1b) r that it was"frustrating”(or wrong answers effect in Asch-like ituati a(e.."looked for intelligible ette then supposed what ve expectation of agreement with eing and agre ers.The firs partic tions.E n normative expectati The for the two exp ith A and Be ansy ve deserihed them e than six time nany people to see what they would predict.Invariably.they vere introduced in later studies to provide more strineent asses e statistical procedure we a vard:Sh d the nat A and as an ov ersed ut likely to be in orrect?As ted,the mos umber ofesero-inflated negative binomial regression was common solution was always to agree with the answers of b bette

although not necessarily within the word that was to be identified. In short, there was just enough information to encourage partici￾pants at C to try, and not enough to succeed in, reading the target word. Participants and replication. Twenty-seven undergraduate students at a private liberal arts college participated in Experiment 1a, receiving extra credit in a course or $5 for their participation. The same was true in Experiment 1b, although three participants were dropped for not following instructions. Ten participants in Experiment 1a asked questions or offered spontaneous comments indicating their ignorance and/or uncer￾tainty (e.g., “I can’t see. What am I supposed to do?”) during one of the early trials. Experimenters tried to answer these questions in a helpful, noncommittal way, but since there were slight variations in responding that might have affected participants’ answers, we replicated the experiment, standardizing the experimenter’s re￾sponse. Thus, in Experiment 1b, if participants in Position C made spontaneous comments about their task or asked questions, the experimenter instructed them to give a single word that they thought was “the most correct answer that you can give,” effec￾tively repeating the original instructions. Results Manipulation and suspicion checks. Only one participant in either experiment gave any evidence of suspecting that Partici￾pants A and B were confederates, and she agreed on all critical trials. We retained her in analyses since her responses did not provide support for our hypothesis. Participants believed that A and B had answered all or very nearly all of the questions correctly (M 17.6 out of 18) and that they themselves had answered many fewer questions correctly (M 7.8 for Experiment 1, M 6.3 for Experiment 2). When asked to describe their task in the experi￾ment, participants indicated that that it was to give the correct answer, with two commenting on the difficulty of the task. Participants were well aware that they “could see little or nothing” at Position C (M 16.5 out of 18 items in Experiment 2). When asked what it felt like to be unable to see words clearly, 79% (Experiment 1b) wrote that it was “frustrating” (or something similar), often indicating their dependence on A and B. Others described straightforwardly the strategy they used to solve the dilemma (e.g., “I looked for intelligible letters, then supposed what the word might be”). Disagreeing and agreeing with correct answers. The first and second columns of Table 1 show the distribution of partici￾pants’ disagreeing answers for Experiments 1a and 1b, respec￾tively. The distribution was similar for the two experiments: In both, approximately half of all participants always agreed with A and B, a quarter disagreed one to six times, and a quarter disagreed more than six times. To test for an SFI effect, we compared observed disagreements on the 12 critical items (i.e., those that A and B had answered correctly prior to C giving an answer) against an expected value of zero. In addition to this simple standard, experimental controls were introduced in later studies to provide more stringent assess￾ments. The statistical procedure we adopted to test for the presence of an SFI effect was very conservative. Because number of dis￾agreements was an overdispersed count variable with a large number of zeros, zero-inflated negative binomial regression was employed for this analysis. Consistent with our hypothesis, the number of disagreements on critical trials differed significantly from zero (in Experiment 1a: b 1.84, Wald 2 29.70, p  .001; in Experiment 1b: b 1.79, Wald 2 50.41, p  .001). Disagreeing answers were distributed relatively evenly across all 12 critical trials; there was no relation between order of trials and number of agreements, Experiment 1a: Q(11) 10.00, p .53; Experiment 1b: Q(11) 11.37, p .41. Discussion Two exploratory studies provided clear evidence for an SFI effect with 29% (94 out of 324) and 26% (75 out of 288) of responses disagreeing with the correct answers of better positioned others in Experiments 1a and 1b, respectively. The size of this effect is comparable to the average size of the agreeing-with￾wrong answers effect in Asch-like situations (about 25%), based on results of meta-analysis (Bond & Smith, 1996), although in Asch’s (1951) initial study, 34% answered incorrectly. This is true despite the fact that the normative expectation of agreement with correct answers in the SFI situation is stronger than is the expec￾tation of disagreement with incorrect answers in the Asch-type situations. Evidence for this difference in normative expectations is found in the fact that over half of participants always agreed with A and B’s answers on critical trials, while about a quarter of Asch’s participants always disagreed. Prior to running these experiments, we described them carefully to many people to see what they would predict. Invariably, they thought the experiments hardly worth running because obviously people would always agree with their more knowledgeable peers. Intuitively, it does not seem like there should be any conflict in an SFI situation, unlike the Asch dilemma. Nevertheless, results in￾dicated that participants found the situation frustrating and awk￾ward: Should they repeat what A and B have said, or should they make up a word that is theirs, one unique to their position and perspective but likely to be incorrect? As expected, the most common solution was always to agree with the answers of better Table 1 Distribution of Disagreements and Percentages per Participant for Experiments 1a and 1b Disagreements Experiment 1a Experiment 1b 0 14 (51.9%) 12 (50.0%) 1 4 (14.8%) 3 (12.5%) 2 0 1 (4.2%) 30 0 4 1 (3.7%) 0 5 0 2 (8.3%) 6 1 (3.7%) 0 70 0 8 0 1 (4.2%) 9 0 1 (4.2%) 10 1 (3.7%) 1 (4.2%) 11 2 (7.4%) 3 (12.5%) 12 4 (14.8%) 0 M 3.48 3.13 Note. Means are calculated by dividing total number of disagreements by total number of critical trials and multiplying by the number of critical trials per participant (i.e., 12). This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. SPEAKING FROM IGNORANCE 223

224 HODGES.MEAGHER.NORTON.McBAIN.AND SROUBEK Whsitioncd others.Howevercable Sl efectoccd one epB Predictions First,in addition to hypotheses t sted earlier (i.e.an SFl effect Experiment 2 was designed to begin to address these issues. sat Position 6(Not see)than Position 5(See).If this Experiment 2 h nd,there evidence of the attr rious alternative hypoth two important not in placed not only in s.social solidarity.and pragmatics in ways that indicated n a p tion of knowled hich hee con for Participant A.B. or C (real p with C ment than motivations to act on me were not informed i of self- ort mea s that as o the dem ands of speakin and cally thar ing wi AeFree On the othe hand, we did o that ther oul 、D agreers do not care about other part allowed us to test sever al altemative hypot cial so tly.we mig H se they are speaking truthfullyfro d persp n eople want to distinguish th nd a pr d tha s of pe he fact that hird possibility is that those people who lisagree with corre thers.Fifth. 1981).A fourth pe sibility is that the in these term Furthermore,we expected that Agreer ackno hen that an ons (OmeWhit 2000)A fift ndicating anger about their position would not be correlated with d be nc ant s who then act in ith th pants are ry at the Method and byn Participant ght be was omitted from analyses due to a procedural led he and I anc stud &H e of th were addre ed b yincluding tw ware of the nc however.were addressed by comparing the number of disagrecing that they were not aware of how many people were expected at

positioned others. However, a sizeable SFI effect occurred, one which cannot be explained by a lack of trust in others or a misperception of one’s own (lack of) knowledge. Participants knowingly disagreed with correct answers. Although Experiments 1a and 1b indicated that an SFI effect can occur and that it is likely not due to mistrust or inattention, they did not provide positive evidence in favor of a values-pragmatics account, nor did they consider possible alternative explanations. Experiment 2 was designed to begin to address these issues. Experiment 2 The focus of Experiment 2 was to evaluate evidence that could provide clearer support for a values-pragmatics account of the SFI effect and that could begin to rule out various alternative hypoth￾eses that might explain why some people some of the time choose to disagree with correct answers rather than conforming. The design of Experiment 2 included two important features not in￾cluded in the earlier experiments. One was the addition of a new control condition: Participants were placed not only in a position of ignorance but also in a position of knowledge in which they could see, as well as hear, the correct answer. We set up six positions for Participant A, B, or C (real participant), with C alternating between Position 5, in which the participant could see clearly, and Position 6, in which he or she could not (see Figure 1). Participants were not informed in advance that they would be placed in both positions. A second feature was the addition of a series of self-report measures that assessed various possible moti￾vations for how participants answered or other perceptions or feelings that might contribute to participants disagreeing with their peers. These two features allowed us to test several alternative hypoth￾eses for why an SFI effect occurs. One possibility is that people do not want to look like copycats, merely repeating what others have said. That is, perhaps disagreeing answers are refusals to imitate others (Wheeler & Arrowood, 1966). Another possibility is that people want to distinguish themselves from others, and answering correctly does not do that in this context (Brewer & Roccas, 2001; Griskevicius, Goldstein, Mortensen, Cialdini, & Kenrick, 2006). A third possibility is that those people who disagree with correct answers are simply more independent than those who always agree (Perrin & Spencer, 1981). A fourth possibility is that the experi￾mental situation somehow invites or demands that people give differing answers when they have been placed in an odd position and asked to answer questions (Orne & Whitehouse, 2000). A fifth possibility is that experimenters expect people to disagree and somehow subtly and nonconsciously signal this expectation to participants who then act in accordance with those signals (Hazel￾rigg, Cooper, & Strathman, 1991). A sixth possibility is that participants are angry at the experimenter for asking them impos￾sible questions and they respond aggressively by giving incorrect answers (Brehm & Brehm, 1981). A seventh possibility, which might be framed in attribution terms, is that participants think the experimenter knows they cannot see and is testing them to see if they will acknowledge it (Ross, Bierbrauer, & Hoffman, 1976). Some of these alternatives were addressed by including relevant measures in the postidentification task (e.g., how angry did par￾ticipants say they felt while in the position of ignorance?). Others, however, were addressed by comparing the number of disagreeing answers given in the position of ignorance with those given in the position of knowledge (e.g., are participants refusing to imitate?). Predictions First, in addition to hypotheses tested earlier (i.e., an SFI effect, despite trusting others’ answers, and experiencing the situation as a dilemma), we predicted that there would be significantly more disagreements at Position 6 (Not see) than Position 5 (See). If this occurs it would undercut alternative hypotheses mentioned earlier, such as refusing to imitate, experimenter expectancy, wanting to appear distinctive, and demand characteristics. Second, there should be some evidence of the attraction of truthfulness, social solidarity, and pragmatics on most if not all participants in a position of ignorance, not just those who choose to give disagreeing answers one or more times. Thus, we predicted that all participants, both those who always chose to agree with A and B (Agreers) and those who sometimes chose to disagree (Disagreers), would answer items directed at concerns for truth￾fulness, social solidarity, and pragmatics in ways that indicated their actions were constrained by values-realizing dynamics. In￾dexes of truthfulness, social solidarity, and pragmatics were ex￾pected to receive stronger endorsement than motivations to act agreeably or answer independently. Third, we expected that Disagreers would score higher on mea￾sures of truthfulness and pragmatics than Agreers. Disagreers, we believed, would likely show greater sensitivity in an SFI situation to the demands of speaking truthfully and pragmatically than Agreers. On the other hand, we did not expect that there would be a difference in expressions of social solidarity between Agreers and Disagreers. If Disagreers do not care about other participants and are simply acting independently, we might expect expressions of social solidarity to be less than for Agreers. However, if they disagree primarily because they are speaking truthfully from their assigned perspective, then there is no reason to expect that they would express significantly less social solidarity than Agreers. Fourth, we predicted that measures of agreement and indepen￾dence would be relatively weaker than other motivations, despite the fact that participants would probably often choose to agree with others. Fifth, we expected that concerns to answer correctly would be expressed by virtually all participants since the task was defined in these terms. Furthermore, we expected that Agreers would be more likely than Disagreers to acknowledge correctness as a motivation for their actions. Finally, we predicted that an item indicating anger about their position would not be correlated with disagreeing answers at Position 6 and that there would be no difference between Disagreers and Agreers in this regard. Method Participants. Forty-seven undergraduate participants at a pri￾vate liberal arts college completed the experiment for extra credit. One participant was omitted from analyses due to a procedural error, leaving 36 women and 10 men in the study. Procedure. In order to test for possible experimenter expec￾tancy effects, two experimenters unaware of the previous studies, including their rationale and results, were used, while two others were aware. Participants were recruited by e-mail solicitations so that they were not aware of how many people were expected at This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 224 HODGES, MEAGHER, NORTON, MCBAIN, AND SROUBEK

SPEAKING FROM IGNORANCE 225 mly ssigned either to Positions 1.3.and 5 (or 2.4.and 6) rticinants were debriefed with written statements and verha true partic moved to their positionsand were told the following. e seen as appro s that ed pos s that could have affecte cd.the er ple Social Solidarity.and Independence. and the items used to assess The rate wat to give your rather than rewordings of the same question.All items were rated The ex enter then presented the first nine iten tem.the word Break appeared.and the questions were.when appropriate.worded to highlight the position pa impressions of tha en (if they began in 6).This change of osition was of other varis vide pporting information.First,there were items desi ated as na index ("Sometimes during this experment was face reverse All question Participa tA to B to C.Thus,there were nine trials in which Trust index thousht the people in the other were giving hor index ("I thought the wers (ie..in pretesting.participants w imen.There was also n Anger inde when I could not see well enough to answer correctly by him-or herself not see as clearly and others could"). Index veryclmy answers.I thought it was important to give the correct answer.even if I could not see felt o did ne n giving my a others in better positions would want me to tell them my best when I coul id and said w er I felt like CPknnhe,oterposiosweemindinghcronab s and were not paving

each session. Six chairs were arranged in a semicircle as shown in Figure 1. Participants were instructed as in Experiment 1b, then randomly assigned either to Positions 1, 3, and 5 (or 2, 4, and 6), with the true participant sitting in Position 5 or 6. After seeing a sample question while seated in front of the screen, they were moved to their positions and were told the following. First, please sit straight in your chair without leaning to the left or right. Obviously, we have you in your positions for a reason. Second, as the example we just looked at illustrated, the answers you should give will always be a word. When giving your answer, please give the word that is the best answer. Please do not offer answers like “I’m not sure” or “It’s hard to tell.” Instead, please give a single word for your answer. Do the best you can to give the correct answer. Third, please wait to give your answer until I call on you. I will call on you in turn. You [pointing to A, etc.] I will call A, and I’ll call you B, and you C. The experimenter then presented the first nine items, recording answers. After the ninth item, the word Break appeared, and the experimenter instructed participants to move to a new position, one position to the right (if they began in 1, 3, 5) or one position to the left (if they began in 2, 4, 6). This change of position was unforewarned. Thus, the true participant moved either from a position in which he or she could see the correct answer to one in which he or she could not, or the reverse. All questions were answered verbally, and the order of answers always moved from Participant A to B to C. Thus, there were nine trials in which C answered last but could see, perhaps not quite as well as A and B but sufficiently well that the participant had personal visual war￾rant for his or her answers (i.e., in pretesting, participants were able to answer correctly alone from this position). On the other nine trials, the participant was in a position where he or she could not see well enough to answer correctly by him- or herself. After completing the 18 trials of the identification task, partic￾ipants answered a series of questions described below. Finally, all participants were debriefed with written statements and verbal reassurances, indicating that choosing to agree and/or disagree with others’ answers in the SFI situation could be seen as appro￾priate responses to the situation. Motivational indexes. Participants answered 6 two-item mea￾sures that indexed possible motivations that could have affected agreeing or disagreeing during the identification task. The six indexes were Agreement, Correctness, Truthfulness, Pragmatics, Social Solidarity, and Independence, and the items used to assess them are given in Table 2. The two items used for each index were intended to be separate samplings of the motivational domain rather than rewordings of the same question. All items were rated on scales from 1 (Disagree) to 7 (Agree). Although participants answered these questions after having been in both positions, questions were, when appropriate, worded to highlight the position of ignorance and participants’ impressions of that situation. Other indexes. Other items were constructed to serve as in￾dexes of other variables relevant to predictions, or to provide supporting information. First, there were items designated as a Dilemma index (“Sometimes during this experiment I was faced with a dilemma between saying what I knew [from what others had said] and guessing incorrectly since I could not see clearly”), a Trust index (“I thought the people in the other positions were giving honest answers”), and a Suspicion index (“I thought the people in the other positions were actually working with the experimenter and that I was the only real participant in the exper￾iment”). There was also an Anger index (“I felt angry when I could not see as clearly and others could”). Table 2 Motivational Indexes Across Positions of Ignorance and Knowledge Index Agreement When giving my answers, I thought it was important to agree with others as much as possible, regardless of what I could or could not see. Regardless of whether I could see the material on the screen clearly or not, I thought it best to give the answer others had given. Correctness When giving my answers, I thought it was important to give the correct answer, even if I could not see very clearly. I was most concerned to give the correct answer, even if it meant I had to depend on others to provide me with the necessary information. Truthfulness I felt obligated both to give the correct answer and to indicate truthfully what I could or could not see. It felt like it was “cheating” to give the answer that others had given when I could not see clearly. Pragmatics I did not feel like I was contributing much to the experiment if I just repeated what the others had said. In giving my answers, I tried to make a distinctive contribution to “the conversation” rather than just doing what everyone else was doing. Social Solidarity I thought that others in better positions would want me to tell them my best guess, even when I could not see very clearly. Even though I did not know the other persons in this experiment very well, I felt like I should pay attention to their answers and learn from them. Independence I mostly ignored what others said and said whatever I felt like saying. I thought the people in the other positions were minding their own business and were not paying particular attention to my actions. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. SPEAKING FROM IGNORANCE 225

226 HODGES.MEAGHER.NORTON.McBAIN.AND SROUBEK agre 022.p make the Finally the two ite sible he ugh bout whether I had the not to conform to the eers in fac 5.76)felt like they were in ht th ma greers (M 43.9 tha rather tha g when I wa a bad po on to see cle 2.91)a 44)=0.06. 96.but therew a significant differe e of ces about how often to disagree or urther eviden of participants'trust in the integrity of the situ Results and Discussion rectly by A and B(17.8 out 1).By contrast.their mea The distr ng a ()( nno ce) 42=634.p<.01 The mot on position 6 order of s(5 first vs.6 first),and .nallwo-way 5740r 5.14) way n ontrast.indexes of Ind M 331 2.67)e s at Position 6 (M of 9 pe and Dis self- hental hypotheses did not have an effect on disagr f va e(MANOVA)o Position 6.30.2%were incorrect,disagn g answers.which i igure 3.The figure presents dexes c 1.76 nd Pr he on highlighted by valucs-pragmatics theory.while the scond Distribution of Disagreements and Percentages per Participan e at Differing Positions in Experiment 2 for these ns is as follow Disagreement Position 5(54.3%) cople are a that they (truthfully)cannot personally see the the demands of truthfulnessare ted to be more likely to for motivated to c about others would lead one to ty. ever tha 2.72 0.35 Note. in which would tend to yield How

Two items indexed beliefs about experimenter expectations: (a) “I thought the experimenter wanted me to agree with the other people in the group, especially if they were in a better position than me,” and (b) “I thought the experimenter wanted me to make the best guess I could about what I saw, even if it was incorrect.” Finally, there were two items posing possible hypotheses about the experiment, one related to courage (“I thought that, although the experiment appeared to be about visual perception, it was really about whether I had the courage not to conform to the answers of the others”) and the other related to humility (“I thought that the experiment was less about what I could see and more about whether I had the humility to depend on others, rather than trying to “do my own thing” when I was in a bad position to see clearly”). We had no explicit expectations about how much agreement these four items might attract, but we thought none would bear a close relation with participants’ choices about how often to disagree or agree with A and B’s answers. Results and Discussion Trust, dilemma, and going beyond correctness. The distri￾bution of agreeing answers for Positions 5 (can see) and 6 (cannot see) is shown in Table 3. To assess the influence of the experi￾mental manipulations, frequency of disagreements was regressed on position (5 vs. 6), order of positions (5 first vs. 6 first), and experimenter (informed vs. uninformed) using a mixed-effects, zero-inflated negative binomial regression, including all two-way and three-way interactions. The only significant effect was a main effect for position (b 1.04, Wald 2 34.69, p  .001): Disagreements at Position 6 (M 2.72 of 9 possible) were significantly greater than at Position 5 (M 0.35 of 9 possible). The lack of any experimenter effects indicates that knowledge of experimental hypotheses did not have an effect on disagreeing answers given at Position 6. There were no order effects either, indicating that the SFI effect was not moderated by the first position in which participants sat. As predicted, a sizeable and significant SFI effect occurred at Position 6: Of answers given at Position 6, 30.2% were incorrect, disagreeing answers, which is significantly different from the expected value of zero (b 1.76, Wald 2 158.76, p  .001). In contrast, disagreeing answers at Position 5 (M 3.86%) did not significantly differ from zero (b 0.22, p .60). Twenty-five participants (54.3%) always agreed (Agreers) with A and B while sitting at Position 6, while 21 participants (45.7%) disagreed sometimes (Disagreers), making up their own incorrect answers. As predicted, participants strongly endorsed the Dilemma index. It is important to note that this was not limited to Dis￾agreers; in fact, Agreers (M 5.76) felt like they were in a dilemma as much as Disagreers (M 5.43), t(44) 0.67, p .50. Analyses of self-report measures revealed that participants showed high levels of Trust (M 6.67 out of 7), and low levels of Suspicion (M 2.91) about the other participants and the task. There was no difference between Agreers and Disagreers on Trust, t(44) 0.06, p .96, but there was a significant difference on Suspicion, with Disagreers (M 2.14) actually expressing less suspicion than Agreers (M 3.56), t(44) 2.29, p .027. Further evidence of participants’ trust in the integrity of the situ￾ation is provided by their estimates of questions answered cor￾rectly by A and B (M 17.8 out 18). By contrast, their mean estimate of questions they answered correctly at Position 6 was 4.9 (out of 9), much lower than their estimate for Position 5 (M 8.6), t(42) 6.34, p  .001. Truthfulness, social solidarity, and pragmatics. The moti￾vational indexes with which participants most agreed were, in order, Truthfulness (M 5.74 out of 7), Correctness (M 5.14), Social Solidarity (M 5.08), and Pragmatics (M 4.25). By contrast, indexes of Independence (M 3.31) and Agreement (M 2.67) received ratings indicating a lack of agreement, if one assumes that the midpoint of the scale, 4, is neutral. In order to compare Agreers and Disagreers’ self-reports regarding these mo￾tivational indexes, we conducted a 2 (level of agreement: Agreers vs. Disagreers)
2 (order)
2 (experimenter) multivariate anal￾ysis of variance (MANOVA) on all indexes. No significant order or experimenter effects were found for any of the indexes. Thus, we focused on direct comparisons of three sets of means presented in Figure 3. The figure presents comparisons between indexes of Truthfulness and Correctness, Social Solidarity and Agreement, and Pragmatics and Independence. The first index in each pair is the one highlighted by values-pragmatics theory, while the second in each pair addresses concerns traditionally considered as moti￾vating conformity (Agreement, Correctness) or nonconformity (In￾dependence). The reason for these specific comparisons is as follows: First, Truthfulness is contrasted with Correctness because, ordinarily, being motivated to be truthful implies wanting to give the correct (i.e., true) answer. However, in the context of the SFI situation, people are aware that they (truthfully) cannot personally see the correct answer. Thus, participants who are particularly sensitive to the demands of truthfulness are expected to be more likely to forgo correctness in order to express the truth of their ignorance. Second, Social Solidarity is compared with Agreement because, ordinarily, being motivated to care about others would lead one to expect agreement. Social solidarity, however, involves more than simply agreeing with others and instead entails playing one’s role to do what is most helpful for the group. Third, Pragmatics is juxtaposed with Independence because, in the SFI situation, Pragmatics tends to encourage making a unique, appropriate contribution to the task, which would tend to yield occasional disagreeing answers. How￾Table 3 Distribution of Disagreements and Percentages per Participant at Differing Positions in Experiment 2 Disagreements Position 6 Position 5 0 25 (54.3%) 35 (76.1%) 1 3 (6.5%) 7 (15.2%) 2 2 (4.3%) 3 (6.5%) 3 2 (4.3%) 1 (2.2%) 4 0 5 1 (2.2%) 6 1 (2.2%) 7 2 (4.3%) 8 3 (6.5%) 9 7 (15.2%) M 2.72 0.35 Note. Means are calculated by dividing total number of disagreements by total number of trials and multiplying by the number of trials per partici￾pant (i.e., nine). This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 226 HODGES, MEAGHER, NORTON, MCBAIN, AND SROUBEK

SPEAKING FROM IGNORANCE 227 Finally,for the Pragmatics versus Independence com found for mt p ain found fo ipant behavio and rar Fl,4= atics index 24) 69. atic than agre s an ors of motivational indexes for Ag r.the disagreement would be based on paying close attention to tion nstrained by truth ness and com both fo are equally i s are sensib social solidarity and pragmatics have e pnonty wit nd he impo iven that the Correctness)and the participant behavior (Agreers vs.Dis xplicitl tness in ansv which can only b 015 There wa o participan choices.which n a large number of partic D).A significan was found for Possible alternative While the vior.F(1.44)- port for atings for Di .20 458, 001 and reasons for ereingan wers bein offered at Positi of P f the cers Di 月d he the comp al So accounts.such accounts that 3193.p different e .00s between motivation and participant beh gree in order to differ ntiate themselves from others and/ r t then or th Social <001 and Agreers (24)=304=006.Thus even thoush (M=3.81)on the item"Iwanted to give the correct answer each Agreers always chose to agree.they indicated the priority of Socia ime,but I a in their Social Solidar en disa vers at Positions 5 and 6 is ugh Disagreers expressec or hypot that people do not like repeating other peopl

ever, the disagreement would be based on paying close attention to others and the conversation, rather than ignoring others and their contributions. Finally, it should be noted that while these compar￾isons are sensible, there are other plausible comparisons that could be made, given the interrelatedness of truthfulness, pragmatics, and social solidarity. For each pair of motivations we compared, we first tested for a significant interaction between the motivation (e.g., Truthfulness vs. Correctness) and the participant behavior (Agreers vs. Dis￾agreers). For Truthfulness versus Correctness, a significant main effect was found for motivation type, F(1, 44) 6.37, p .015, with higher ratings for the Truthfulness index. There was no main effect for participant behavior (Agreers vs. Disagreers) on these items (F  1). A significant interaction was found for motivation and participant behavior, F(1, 44) 18.10, p  .001. As shown in Figure 3, Truthfulness ratings were significantly greater than Cor￾rectness ratings for Disagreers, t(20) 4.58, p  .001, and there was no difference between the two indexes for Agreers, t(24) 1.28, p .21. Disagreers expressed significantly more concern for truthfulness than Agreers, t(44) 2.24, p .030, while Agreers indicated greater concern for correctness than did Disagreers, t(44) 3.00, p .004. Thus, the pattern suggests that participants were concerned for both truthfulness and correctness but that truthful￾ness took the lead for Disagreers, while correctness and truthful￾ness were equally important for Agreers. For the comparison between Social Solidarity and Agreement, a significant main effect was found for motivation type, F(1, 44) 31.93, p  .001, with higher ratings for the Solidarity index. A significant main effect was also found for participant behavior (Agreers vs. Disagreers) on these items, F(1, 44) 7.65, p .008. The interaction between motivation and participant behavior (Agreers vs. Disagreers) was not statistically significant, F(1, 44) 2.45, p .125. As shown in the middle panel of Figure 3, ratings for the Social Solidarity index were significantly greater than ratings for the Agreement index for both Disagreers, t(20) 4.86, p  .001, and Agreers, t(24) 3.04, p .006. Thus, even though Agreers always chose to agree, they indicated the priority of Social Solidarity as a motivating concern. There was not a significant difference between Agreers and Disagreers in their Social Solidar￾ity ratings, t(44) 1.72, p .093, although Disagreers expressed significantly less agreement with the Agreement index than did Agreers, t(44) 2.80, p .008. Finally, for the Pragmatics versus Independence comparison, a significant main effect was found for motivation type, F(1, 44) 13.87, p .001, with higher ratings for the Pragmatics index. A significant main effect was also found for participant behavior (Agreers vs. Disagreers) on these items, F(1, 44) 11.56, p .001. The interaction between motivation and participant behavior (Agreers vs. Disagreers) was not statistically significant, F(1, 44) 0.55, p .463. As shown in the right panel of Figure 3, both Agreers and Disagreers’ ratings on the Pragmatics index were significantly greater than ratings on the Independence index, t(24) 2.69, p .013, and t(20) 2.57, p .018, respectively. Disagreers were significantly more motivated by pragmatic concerns than Agreers, t(44) 2.24, p .030, while the difference between Disagreers and Agreers in terms of expressing independence did not reach statistical significance, t(44) 1.85, p .071. Overall, the pattern of results regarding motivational indexes shown in Figure 3 is remarkably clear. First, choices in the SFI situation are constrained by truthfulness and correctness, both for Agreers and Disagreers, but truthfulness takes the lead over cor￾rectness for Disagreers, while for Agreers they are equally impor￾tant. Second, social solidarity and pragmatics have priority with respect to agreement and independence, respectively, as motiva￾tions for actions in the SFI situation, particularly for Disagreers. Third, the importance of truthfulness, social solidarity, and prag￾matics is particularly impressive given that the situation is one that explicitly calls for correctness in answers, which can only be achieved by agreement. This suggests there was an important role for values-pragmatics considerations in the dynamics constraining participants’ choices, which resulted in a large number of partic￾ipants sometimes giving disagreeing answers. Possible alternative explanations for SFI effect. While the results described thus far have provided encouraging support for a values-pragmatics account of SFI effects, it is possible that there are other reasons for disagreeing answers being offered at Position 6. The comparison of Positions 5 and 6, described earlier, provided a way to test the viability of some of these alternatives, while others were addressed by self-report measures. As noted earlier, all indexes were initially submitted to a MANOVA in which the only significant effects were due to differences between Agreers and Disagreers. Demand characteristics, imitation, and uniqueness. While the strong SFI effect supports values-pragmatics theory, it poses problems for alternative accounts, such as accounts that claim disagreeing answers are to be expected when people are put in different positions in a perception experiment. If this were correct, we would expect as many disagreeing answers in Position 5 as in Position 6. A similar argument holds for the hypothesis that people disagree in order to differentiate themselves from others and/or to express their uniqueness: If disagreement is motivated by a differ￾entiation motive, then there should be no difference between Positions 5 and 6. This finding is reinforced by the fact that there was no difference between Agreers (M 4.32) and Disagreers (M 3.81) on the item “I wanted to give the correct answer each time, but I also wanted to make a unique contribution to the experiment,” t(44) 0.92, p .36. Finally, the large difference between disagreeing answers at Positions 5 and 6 is problematic for the hypothesis that people do not like repeating other people Figure 3. Means and standard errors of motivational indexes for Agreers and Disagreers for Experiment 2. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. SPEAKING FROM IGNORANCE 227