My teacher is wrong: Preschoolers’ opposition to non-conventional statements

Early Childhood Research Quarterly 39 (2017) 1–13

Contents lists available at ScienceDirect

Early Childhood Research Quarterly

My teacher is wrong: Preschoolers’ opposition to non-conventional statements夽 Silvia Guerrero a,∗ , Cristina Cascado a , Melisa Sausa a , Ileana Enesco b a b

Universidad de Castilla-La Mancha, Facultad de Educación de Toledo, Campus Fábrica de Armas, Avda. Carlos III, s/n, 45071, Toledo, Spain Universidad Complutense de Madrid, Facultad de Psicología, Campus de Somosaguas, 28223, Pozuelo de Alarcón, Madrid, Spain

a r t i c l e

i n f o

Article history: Received 3 July 2015 Received in revised form 10 November 2016 Accepted 15 November 2016 Keywords: Epistemic authority Conventional norms Majority pressure Testimony Teacher Preschoolers

a b s t r a c t We examined the influence of teachers’ conflicting opinions on preschoolers’ decisions in two contexts: (a) alternative, non-conventional uses of common objects (e.g., using a fork to comb hair), and (b) labeling new objects. In the first context, teachers’ conflicting claims involved the acceptability of the alternative use of objects, either rejecting them (conventional view) or accepting them (non-conventional view). In Study 1 (N = 36, 3 and 5 year-olds), the aim was to evaluate the pressure of a familiar epistemic authority: Children were presented with the opposing claims of their own teacher and a stranger. In Study 2 (N = 91, 5-year-olds), the aim was to evaluate the pressure of a majority of teachers and the role of dissenters. Children were assigned to two conditions: in the dissenter condition (DC), they faced the conflicting opinions of three teachers vs. one teacher, and in the non-dissenter condition (NDC), they faced the unanimous opinion of three teachers. The general results showed that the responses of the 3- year-old children (study 1) were not influenced neither by their teacher nor by the context. By contrast, the 5year-olds strongly opposed the teachers when they supported a non-conventional way to use common objects, regardless of the informant’s familiarity with the informant (their own teacher, study 1), and the degree of consensus among teachers (unanimous or partial majority, study 2). Most children only sided with the informant/s whose claims were conventional (disapproving the alternative use of objects). In the labeling context, 5-year-olds’ decisions were influenced more by the unanimous majority of teachers (NDC) than by the partial majority (DC), or by their own teacher. Overall, the findings show that children’s previous beliefs have more strength than their compliance with the authority represented by teachers. Moreover, this work provides evidence of preschoolers’ resistance – or indifference- to majority pressure. © 2016 Elsevier Inc. All rights reserved.

1. Introduction During childhood, essential knowledge of the physical and social world is acquired. A part of this knowledge is learned from other people’s testimony rather than direct observation or experience (for a review of the literature, see Harris, 2012; Robinson & Einav, 2014; Rotenberg, 2010). Children can learn through their own experience that water is wet but not that the Earth is a sphere. In the first case, testimony on the properties of water is not essential to acquire this knowledge, but it is in the second case. In other types of learning, such as vocabulary acquisition or the use of utensils (e.g., cutlery), children can learn new words or the function of a new object, not only from verbal testimony but also

夽 The study presented in this article was supported by the Ministerio de Economía y Competitividad of Spain (PSI2012-31477), awarded to Ileana Enesco. ∗ Corresponding author. E-mail address: [email protected] (S. Guerrero). http://dx.doi.org/10.1016/j.ecresq.2016.11.001 0885-2006/© 2016 Elsevier Inc. All rights reserved.

through behavioral cues which complete the information (e.g., an adult names the object while pointing at it or showing how it is used). These cues tend to be regularly repeated by other people. In all these learning processes, a key factor in accepting the information is trust in the testimonial source and in particular, how much knowledge or epistemic authority is attributed to the source in the given context. Some authors have specifically highlighted the importance of this dimension (Mills, 2013). Adults do not tend to invest unlimited trust in a single source or person. We trust an expert in economics for matters regarding the economy but not for questions of linguistics or building materials. Is the same true for children? A number of recent studies show that, from 5 years, children trust more the information provided by an expert in a particular subject or someone who has privileged access to relevant information, than a non-expert or a person with no access to relevant information (Einav, 2014; Lane & Harris, 2014). However, there are few studies evaluating children’s trust in the testimony of the quintessential epistemic authority: the teacher.

2

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

Preschool teachers are undoubtedly an emotional reference in early childhood and can become an attachment figure for children (Van Ijzendoorn, Sagi, & Lanbernon, 1992). In addition to this close relationship, young children also take for granted that their teacher knows all that needs to be known, and furthermore, transmits this knowledge in class (Olson & Bruner, 1996). Children trust their teacher for various reasons: The teacher is a figure of authority, has academic qualifications, and is treated as an expert by other teachers and by parents, too (Siegel, 2005). In the same vein, recent studies identify benevolence and competence as the two principal dimensions an informant must possess to be trustworthy (Johnston, Mills, & Landrum, 2015; Koenig & Stephens, 2014): The figure of the teacher embodies both dimensions and is thus a testimonial source who can, and in most cases should, be trusted. However, what happens when the information supplied by a teacher conflicts with children’s prior knowledge? Do children defer to the epistemic authority of a teacher or, by contrast, trust their own judgement? This is one of the questions addressed in this work. Chan (2011) and Chan and Tardif (2013) are one of the few authors to study the importance of the teacher’s opinion in children’s decisions. In these studies, kindergartners and second graders in the United States and Hong Kong were asked to categorize a series of ambiguous objects (e.g., a button which could also be a wheel) or non-ambiguous objects (e.g., a button that was clearly a button) after a fictitious teacher had provided information that conflicted with the opinions previously expressed by the children. Results showed that participants generally endorsed more of the teacher’s labels for ambiguous objects than for non-ambiguous objects, although there were some differences across ages and cultures. When the object was non-ambiguous, kindergartners in the United States endorsed more conflicting labels – giving up their prior beliefs and siding with a teacher – than their peers in Hong Kong. The cross-cultural differences disappeared by second grade. These results suggest that factors such as prior knowledge, age and culture influence the trust children invest in a teacher. In another study Chan (2011, Study 3) asked children aged from five to eight to assess how right a teacher’s behavior in a story was when she did not behave as expected, but instead breached accepted moral and socio-conventional rules. The results were clear: In 99% of the tests, the children judged the teacher’s behavior to be wrong. Other authors have studied the teacher’s influence on the learning of new words or new functions for novel objects (Corriveau & Harris, 2009). Their findings show that from 3 years, children trust a familiar teacher from their school more than an unfamiliar one. However, this level of trust decreases from 4 years, and more clearly from 5 years, when a familiar teacher labels something inaccurately (e.g., labeling a duck a spoon). These results show that other dimensions such as past accuracy may have more influence than a familiar figure of authority, at least when children have no prior knowledge to assist their decision. In all these previous studies, the teacher was either a stranger artificially given the status of teacher, or was a teacher familiar to the children, but not their own teacher. Some authors suggest that the setting in which studies analyzing trust in testimony are conducted may influence the results. They specifically state that children might be especially receptive to their teachers’ claims in the school context (Lane & Harris, 2014, p. 6). Following this suggestion, in our first study, one of the informants is the participants’ own teacher. The aim of Study 1 is to assess children’s trust in their own teacher versus a stranger in two conventional domains of knowledge: (a) Alternative use of a common object: Participants had to decide if the non-conventional use of different familiar objects was right (e.g., a fork used to comb hair); (b) labeling new objects: Taking

labeling tasks used in previous studies (Corriveau, Fusaro, & Harris, 2009) as a reference, children had to relate unfamiliar objects with novel labels. In both contexts, children were confronted on film by contradictory information supplied by each informant (their own teacher vs. an unknown woman). Only after seeing the two informants and listening to the different judgements, were children asked their own. In the labeling trials, we supposed that children would trust the information supplied by their own teacher more than that of the stranger. This prediction took into account Koenig and Stephens’ proposal regarding the key dimensions of informants, as well as previous studies on this question with adult informants. Predictions for the alternative use of a common object are more complicated given the nature of the situations. Children observe totally plausible non-conventional uses of familiar objects (e.g., a beach bucket used as a salad bowl) and listen to the opposing opinions of two informants (it is ok to use for vs. it isn’t ok to use for ). We ruled out absurd situations in which physically impossible functions are defended (e.g., claiming that an object with no containment capacity is for drinking), since previous studies (Schillaci & Kelemen, 2014) have shown that from at least 4 years children actively reject absurd suggestions even when defended by an adult majority. They similarly lose trust in an adult seen to make mistakes in simple tasks (e.g., wrongly labeling common objects). Consequently, a likely hypothesis in this setting is that children will be more disposed to accept the non-conventional – but plausible- use of an object when defended by their teacher rather than by a stranger. Alternatively, if we assume that preschoolers are critical of the breaching of rules, be they socio-conventional rules like ways of eating, dressing, greeting, etc., (Nucci, 1981; Smetana, 1981), or game rules (Rakoczy, Warneken, & Tomasello, 2008), it cannot be ruled out that they follow their own normative understanding more than the teacher’s authority. Few studies analyze developmental differences when using teachers as informants (Chan, 2011; Corriveau & Harris, 2009) and their differing conclusions prevent us making robust predictions. Broadly speaking, we predict that as participant age, and consequently, their knowledge of the world increases, trust in the teacher will decrease in situations where their prior knowledge contradicts the teacher’s testimony. Some authors explain this tendency by allusion to developmental changes in the heuristics children use to assess testimony (see Harris et al., 2012). Thus, children’s increasing knowledge and greater capacity to weigh the information and the informant, allow us to predict a reduced trust in the teacher. The results of the first study left some questions open. This motivated the design of the next study in which we continued to explore the influence of teachers’ conflicting opinions on preschoolers’ decisions. We used the same paradigm of conflicting claims from Study 1, but included a relevant dimension in the research about trust in testimony: Consensus in the information provided by different people. Some recent studies have assessed children’s preference for consensus testimony when presented in the form of a majority of adults who voice an opinion in opposition to a lone dissenter (Chen, Corriveau, & Harris, 2011; Chen, Corriveau, & Harris, 2013; Corriveau, Fusaro et al., 2009; Schillaci & Kelemen, 2014). The results of these studies have shown that from the age of 4, children are sensitive to the consensus opinion in the acquisition of new knowledge as, for example, when deciding the name or the functions of new objects. In these studies, the consensus opinion is provided by unfamiliar adults. No information is given about who they are or what they know. We adapted this paradigm to our purposes. Accordingly, the informants in Study 2 were presented as teachers from another school, so as to reinforce the idea they had expert knowledge. An additional goal of this second study was to explore the sensitivity of preschoolers to the presence or absence of a dissenter teacher faced with a majority. To the best of our

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

knowledge, studies assessing the weight of the majority in children have worked either with unanimous majorities (Corriveau & Harris, 2010; Haun & Tomasello, 2011), or non-unanimous majorities (Corriveau, Fusaro et al., 2009). However, there are no previous studies analyzing the degree to which this factor is relevant for children. 2. Study 1 The objective of Study 1 was to assess children’s trust in the information provided by their own teacher versus a stranger in two conventional domains of knowledge: (a) Alternative use of a common object (a fork, pajamas, a bucket, a bottle), and (b) labeling new objects. In order to confirm that the situations presented in the context of an alternative use of a common object were wholly familiar to the preschoolers, fourteen 3-year-old children (M = 43 months, SD = 3 months, range: 38–49, 7 males) participated in a pilot study. Children were interviewed to evaluate their knowledge of the common function of the objects to be used in this research and to be sure that children as young as 3 had prior knowledge about the conventional use of these objects. First, children saw four real objects in succession: a fork, pajamas, a bucket and a bottle. They were asked the following questions about each object: 1) “Do you know what it is? Do you know what it’s for?” 2) “Could you use the fork/pajamas/bucket/bottle for something different/in another way?” 3) “Would it be ok to use the fork/pajamas/bucket/bottle for combing your hair/going to school/serving salad/drinking soup?” and “Why?” The results of the pilot study showed that all participants knew the objects and their conventional use, and only three of them (21.4%) also accepted a non-conventional use of some objects (question 3): Two children accepted the use of a bottle to drink soup, and one child the use of a bucket to put salad in. Thus, virtually all the children rejected non-conventional uses and their judgements conformed to socio-conventional norms. 2.1. Method 2.1.1. Participants Thirty-six preschool children participated in this study (M = 57 months, SD = 13 months, range: 38–75, 17 male). The sample was composed of two age groups: Younger, 18 3-year-olds (M = 44 months, SD = 4 months, range: 38–49, 8 male) and older, 18 5-year-olds (M = 68 months, SD = 3 months, range: 66–75, 10 male). All children were recruited from the same school in Mocejón (Toledo, Spain), serving middle socioeconomic status (SES) families. Most (83.3%) belonged to the majority ethnic group with a minority of other groups (11.1% Moroccan, 2.7% Rumanian, and 2.7% Peruvian). All children were fluent in Spanish. Children were individually interviewed by a female researcher. Written parental consent, as well as children’s verbal assent, was obtained for all participants. All the children in the classes taught by a specific teacher, Angelines, took part in the study. 2.1.2. Procedure All participants watched videos of their own teacher, Angelines (T.), and of a stranger, Nieves (S.), who made conflicting claims in two different contexts: (a) Judging an alternative use of common objects, and (b) novel object labeling. T. was the teacher for all of the participants because the data were collected in two successive academic years: In 2013, T. was the teacher in the 5-year-old’s class and the following year, 2014, she was the 3-year-olds’ teacher. T. was introduced as their teacher, and all the children recognized her. The stranger was presented as a woman unknown to the children. Both informants were adult females from the same racial group (Caucasian), with neutral facial expressions, and dressed differently: T.

3

was wearing the clothes she usually wore on a school day and the stranger wore a dark T-shirt and blue jeans. The interviewer began by saying: “Let’s watch some videos of your teacher, Angelines (T.), and another woman you don’t know. Her name is Nieves (S.). Both are going to give us their opinions about different things. Pay attention to what they say because I would like to know what you think.” The stimuli were presented on a laptop and the interviewer was the only person responsible for operating the presentation. See the Supplementary materials for details. 2.1.3. Contexts 2.1.3.1. Alternative use of common objects. The situations were presented on a computer screen and the stimuli appeared in drawings and/or videos. Participants completed four trials. The procedure was always the same. Four different stories were presented: (a) Soup-Bottle: Children saw three vignettes and listened to a story: “Well, now we are going to see some pictures. Look, this is Héctor. He is going to have some soup. Héctor wants to drink the soup from a bottle but not from the plate. Héctor has put the soup in the bottle and is drinking the soup from the bottle”; (b) Salad-Bucket: Children watched a video of a woman putting salad in a bucket, using the bucket as a salad bowl. During the video the interviewer said, “Look, this woman is putting the salad in a bucket”; (c) Fork-Comb: Children watched a video of a woman using a fork to comb her hair. During the video the interviewer said, “Look, this woman is using a fork to comb her hair”; (d) Pajamas-School: Children were shown three vignettes and listened to a story: “Well, now we are going to see some pictures. Look, this is Samuel. He has to get dressed to go to school. He’s thinking about what he’s going to wear today. Samuel has decided that he’s going to wear pajamas. Samuel is at school in pajamas”. After each story, children were told “Let’s see what T. and S. say about this”. Then they watched a video where T. was non” and conventional. In these situations, T. said, “Yes, it is ok to S. said, “No, it isn’t ok to ”, for soup-bottle and salad-bucket stories; or the T. was conventional. In these situations T. said, “NO, it isn’t ok to ” and S. said, “Yes, it is ok to ”, for fork-comb and pajamas-school stories. Before asking the child’s opinion, the interviewer always reminded them what the informants had said in each story. “Ok. T., has said .and S. has said . Who do you think is right? And ‘Why do you think she is right?’ 2.1.3.2. Labeling new objects. The presentation was shown on a computer screen. Participants completed two trials. The procedure was always the same and analogous to alternative use of common objects tasks. In the first trial, they saw a photo of one novel object in the middle of the computer screen and the interviewer said, “Look, this is a very strange thing. I don’t know what it is, do you? Let’s see what T. says, and what S. says, but wait until they both speak before you say what you think”. The children then watched a video in which T. said, “It is a Reso” and S. said, “No, it is a Firto”. After that, and before asking the child’s opinion (as in the alternative use of common objects task), the interviewer always reminded them what the informants had said in each story: “Ok. Your teacher has said this is a Reso and Nieves has said it’s a Firto. Who do you think is right?” The second trial was similar but a new object and new labels were presented. 2.1.4. Order of presentation (see Table 1) The order of trials was maintained across participants to preserve clarity: 1) First, novel object labeling (Reso-Firto), 2) Soupbottle (T. is non-conventional), 3) Fork-comb (T. is conventional), 4)

4

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

Table 1 Order of trials across studies. Order 1

Order 2

The teacher said (Study 1)/The three teachers said (Study 2): 1◦ ) It is a RESO (Labeling) 2◦ ) It is OK to use a Bottle to drink soup (Non-conventional) 3◦ ) It isn’t OK to use a fork to comb your hair (Conventional) 4◦ ) It is a FERPO (Labeling) ◦ 5 ) It isn’t OK to wear pajamas to go to school (Conventional) 6◦ ) It is OK to use a bucket to put salad in (Non-conventional) And the stranger (Study 1)/The dissenter (Study 2) always said the other choice for each trial in the DC. No other choice was offered in the NDC.

The three teachers said (Study 2): 1◦ ) It is a FERPO (Labeling) 2◦ ) It isn’t OK to wear pajamas to go to school (Conventional) 3◦ ) It is OK to use a bucket to put salad in (Non-conventional) 4◦ ) It is a RESO (Labeling) 5◦ ) It is OK to use a bottle to drink soup (Non-conventional) 6◦ ) It isn’t OK to use a fork to comb your hair (Conventional) And the dissenter (Study 2) always said the other choice for each trial in the DC. No other choice was offered in the NDC.

Second, novel object labeling (Ferpo-Linso), 5) School-pajamas (T is conventional), and 6) Bucket-salad bowl (T. is non-conventional). 2.1.5. Coding Given that in the context of alternative use of a common object T. could adopt a conventional view or a non-conventional view, three possible situations were analyzed: (a) T. is not conventional (she agrees with the transgression), (b) T. is conventional (she disagrees with the transgression), and (c) novel object labeling. Participants’ dichotomous responses across the two trials within the same situation were compared (McNemar tests) and no significant differences were found between agreement with T. in trial one and trial two. Thus, for all analyses, participants’ dichotomous responses for each trial were scored 1 point, when children chose their teacher’s option, and 0 points, when children chose the stranger’s option. Scores for the three situations were also combined to yield a total score for each participant ranging from 0 to 6. 2.2. Results and discussion Results across studies for questions of justification will be presented later. All analyses on gender differences yielded nonsignificant results in their trust in T. Therefore, subsequent analyses were run on data collapsed across that factor. Given the categorical nature of the data in the two studies for the statistical analyses, we used non- parametric tests, providing the nonparametric effect sizes (the phi coefficient) when required (American Psychological Association, 2011, p. 34). 2.2.1. Trust in the own teacher Considering the total score for each subject in the three situations (six trials, range 0–6), children were significantly more willing to endorse the T. view than the S. view, ␹2 (4, N = 36) = 23.17, p < .001. In order to analyze whether this tendency depended on the situation, a Friedman test was conducted with the whole sample. It was statistically significant, ␹2 (2, N = 36) = 14.92, p < .001. Pair-wise comparisons using Wilcoxon tests with Bonferroni correction showed that children were significantly more likely to side with their T. in the labeling than in the non-conventional situation (p < .001) and also more likely in the conventional than in the nonconventional situation (p < .005); there was no difference between conventional and labeling situations (p = .313), (see Fig. 1, black columns). 2.2.2. Differences by age in trust in the own teacher Mann-Whitney tests were conducted (see Fig. 2) to find whether these results varied by age. Broadly speaking, throughout the test,

the younger children sided more with the teacher than the older children. This difference was marginally significant, U = 109.500, n = 36, p = .076. Regarding each situation, differences between 3 and 5 year olds were found in the alternative use of common object: Younger children sided with their T. more than the older ones in the non-conventional situation, U = 55.50, n = 36, p < .001, while older children sided with their T. more than the younger ones in the conventional situations, U = 81.00, n = 36, p < .05. No differences by age were found in the labeling situation, U = 149.00, n = 36, p = .67. In order to see the degree to which the tendency to side with T. in each group depended on the situation, a Friedman test was conducted for each age group. Thus, while the 3-year-old group was not influenced by the situation ␹2 (2, N = 36) = .448, p = .80, the older children were, ␹2 (2, N = 36) = 30.14, p < .001. Younger children accepted their T. at similar rates across the three situations. However, pair-wise comparisons using Wilcoxon tests with Bonferroni correction show that older children were significantly more likely to side with their T. in the conventional situation than in the non-conventional situation (p < .001) and in the labeling situation (p < .005). Furthermore, they sided with T. more in the labeling situation than in the non-conventional situation (p < .001). Finally, binomial tests were conducted for each trial in each situation to check whether children responses were different from chance across age group (see Fig. 3, columns for Study 1). In the alternative use of common objects, the 3-year-olds responded at chance in all the trials (conventional: n = 9, p = 1.00, first trial, n = 12, p = .24, second trial; non-conventional: n = 10, p = .82, first trial, n = 13, p = .10, second trial). By contrast, all the 5-year-olds sided with their teacher in conventional situations and all (except one) opposed her in non-conventional situations (ps < .001). In the labeling situation, both groups responded once at chance, and once above chance (3-yr-old, n = 11, p = .48, first trial, n = 15, p < .005, second trial; 5-yr-old, n = 15, p < .01, first trial, n = 9, p = 1.00, second trial).

2.2.3. Degree of agreement with their own teacher In an additional and complementary analysis, participants’ dichotomous responses to the two trials within the same situation were scored on a 2-point scale: Score = 0 when participants never agreed with the T.; score = 1 when participants agreed once with T. and once with S.; and score = 2 when participants agreed with T. twice. Table 2 shows the percentages of children in these three categories by age and situation. Regarding the labeling context, in both age groups half of the sample always sided with T. and half of the sample never or only once. The patterns for the alternative use of common objects were very different. Thus, when T. said, “it’s ok to use a bucket to put

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

5

Fig. 1. Proportion of children’s responses siding with their own teacher (Study 1) or the majority of teachers (Study 2).

Fig. 2. Proportion of children’s responses siding with their own teacher by Age (Study 1).

*p < .05; *** p < .001.

Fig. 3. Chance level for each trial in each situation (Study 1, by Age group; Study 2, by Condition). DC: dissenter condition; NDC: non-dissenter condition.

6

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

Table 2 Percentages of children who sided (always, once, or never) with their own teacher (T.), by age and situation (Study 1) and with the majority of teachers (M.) by condition and situation (Study 2). Study

Situation Alternative use of common objects Non-conventional

Labeling Conventional

Study 1 Never side with T. Side once with T. Always side with T.

3 yrs 27.7 16.6 55.5

5 yrs 94.4 0 5.5

3 yrs 33.3 16.6 50

5 yrs 0 0 100

3 yrs 5.5 44.4 50

5 yrs 11.1 44.4 44.4

Study 2 Never side with M. Side once with M. Always side with M.

DC 75.6 7.3 17.1

NDC 60 12 28

DC 12.2 4.9 82.9

NDC 0 10 90

DC 24.4 39 36.6

NDC 16 24 60

DC: dissenter condition; NDC: non-dissenter condition.

salad in” and “it’s ok to use a bottle to drink the soup” only one older child always sided with T. whereas 55.5% (10 out of 18) of younger children did so. However, in the cases in which T. was conventional the results were different: All of the older children always sided with T., while this percentage decreased to 50% among younger children. This first study reveals the limits of the influence of a teacher on preschoolers’ decisions. The results show that children do not blindly accept their teacher’s claims but they seem to analyze the information provided. This pattern becomes evident among the 5-year-old children: their decisions were guided by the type of claim (conventional or non-conventional), and not by the source of information (teacher or strange). In other words, the older children massively sided with the conventional perspective, regardless of whether this person was the teacher or the stranger (see Chan, 2011; for similar results with 5-year-olds). By contrast, half of the 3-year-olds always endorsed their teacher’s claims, regardless of whether she held a conventional view or not. It must be remembered that the pilot study showed that children as young as 3 knew the objects and their conventional uses, and virtually all of them were against using them in a different way. If we analyze only the results for when T. is non-conventional, it could be argued that many children were willing to abandon their prior beliefs and consider a new possibility presented by a teacher (45.5% of children sided once or twice with the teacher’s alternative view). We could then conclude that they have less confidence in their own knowledge. However, this explanation is less convincing when we analyze the situations in which T. was conventional: 33.3% of the children never sided with her and 16.6% sided with her only once. If we look at the results as a whole, another possible explanation is that the task is over-complex and not properly understood by the younger children. In this case, individual differences in task comprehension would disappear from 5 years, as is the case. Nevertheless, research on children’s functional fixedness (German & Defeyter, 2000; Kelemen, 1999) can also give us some clues to explain the behavior of the younger children who understood the task but also sided with the teacher when she represented non-conventional opinion. Some studies have found that younger children, particularly under the age of 5, are less rigid than older children and adults regarding the original intended function of an object (common artifact). Thus, young children are more likely to accept a novel function for a familiar object than older children and adults. Our results fit well with this hypothesis since the older participants gave stronger responses in favor of the proper function of a common object than the 3 year-olds, who apparently showed more functional flexibility (German & Defeyter, 2000, p. 708). However, the results do not allow us to assert whether participants’ responses were guided by the influence of the teacher’s authority or by this supposed “flexibility”. Some children arguably

sided with the teacher simply because of who she was, without judging what she said. Nevertheless, this hypothesis needs further investigation to evaluate its explanatory power. In the following study, the 3-year-old group is not included and the target group is the older preschoolers. Regarding the labeling situation, we found that children’s agreement with their teacher was lower than expected. Around half of the participants always sided with the teacher but the rest of children did so only once or never. Moreover, in both age groups children’s agreement with their teacher was above chance in only one of the two trials (the second trial in the case of the 3-yearolds, the first trial in that of the 5-year-olds), but at chance in the other trial. These findings differ from past studies showing that preschoolers tend to trust familiar teachers more than unfamiliar ones when endorsing new labels (Corriveau & Harris, 2009). However, it must be noted that previous studies did not analyze different contexts at the same time (they only focus on assigning novel labels or functions to unfamiliar objects), while in the present study different situations were interwoven in each task. Furthermore, it is worth noting that some studies have found that 3-year-olds begin to doubt an informant after a single error (Pasquini, Corriveau, Koenig, & Harris, 2007). In the present study, the informants do not make such obvious “mistakes” as in the previously mentioned works (e.g., calling a spoon a duck); their non-conventional proposals are perfectly plausible. Nevertheless, we cannot rule out that in some case children interpret them as mistakes. In any event, these differences make it difficult to directly compare our results with previous ones in the field. Regardless of these results of the labeling task, in general terms, we expected the judgement of their own teacher to have greater influence than was actually found. The older children’s extreme opposition to non-conventional use of objects even when this conflicts with the teacher’s plausible proposal is especially surprising. These results motivated the design of the next study. In order to continue exploring the influence of teachers’ opinions on preschoolers’ decisions, we adapted a paradigm that analyzes the influence of consensus, frequently used in works on trust in testimony. These studies usually present unfamiliar, adult, female informants defending a majority consensus opinion. In our study, the informants appeared in a video set in a classroom and were introduced as teachers from another school. Thus, in Study 1 the aim was to evaluate the pressure of a familiar epistemic authority (the teacher) on children’s decisions, while Study 2 aims to evaluate majority pressure. Although some previous works have shown that preschoolers are influenced by a majority formed by unfamiliar adults (Chen et al., 2011, 2013; Corriveau, Fusaro et al., 2009; Schillaci & Kelemen, 2014), children’s trust in consensus opinions has its limits. There is reasonable empirical evidence that children consider

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

the plausibility of information, and that when this information conflicts with their own direct experience, they are reluctant to defer to the majority opinion. Studies with children based on the classic work by Asch (1956) on simple perceptual judgments (i.e., identifying the correct size of a model) indicate that less than a third of preschoolers defer to the obviously incorrect consensus opinion, regardless of whether the majority opinion is that of adults (Corriveau & Harris, 2010) or of children (Haun & Tomasello, 2011). Will the majority opinion of teachers have a greater influence in situations such as those included in Study 1 regarding the use of common objects? 3. Study 2 The objective of Study 2 was to assess preschoolers’ sensitivity to the information provided by a numerical majority of teachers regarding the two conventional domains of knowledge explored in Study 1. Moreover, an additional objective was to explore whether the presence of a dissenter modulates their trust in majority. A number of studies with adults show that opposing a majority is easier when consensus is not unanimous: the presence of just one person who answers correctly is enough to greatly reduce conformity to the erroneous or dubious judgment of a majority (Allen & Levine, 1971; Morris & Miller, 1975). However, to our knowledge, there are no previous studies controlling the influence of a dissident on preschooler’s judgements and little is known of this effect in children. Until now, studies on majority influence have explored how the pressure of consensus affects children’s decisions either in the absence of a dissident (Haun & Tomasello, 2011; between peers; Corriveau & Harris, 2010; with adults) or with the presence of a dissident (Chen et al., 2013; Corriveau, Fusaro et al., 2009). Nevertheless, the characteristics of the domains of knowledge explored in these works (in the first case, simple perceptual judgments, in the second case, unfamiliar objects) and of the informants, make it difficult to compare these studies and determine the weight of the dissident’s opinion. In this second study, there were two between-subjects conditions: Presence or absence of a dissenter. Children had to make decisions after listening to the conflicting claims of three teachers vs. one teacher (dissenter condition, hereinafter DC), or to the unanimous opinion of three teachers (non-dissenter condition, hereinafter NDC). Presumably, it might be easier for children in the dissenter condition to go against a majority of teachers since the dissident is also a figure of authority. In contrast, children in the NDC might feel less inclined to oppose a majority of teachers without the support of a dissident. Overall, and taking into account the results of Study 1, as well as previous findings on the weight of the majority on children’s opinion, in the labeling contexts, participants were predicted to side more with a majority of teachers than with a single dissenting teacher. In situations regarding the use of common objects, they were predicted to be more likely to accept a non-conventional use proposed by the majority than by a single dissenter. Regarding the role of a dissenter teacher, we predicted in all situations that children would show greater deference to a unanimous majority (NDC) than to a partial majority (DC). 3.1. Method 3.1.1. Participants Ninety-one preschoolers participated in this study (M = 62 months, SD = 7 months, range: 48–72, 44 male). Children were recruited from public and semi-private schools serving lower- to middle-SES families in Madrid and Toledo province, (both in Spain). Most (90.3%) belonged to the majority ethnic group (the

7

rest were 4.2% Rumanian, 1.4% Cameroonian, 1.8% Moroccan, 1.4% Ecuadorian and 0.9% Peruvian). All children were fluent in Spanish. Children were individually interviewed by a female researcher. Written parental consent, as well as children’s verbal assent, was obtained for all participants, as well as from the women shown in the films. Participants in the study were randomly selected from among those who presented parental consent and accepted to participate. 3.1.2. Procedure The procedure, contexts and materials were analogous to Study 1, and the videos and instructions were adapted. Thus, instead of referring to the informants as “your teacher/the stranger”, they were presented as “These three teachers/this teacher”, depending on the condition. See Supplementary material for details. 3.1.2.1. Dissenter and non-dissenter condition. Children were randomly assigned to one of the two experimental conditions in a between-subject design: the dissenter (DC) and non-dissenter conditions (NDC). In the DC (3 teachers vs. 1 teacher), all participants watched films showing a three-person consensus and a dissenter, who made conflicting claims about the same situations as in Study 1. To reinforce the epistemic authority of the informants, all of them were presented as teachers from another school. They were shown with a blackboard behind them and a table in front of them. All informants were young, female and from the same racial group (Caucasian). They had similar body types, presented neutral facial expressions, and were recorded wearing white T-shirts and blue jeans with their hair loose and away from their faces. Children watched the teachers in two trials of labeling tasks and in four trials judging the alternative use of common objects, in the same way as in Study 1: In two of them, the three teachers were nonconventional1 (and the dissenter claimed the opposite) and in the remaining two, the three teachers were conventional (and the dissenter claimed the opposite). In the NDC (3 teachers in consensus), the films presented were the same as in the DC but without the dissenting teacher, who was deleted using video editing software. See Supplementary material for details. In Study 1, the order of trials was the same for the whole sample. In order to control for the possible effect on participants’ responses, in Study 2 participants were randomly assigned to one of two sequences of stimulus presentation (See the Table 1). The coding procedure was the same as in Study 1. Thus, for all analyses, participants’ dichotomous responses to each trial were scored 1 point, when they chose the majority option and scored 0 points, when they chose the alternative option (the dissenter view, in the DC; the alternative view, in the NDC). Participants’ dichotomous responses across the two trials within the same situation were compared (McNemar tests) and no significant differences were found between agreement with the majority in the two trials

1 We thank the anonymous reviewer who raised a concern about the overlap between frequency of information and consensus in this design. The frequency of the responses of the consensus vs. dissident was a deliberate decision in the study design: the notion of consensus (or numerical majority) necessarily implies “greater frequency”. It is unusual to find situations in daily life where consensus (majority) opinion is comparable in frequency to dissident opinion (minority). The design we used therefore contributes to the ecological validity of the study. Furthermore, previous works also used the design we have: 3 informants opposed to 1 dissident (Chen et al., 2011, 2013; Corriveau, Fusaro et al., 2009). Although in these cited works the informants do not verbally repeat the label (because they point to it), there is a higher frequency of subjects in one option than in the other, and more “fingers” pointing to one option than the other. Additionally, in some of these works in which the informants pointed to the object (Corriveau, Fusaro et al., 2009), the authors controlled for the frequency of hands pointing to the object over consensus. Even then, the results showed that the children responded to the consensus and not to the frequency.

8

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

within-situation. Scores for the three situations were also combined to yield a total score for each participant ranging from 0 to 6. 3.2. Results and discussion Considering the total sample, analyses of gender differences yielded non-significant results. Regarding the order of trials, no significant differences were found between agreement with the majority in first trial and second trial within-situation (McNemar tests). Moreover, considering the possible influence (previously mentioned in the discussion on study 1) of the inter-weaving of conventional/non-conventional trial types before the second labeling trial, the second labeling task in both orders (U-Mann Whitney tests) was compared, and no recency effect (the last thing the children had heard was non-conventional and could have influenced their responses in the next labeling trial) was observed in order 2. The same results were found for all the between-subject conditions (DC and NDC). Therefore, subsequent analyses were run on data collapsed across gender and order. 3.2.1. Trust in the majority of teachers Considering the total score for each subject in the three situation (six trials, range 0–6), children were significantly more willing to endorse the majority view than the alternative view, ␹2 (5, N = 91) = 32.36, p < .001. A Friedman test was conducted with the whole sample in order to analyze whether this tendency depended on the situation. It was statistically significant, ␹2 (2, N = 91) = 75.27, p < .001. Pair-wise comparisons using Wilcoxon tests with Bonferroni correction showed that all comparisons were significant (ps < .001), (see Fig. 1, white columns). 3.2.2. Influence of the dissident on trust in the majority Given that one of the aims of the second study was to explore whether children’s responses depended on the type of majority (with or without a dissident), Mann-Whitney tests were conducted to evaluate this possibility (see Fig. 4). The results revealed that, in general terms, siding with the majority was more likely in the NDC than in the DC, U = 692.500, n = 91, p < .01. Regarding each situation, differences between DC and NDC were found in the labeling situation: Children in the NDC condition sided with the majority choice more than children in the DC, U = 789.00, n = 91, p < .05. This same tendency was found for alternative use of common objects but the results were not statistically significant (non-conventional situation, U = 865.00, n = 91, p = .12; conventional situation, U = 940.00, n = 91, p = .25). In order to see the degree to which the tendency to side with the majority in each condition depended on the situation, a Friedman test was conducted for each condition. In both the DC and the NDC, siding with the majority depended on the situation, ␹2 (2, N = 41) = 34.16, p < .001, ␹2 (2, N = 50) = 41.30, p < .001, respectively. Pair-wise comparisons using Wilcoxon tests with Bonferroni correction show that all comparisons between situations were significant in both conditions (ps < .001). Binomial tests were conducted in each experimental condition, for each trial in each situation to check whether children responses were differently from chance (see Fig. 3, columns for Study 2). In the conventional trials, the children in both conditions sided with the majority above chance in all the trials (DC, n = 35, first trial, n = 35, second trial, ps < .001; NDC, n = 47, first trial, n = 48, second trial, ps < .001). In the non-conventional situation, in the DC, children’s responses were below chance in both trials (n = 9, first trial, n = 8, p = 1.00, second trial, ps < .001) while only in the second trial in the NDC (n = 18, first trial, p = .065, n = 16, p < .05, second trial). In the labeling situation, children in the DC responded at chance in both trials (n = 26, first trial, p = .12 n = 20, p = 1.00, second trial) while

responses were above chance in both trials in the NDC (n = 33, first trial, p < .05, n = 39, p < .001, second trial). 3.2.3. Degree of agreement with the majority of teachers As in Study 1, participants’ dichotomous responses to the two trials within the same situation were scored on a 2-point scale: Score = 0 when participants never agreed with the majority; score = 1 when participants agreed only with the majority; and score = 2 when participants agreed with the majority twice. Table 2 shows the percentage of children who sided with the three teachers by situation in each condition. When the three teachers were conventional, 82.9% in the DC and 90% in the NDC always sided with them (majority), and 12.2% in the DC and none of them opposed the teachers in the NDC. However, when the three teachers were non-conventional (i.e., it is ok to use a bucket to put salad in/drink soup from a bottle), only 17.1% of children always sided with them in the DC and this percentage increased to 28% in the NDC; in this situation, 75.6% (DC) and 60% (NDC) of children always opposed the three non-conventional teachers. In the labeling context, in the NDC the number of children always siding with the majority of teachers (60%) was almost double the number in the DC (36.6%), although there was still a considerable percentage of children who never chose the label proposed by the three teachers (24.4% and 16%, DC and NDC, respectively). The results of our second study show that 5 year-old children explicitly reject the consensus opinion of a majority of teachers if they do not agree with them, regardless of the presence or absence of a dissenter. That is, in the situations in which the majority of teachers presented non-conventional opinions, the children in both conditions opposed the majority and defended a conventional perspective of the object function. The findings regarding the labeling task reveal a different picture. In this context, the condition (presence or not of dissenter) had a significant influence on children’s decisions: Only when there was no dissenter did children endorse the majority claims above chance. In contrast, when a dissenter was present, they did not rely on the majority more than on the dissenter, that is, they seemingly behaved at chance. These findings are surprising in that they differ from the results of most previous studies (Chen et al., 2011, 2013; Corriveau, Fusaro et al., 2009; Fusaro & Harris, 2008). Other studies with Spanish children (Guerrero, Elenbaas, Enesco, & Killen, 2013; Guerrero, Elenbaas, Enesco, & Killen, 2017) also show that they tend to conform less to the opinion of a majority of adults than their peers in the United States. We will address this issue in the general discussion. Returning to the issue of dissenters, the results of this second study suggest that dissidents may have an important role in children’s decisions, even under uncertainty. Thus, when children are faced by problems to which they do not know the solution, as in the labeling situation, the presence of a dissenter proposing an alternative view makes children more reluctant to endorse the majority claim. In this respect, the percentages of children who always sided with the majority when learning new words are indeed illustrative: 36.6% and 60%, in the DC and NDC, respectively. When it comes to the context of objects functions, these percentages are much lower: 17.1% and 28% always sided with the majority favoring a non-conventional use of objects. Thus, it seems that prior knowledge of how things are used and reluctance to accept alternative proposals have a stronger influence on children’s decisions than the opinion of a majority of teachers. 3.2.4. Comparisons of children performance across Studies 1 and 2 3.2.4.1. Children’s deference to their own teacher or consensus. As seen in Fig. 1, the deference children show to their own teacher (Study 1) or a majority of teachers (Study 2) is simi-

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

9

1 0.9

*

0.8 0.7 0.6 0.5 0.4

0.3 0.2 0.1 0 Non-Convenonal

Convenonal DC

Labeling

NDC

Fig. 4. Proportion of responses that sided with the majority of teachers by Condition (Study 2). DC: dissenter condition; NDC: non-dissenter condition.

lar in all the situations studied. No significant differences were found (non-conventional situation, U = 1523.00, n = 127, p = .46; conventional situation, U = 1431.00, n = 127, p = .09; labeling situation, U = 1552.50, n = 127, p = .70). Thus, in both studies children behave in a reasonably consistent way: Their degree of confidence in their teacher (Study 1) or the majority of teachers (Study 2) largely depended on the content of information supplied by the teacher/s. When only the results for the 5-year olds in both studies are analyzed, this trend is maintained for both the conventional situation, U = 711.00, n = 109, p = .11, and the labeling situation, U = 816.00, n = 109, p = .98. The difference was significant in the non-conventional situation, U = 599.00, n = 109, p < .05. 3.2.4.2. Maintenance of trust from the first to the second labeling trial in Study 1 and 2. As we have pointed out in the discussion of the Study 1, previous works have demonstrated the sensitivity of preschoolers to the accuracy of informants and how this sensitivity can influence future trust in them (Bernard, Proust, & Clément, 2015; Corriveau & Harris, 2009). Although no significant differences were found regarding order effect in the Study 1, it is interesting to offer the analysis of the degree to which children lose trust in their own teacher and in the majority of teachers in the second labeling trial. It must be remembered that in all the cases the children heard all the informants defend a conventional and a non-conventional opinion before the second trial. Consequently, inaccuracy was evenly distributed. In Study 1, 47.2% of the children maintained their trust in their teacher from the first to the second trial (50% and 44.4% in the 3- and 5-year-olds, respectively), while only 25% abandoned this trust (11.1% and 38.8% in the 3and 5-year-olds, respectively). The percentages are similar in Study 2, where 49.4% of the children maintained their trust in the three teachers from the first to the second trial, while very few abandoned this trust in the second trial (15.4%). 3.2.4.3. Justifications across studies. In the two studies children were asked to justify their choices (see Table 3). For the stories regarding the transgression of socio-conventional rules (“Why do you think she/they is/are right?”), justification categories were based on previous studies on children’s thinking about conventions, drawing on social domain theory (Turiel, 1983). The categories were (a) knowledge attributed to the informant/s (“My teacher is very smart and knows about everything”), (b) conventions and social traditions (“No child goes to school in pajamas”), (c) consequences of breaking the rule (“Children will laugh at him”), (d) circular or undifferentiated responses with no additional reasons

*p < .05.

(“Because it’s not ok”). The coding was conducted by two coders blind to the hypotheses of the studies. On the basis of 25% of the interviews (n = 9 data points for Study 1, n = 24 data points for Study 2), inter-rater reliability was calculated at Cohen’s K = .94 (Study 1) and Cohen’s K = .91 (Study 2). In this context, no justifications were given in Study 2 based on the consensus or numerical composition of the group of informants. Participants’ justifications for their decisions in the labeling task (“Why do you think she is/they are right?”) fell into three categories: (a) knowledge attributed to the informant/s: (e.g., “she knows a lot about things. . .”), (b) physical attributes of the object (e.g., “it has a circle”), and (c) circular or undifferentiated responses with no additional reasons (“Because it’s a reso”). In Study 2, we expected some children to give justifications based on the consensus or numerical composition of the group of informants such as, for example, “they all say the same thing”. However, none of the participants gave this type of response, not even when the majority was unanimous. 3.2.4.4. Justifications in the alternative use of common objects context. Table 4 shows children’s justifications for their decisions in the two contexts (use of common object and labeling) across the two studies, according to whom they chose among the informants. In the context of the alternative use of objects, both in Study 1 and Study 2 the distribution of the justifications across the different categories was related to the informant chosen by the children. This difference was statistically significant in all situations (Study 1), non-conventional situation 2 (3, N = 72) = 22.43, p < .001, : 0.558, conventional situation 2 (3, N = 72) = 24.84, p < .001, : 0.587; Study 2, non-conventional situation 2 (3, N = 182) = 70.10, p < .001, : 0.621, conventional situation 2 (3, N = 182) = 41.77, p < .001, : 0.479. In this context, the most frequent justifications in both studies were based on traditions or conventions (the way common objects are used). This result coincides with the tendency in the two studies to reject alternative uses for objects. This category is followed by responses based on the consequences of breaking the rules. Both types of justifications were specifically given by children who rejected the non-conventional function of the object. In general terms, little reference is made to knowledge attributed to the informant/s. This reference is especially lacking in the Study 1. Most of the circular responses, as well as the knowledge- attributed justifications, were given by children who accepted the non-conventional use of the object.

10

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

Table 3 Examples of justifications for each category by context. Category

Examples of justifications

Alternative Use of Objects Context Knowledge attributed to informant Conventions and social traditions Consequences of breaking the rule Circular or undifferentiated

She/teachers knows/know how to use the fork properly Forks are for eating and combs for combing; No child goes to school in pajamas Children will laugh at him; You could poke the fork in your head Because it is not ok; I don’t know

Labeling Context Knowledge attributed to informant Physical attributes of the object Circular or undifferentiated

She knows a lot about resos; My teacher is very smart and knows about everything; They are teachers, so they know it It has a circle; This is the reso because it has reso-like features Because it is a reso; I don’t know.

Table 4 Percentages of justifications by situation and informant in Study 1 (teacher vs. stranger) and in Study 2 (majority vs. dissenter/non dissenter). Alternative use of common objects Study Non-conventional Study 1 Teacher n = 25 Stanger n = 47 Study 2 Majority n = 51 a Alternative n = 131 Conventional Study 2 Teacher n = 58 Stanger n = 14 Study 2 Majority n = 17 Alternative n = 165

Knowledge attributed

Conventions

Consequences

Undifferentiated

12 10.6

12 55.3

12 21.3

64 12.8

9.8 .8

25.5 64.1

7.8 28.2

56.9 6.9

10.3 14.3

48.3 0

25.9 7.1

15.5 78.6

4.8 5.9

62.4 23.5

24.2 5.9

8.5 64.7

Labeling Study

Knowledge attributed

Physical attributes

Undifferentiated

Study 1 Teacher n = 50 Strange n = 22

42 22.7

14 18.2

44 59.1

Study 2 Majority n = 118 Alternative n = 64

12.7 0

27.1 35.9

60.2 64.1

Note. The percentages are calculated across the total number of responses, not of participants. Study 1, 72 responses; Study 2, 182 responses. a Alternative refers to dissenter and non-dissenter condition for Study 2.

3.2.4.5. Justifications in the labeling context. As in the alternative use of common objects, we also find that justifications in labeling context were related to the informant the children sided with (see Table 4), but it was statistically significant only in Study 2 (2 (3, N = 182) = 9.31, p < .05, : 0.226). Broadly speaking for the two studies, around half of the justifications in this context were circular or don’t know responses. Among the categorizable responses in this context, reference was made to the characteristics of the object. This intriguing result suggests that some participants appear to understand that they have to guess the correct name of the object based on its physical characteristics (i.e., this is the reso because it has reso-like features). The category of knowledge attributed to the informant was more frequent when children sided with their own teacher (Study 1) than when they sided with a majority of teachers (Study 2). Interestingly, the percentage of children that repeated or endorsed the novel object labels in their justifications was similar in the two studies (13–16%). 4. General discussion The general objective of this work was to analyze the influence of the information provided by schoolteachers on preschoolers’ decisions in two contexts: The non-conventional use of objects and learning new words. In the two studies presented here, most of

the children opposed the information provided by teachers when it conflicted with conventional expectations. Overall, our participants were more likely to interpret the events guided by their own knowledge than by information offered by others, regardless of the epistemic authority attributed to these others (coinciding with previous research, Chan & Tardif, 2013). For children, teachers are not only attachment figures they usually see as having good intentions, but also authorities in the dual sense of powerful and knowledgeable (well-informed) persons (Olson & Bruner, 1996; Siegel, 2005). In this sense, it could be expected that children’s decisions are largely influenced by their teachers. However, our results show that children’s previous knowledge may outweigh teachers’ epistemic authority as well as the pressure of the consensus, at least in contexts of informal knowledge such as those presented in these studies. It is difficult to quantify and separate the specific weight of children’s prior knowledge and the nature and content of testimony. Regarding the alternative use of objects, the pilot study confirmed that, from 3 years, children have very clear ideas about the conventional use of the everyday objects used in our research. The results of the two studies are clear as to the children’s attitude towards breaching socio-conventional rules: Their normative orientation is more powerful than both the teacher’s opinion and the pressure of consensus. In the labeling task, although a priori the children

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

were unfamiliar with the object and the label, they were not totally “blind” or naïve with respect to the new information. The children’s responses reveal that they utilize a series of strategies including the use of testimony from their own teacher or a majority of teachers but this is by no means the only or the most commonly used strategy. As previously seen, when a person provides an alternative label (Study 1 and DC in Study 2) several children reject the label for a new object provided by their own teacher or by a majority of teachers. The dissident voice appears to make children doubt their trust in the information. Yet, how can we explain that even in the absence of a dissident (Study 2), several children (40%) opposed the unanimous opinion on at least one occasion? It must be emphasized that the children were asked Who do you think is right? (Study 1 and Study 2, DC) or Do you think they are right? (Study 2, NDC), and they were not asked to provide the correct response themselves (What is this?), a prototypical question raised in previous studies in this field (Chen et al., 2011, 2013; Corriveau, Fusaro et al., 2009; Koenig, Clément, & Harris, 2004). In spite of this specific question (Who is right?), which is designed to evaluate trust in the informants and not children’s opinion, the justifications reveal that several children interpreted the task as having to discover the relationship between the object and its name, which recalls what Piaget (1926) described as nominal realism. The open question is whether the discrepancies between our results and those of previous studies are due to differences in methodology, or to specificities of the culture and education. Some cultural differences in acceptance of consensus opinion have been noted for both children (Chen et al., 2011; Corriveau & Harris, 2010; Diyanni, Corriveau, Kurkul, Nasrini, & Nini, 2015) and adults (Bond & Smith, 1996). It may be that the educational system and the subtleties of the child-teacher relationship are different in Spain and the US, especially as regards “deference” to authority. This is reflected in norms of discipline and conventions both inside and outside the classroom (e.g., the way children speak to teachers and other adults, norms on how to participate in class or how to enter a conversation). For instance, in multi-ethnic classrooms in Spain, teachers tend to typify Spanish children as less disciplined in terms of norms and authority than Asian, Latin American or other European children. However, these are no more than teachers’ observations and there is currently no empirical research to support them. Consequently, it is unwise to draw conclusions on cultural differences even within Western societies. More research is needed on the possible cultural differences underlying this phenomenon. Another limitation of the paradigm used in this work is that we cannot actually know whether children who accept the nonconventional opinion of an informant do so merely in deference to authority or because they are really “convinced”. Although the children’s justifications provide a more complete vision of how children reach their decisions, the information is not sufficiently complete to reveal the real motivation of their choice. It is surprising that children show so little deference to the opinion of both their own teacher and the consensus (unanimous or partial). Aspects of the experiment itself may well have influenced what children perceived to be the task at hand. What was the role of the experimenter in the testing room? Was the experimenter the “hidden” epistemic authority in the study, there to evaluate the accuracy of the children’s answers? What would a preschooler participating in this work see as their goal: Was it about making a decision on trusting information provided by different informants, or about choosing the correct (or least risky) answer for the experimenter? These questions2 could be partly addressed in

2

We thank the anonymous reviewer who raised this interesting point.

11

future studies using other contexts, such as pretend play, where children arguably accept non-conventional uses of objects such as those presented here. Wyman, Rakoczy, & Tomasello (2009) found that 3-year-olds reacted differently according to the context. In pretend play, they accepted the non-conventional, fictitious use of an object and protested when the target object was treated according to its real function. This shows that from an early age children make fine discriminations about how and when objects can be used conventionally or non-conventionally, and that they have a strong normative orientation regardless of whether the context is serious or one of play. Despite the differences between the objectives and procedure of Wyman et al. (2009) and those of the current work, our results show a similar tendency with regard to the normative orientation of children. It is worth noting that none of our participants interpreted the non-conventional situations as a game or as an acceptable alternative in the case of lacking a better option (e.g., not having a comb to comb your hair). It is interesting to note that the highest percentage of children actively opposing authority is found in Study 1, where 94% of the older children questioned their teacher’s non-conventional opinion (this percentage decreases in the Study 2, 75.6% and 60%, with and without dissenter, respectively). Factors related to the dynamics of the teaching-learning process implemented by the teacher herself probably had an impact on this finding. Angelines, the actual teacher of the participants in Study 1, is a kind, approachable teacher, concerned about the emotional development of the children, who are given freedom to express themselves in the class. A number of previous studies which take into account the emotional relationship between children and teachers and the trust invested in teachers (Corriveau & Harris, 2009) provide no conclusive results on the association between these variables. However, if school is a place where preschoolers can express themselves freely, and are encouraged to do so by the teacher, then they may plausibly be critical of authority and even question it. As mentioned in the discussion of Study 1, there are a number of studies on the relevance for children of informants’ past accuracy (Corriveau, Meints, & Harris, 2009; Jaswal & Neely, 2006). Given the design of the tasks in our two studies, it could arguably be interpreted that some results were contaminated by this dimension since during the procedure the children hear the authority saying strange things (“yes, you can use a bucket to put salad in. . .” or “. . . to use a bottle to drink soup”). It is true that although such proposals are strange, they are still plausible (we can put salad in a bucket, drink soup in a bottle). Despite this, the children might be trapped by the salience of their own teacher or a majority of teachers saying something strange. However, this is unlikely to have influenced the results since the statistical analyses show that, on one hand, there are no differences between trials in the same situation; and on the other, the order of the tasks – controlled in Study 2 – did not impact on the responses. Also, it is worth remembering the results of the intra-individual analysis of trust from the first to the second labeling trials: in both studies, overall, there were more children who maintained their trust in authority than those who lost it. This work studied to what extent children choose to trust teachers defending non-conventional statements. To this end, we provided an additional analysis considering the participants as a referee unit and not only a proportion of their responses. In the results we included the number of participants that always, never or only once side with authority. To our mind, studies which only present the mean of the responses provide useful but incomplete information on the phenomenon, since the number of children who present truly conformist behavior cannot be directly inferred, especially if the mean is obtained across a large number of trials (Haun & Tomasello, 2011). Indeed, recent studies present both the mean of the responses and the number of participants following a certain

12

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13

response pattern (Einav, 2014; Schillaci & Kelemen, 2014). In this field of research, it is of great importance to identify individual patterns, as well as sample trends, since sensitivity to the testimony of others may be especially affected by personal variables that go beyond strictly developmental characteristics. Society tends to think that children are sponges who acquire information quasi-automatically, especially if the information comes from a teacher in a school setting. However, developmental research tells us that children do not approach school as tabulae rasae, but hold varied information and theories about the world unmediated by formal school learning. There is no doubt that teachers are an essential part of the learning process, but being their teacher is not enough to instill new knowledge or modify children’s beliefs. Children’s prior ideas, their ways of reasoning, their epistemological beliefs and metacognitive strategies appear to play a central role in the transmission of knowledge in a formal setting such as a school (Campanario & Otero, 2000). Some studies show that even in formal contexts the teacher’s explanation accompanied by empirical evidence or conceptual explanation is not enough to change pupils’ prior erroneous ideas, e.g., in the understanding of floating and sinking (Howe, Rodgers, & Tolmie, 1990) or the comprehension of the logic of counting (Enesco, Rodríguez, Lago, Dopico, & Escudero, 2016). Consequently, it seems that what children know about new information before it is presented has a great influence and can interfere in the learning process. This phenomenon arguably has two readings: On one hand, it explains why some university students hold inaccurate beliefs about simple physical phenomena (Driver, 1988; Kruger, Palacio, & Summers, 1992; Viennot, 1979); and on the other, it has a protective dimension, in that children do not indiscriminately gobble up information. Future research should attempt to generate more realistic situations: Watching unknown adults making different judgments with a serious expression in front of a camera is not a situation close to a child’s world. Similarly, hearing a teacher say that a beach bucket can be used as a salad bowl, without providing a justification, is, at best, a strange situation. The phenomenon needs to be approached in a way which is more natural and meaningful for the children, by, for example, giving reasons why it may or may not be justifiable to breach the conventions, motives or interest underlying different people’s opposing views, or by presenting these situations in the context of pretend play. More realistic settings with different epistemic content in which children’s naïve theories come into conflict with new information (Howe & Tolmie, 2003) could provide new data, which, analyzed from the viewpoint of the trust testimony paradigm, would not only help teachers in their work but would also help to understand the learning process across a child’s development.

Acknowledgments We owe thanks to Irene Arias, Pilar García-Valcárcel, Jennifer González, Ma Ángeles Piedrabuena and Laura Villagarcía for their collaboration in the construction of the materials for this research. ˜ Carla Sebastián-Enesco The authors wish to thank María Núnez, and the anonymous reviewers for providing many helpful comments.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecresq.2016.11. 001.

References Allen, V. L., & Levine, J. M. (1971). Social support and conformity: The role of independent assessment of reality. Journal of Experimental Social Psychology, 7, 48–58. American Psychological Association. (2011). Publication Manual of the American psychological Association (6th edition). American Psychological Association: Washington. Asch, S. E. (1956). Studies of independence and conformity: I. A minority of one against a unanimous majority. Psychological Monographs: General and Applied, 70(9), 1–70. Bernard, S., Proust, J., & Clément, F. (2015). Four- to six-year-old children’s sensitivity to reliability versus consensus in the endorsement of object labels. Child Development, 86, 1112–1124. http://dx.doi.org/10.1111/cdev.12366 Bond, R., & Smith, P. B. (1996). Culture and conformity: A meta-analysis of studies using Asch’s (1952b, 1956) line judgment task. Psychological Bulletin, 119, 111–137. http://dx.doi.org/10.1037/0033-2909.119.1.111 Campanario, J. M., & Otero, J. C. (2000). Más allá de las ideas previas como dificultades de aprendizaje: las pautas de pensamiento, las concepciones epistemológicas y las estrategias metacognitivas de los alumnos de ciencias. Ense˜ nanza de las Ciencias, 18(2), 155–169. Chan, C. C. Y., & Tardif, T. (2013). Knowing better: The role of prior knowledge and culture in trust in testimony. Developmental Psychology, 49(3), 591–601. http:// dx.doi.org/10.1037/a0031336 Chan, C. C. Y. (2011). The role of prior belief in children’s trust in testimony: A social ecological approach. (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses. (UMI No. 3458832). Chen, E. E., Corriveau, K. H., & Harris, P. L. (2011). Children are sociologists. Invited paper for special issue on Prejudice: Sociodevelopmental perspectives. Anales de Psicología, 27, 625–630. Chen, E. E., Corriveau, K. H., & Harris, P. L. (2013). Children trust a consensus composed of outgroup members – but do not retain that trust. Child Development, 84, 269–282. http://dx.doi.org/10.1111/j.1467-8624.2012.01850. x Corriveau, K. H., & Harris, P. L. (2009). Choosing your informant: Weighing familiarity and recent accuracy. Developmental Science, 12, 426–437. http://dx. doi.org/10.1111/j.1467-7687.2008.00792.x Corriveau, K. H., & Harris, P. L. (2010). Preschoolers (sometimes) defer to the majority in making simple perceptual judgments. Developmental Psychology, 46, 437–445. http://dx.doi.org/10.1037/a0017553 Corriveau, K. H., Fusaro, M., & Harris, P. L. (2009). Going with the flow: Preschoolers prefer non-dissenters as informants. Psychological Science, 20, 372–377. http:// dx.doi.org/10.1111/j.1467-9280.2009.02291.x Corriveau, K. H., Meints, K., & Harris, P. H. (2009). Early tracking of informant accuracy and inaccuracy. British of Developmental Psychology, 27, 331–342. http://dx.doi.org/10.1348/026151008X310229 DiYanni, C. J., Corriveau, K., Kurkul, K., Nasrini, J., & Nini, D. (2015). The role of consensus and culture in children’s imitation of inefficient actions. Journal of Experimental Child Psychology, 137, 99–110. http://dx.doi.org/10.1016/j.jecp. 2015.04.004 Driver, R. (1988). Un enfoque constructivista para el desarrollo del currículo de nanza de las Ciencias, 6, 109–120. ciencias. Ense˜ Einav, S. (2014). Does the majority always know best? Young children’s flexible trust in majority opinion. PLoS One, 9(8), e104585. http://dx.doi.org/10.1371/ journal.pone.0104585 Enesco, I., Rodríguez, P., Lago, O., Dopico, C., & Escudero, A. (2016). Do teachers’ conflicting testimonies influence children’s decisions about unconventional rules of counting? European Journal of Psychology of Education, http://dx.doi. org/10.1007/s10212-016-0319-4 Fusaro, M., & Harris, P. L. (2008). Children assess informant reliability using bystanders’ non-verbal cues. Developmental Science, 11, 781–787. http://dx.doi. org/10.1111/j.1467-7687.2008.00728.x German, T., & Defeyter, M. A. (2000). Immunity to functional fixedness in young children. Psychonomic Bulletin & Review, 7, 707–712. http://dx.doi.org/10.3758/ BF03213010 Guerrero, S., Elenbaas, L., Enesco, I., & Killen, M. (2013). Young children’s use of trust and consensus varies by the context. In Symposium presented at Biennial meeting of the Society for Research in Child Development. April. Guerrero, S., Elenbaas, L., Enesco, I., & Killen, M. (2017). Preschoolers’ trust in social consensus varies by context: conventional vs. moral domains. Annals of Pyschology, 33(1) http://dx.doi.org/10.6018/analesps.33.1.230831 Harris, P. L., Corriveau, K. H., Pasquini, E. S., Koenig, M., Fusaro, M., & Clément, F. (2012). Credulity and the development of selective trust in early childhood. In M. J. Beran, J. L. Brandl, J. Perner, & J. Proust (Eds.), Foundations of metacognition (pp. 193–210). Oxford University Press. Harris, P. L. (2012). Trusting what you’re told: How children learn from others. Cambridge, MA: The Belknap Press/Harvard University Press. Haun, D., & Tomasello, M. (2011). Conformity to peer pressure in preschool children. Child Development, 82, 1759–1767. http://dx.doi.org/10.1111/j.14678624.2011.01666.x Howe, C. J., & Tolmie, A. (2003). Group work in primary school science: Discussion, consensus and guidance from experts. International Journal of Educational Research, 39, 51–72. http://dx.doi.org/10.1016/S0883-0355(03)00073-9 Howe, C., Tolmie, A., & Rodgers, C. (1990). Physics in the primary school: Peer interaction and the understanding of floating and sinking. European Journal of Psychology of Education, 5(4), 459–475.

S. Guerrero et al. / Early Childhood Research Quarterly 39 (2017) 1–13 Jaswal, V. K., & Neely, L. A. (2006). Adults don’t always know best preschoolers use past reliability over age when learning new words. Psychological Science, 17(9), 757–758. http://dx.doi.org/10.1111/j.1467-9280.2006.01778.x Johnston, A. M., Mills, C. M., & Landrum, A. R. (2015). How do children weigh competence and benevolence when deciding whom to trust? Cognition, 144, 76–90. http://dx.doi.org/10.1016/j.cognition.2015.07.015 Kelemen, D. (1999). The scope of teleological thinking in preschool children. Cognition, 70(3), 241–272. Koenig, M. A., & Stephens, E. (2014). Characterizing children’s responsiveness to cues of speaker trustworthiness: Two proposals. In E. Robinson, & S. Einav (Eds.), Trust and skepticism: Children’s selective learning from testimony. Cambridge, UK: Psychology Press. Koenig, M. A., Clément, F., & Harris, P. L. (2004). Trust in testimony: Children’s use of true and false statements. Psychological Science, 15(10), 694–698. Kruger, C., Palacio, D., & Summers, M. (1992). Survey of English primary teachers’ conceptions of force, energy and materials. Science Education, 76, 339–351. Lane, J., & Harris, P. L. (2014). Confronting, representing and believing counterintuitive concepts: Navigating the natural and the supernatural. Perspectives on Psychological Science, 9, 144–160. http://dx.doi.org/10.1177/ 1745691613518078 Mills, C. M. (2013). Knowing when to doubt: Developing a critical stance when learning from others. Developmental Psychology, 49, 404–418. http://dx.doi.org/ 10.1037/a0029500 Morris, W. N., & Miller, R. S. (1975). Impressions of dissenters and conformers: An attributional analysis. Sociometry, 38(2), 327–339. Nucci, L. (1981). Conceptions of personal issues: A domain distinct from moral or societal concepts. Child Development, 52, 114–121. Olson, D.R., & Bruner, J.S. (1996/98). Folk psychology and folk pedagogy. In D.R. Olson, & N. Torrance (Eds.), The handbook of education and human development. New models of learning, teaching and schooling (pp. 9–27). Cambridge: Blackwell.

13

Pasquini, E. S., Corriveau, K. H., Koenig, M., & Harris, P. L. (2007). Preschoolers monitor the relative accuracy of informants. Developmental Psychology, 43, 1216–1226. http://dx.doi.org/10.1037/0012-1649.43.5.1216 Piaget, J. (1926). The language and thought of the child. New York: Harcourt Brace & Company. Rakoczy, H., Warneken, F., & Tomasello, M. (2008). Developmental Psychology, 44(3), 875–881. http://dx.doi.org/10.1037/0012-1649.44.3.875 Robinson, E. J., & Einav, S. (Eds.). (2014). Trust and skepticism: Children’s selective learning from testimony. UK: Psychology Press. Rotenberg, K. J. (Ed.). (2010). Interpersonal trust during childhood and adolescence. New York: Cambridge University Press. Schillaci, R. S., & Kelemen, D. (2014). Children’s conformity when acquiring novel conventions: The case of artifacts. Journal of Cognition and Development, 15(4), 569–583. http://dx.doi.org/10.1080/15248372.2013.784973 Siegel, H. (2005). Truth, thinking, testimony and trust: Alvin Goldman on epistemology and education. Philosophy and Phenomenological Research, 71(2), 345–366. http://dx.doi.org/10.1111/j.1933-1592.2005.tb00452.x Smetana, J. G. (1981). Preschool children’s conceptions of moral and social rides. Child Development, 52, 1333–1336. Turiel, E. (1983). The development of social knowledge. Morality and convention. Cambridge, MA: Cambridge University Press. Van Ijzendoorn, M. H., Sagi, A. M., & Lanbernon, W. E. (1992). The multiple caretaker paradox: Data from Holland and Israel. In R. C. Pianta (Ed.), Beyond the parents: The role of other adults in children’s lives (pp. 5–24). San Francisco: Jossey-Bass Publisher. Viennot, L. (1979). Spontaneous reasoning in elementary dynamics. European Journal of Science Education, 1, 202–222. Wyman, E., Rakoczy, H., & Tomasello, M. (2009). Normativity and context in young children’s pretend play. Cognitive Development, 24, 146–155. http://dx.doi.org/ 10.1016/j.cogdev.2009.01.003