1 CHILDREN S UNDERSTANDING OF FALSE BELIEFS ABOUT PEOPLE S TRAITS By SUNGHEE AHN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Sunghee Ahn
3 ACKNOWLEDGMENTS I would like to gratefully and sincerely thank Dr. Scott Miller for his guidance, understanding, and patience. During the last six years of study in graduate school, he provided a well rounded experience consistent to my research interest and encouraged me to pursue my goal. His profound knowledge has been a major driving force for me to complete my work. I would like to express my deepest gratitude to him. I am also very thankful for having a great doctoral committee. I appreciate Drs. James Algina, Jeff Farrar, and John Chambers for serving as my committee members with their consideration and valuable advice. The benefits I gained from them will last throughout my life. Far too many people to mention individually have assisted in so many ways for my diss ertation. They all have my sincere gratitude. In particular, I would like to thank Wendy deBatiste, Nadia Nicolas, Sarah Sinai, Shannon Wilson, and Jillian Wilson. Finally and most importantly, my greatest blessing goes to my beloved husband, Dr. Youngmok Kim. His support, encouragement, quiet patience, unwavering love, and companionship became a strong energy for me to cope with any kinds of obstacles during my days in graduate school.
4 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 3 LIST OF TABLES ................................................................................................................................ 5 ABSTRACT .......................................................................................................................................... 6 CHAPTER 1 INTRODUCTION ......................................................................................................................... 8 The New Trait M easure .............................................................................................................. 12 Mastery of Beliefs ....................................................................................................................... 16 2 METHOD .................................................................................................................................... 19 Participants .................................................................................................................................. 19 Design .......................................................................................................................................... 19 Materials ...................................................................................................................................... 19 Procedure ..................................................................................................................................... 22 3 RESULTS .................................................................................................................................... 24 Analysis of Judgment .................................................................................................................. 25 Judgment of Other Belief .................................................................................................... 26 Judgment of Self Belief ....................................................................................................... 28 Analysis of Justification .............................................................................................................. 29 Justification of Other Belief ................................................................................................ 29 Justification of Self Belief ................................................................................................... 31 4 DISCUSSION .............................................................................................................................. 33 APPENDIX: BELIEF SCENARIOS .............................................................................................. 38 LIST OF REFERENCES ................................................................................................................... 42 BIOGRAPHICAL SKETCH ............................................................................................................. 45
5 LIST OF TABLES Table page 3 1 Proportion of c orrect r esponses to belief judgments on test questions .............................. 27 3 2 Proportion of r esponses to b elief j udgments on s elf q uestions .......................................... 28 3 3 Proportion of c orrect j udgments accompanied by a dequate e xplanations ......................... 30 3 4 Proportion of type of self justifications ............................................................................... 32
6 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHILDREN S UNDERSTANDING OF FALSE BELIEFS ABOUT PEOPLE S TRAITS By Su nghee Ahn August 2009 Chair: Scott A. Miller Major: Psychology Children aged 4 7 years ( N = 79) received 6 false belief tasks, 2 object and 4 trait tasks (2 Trait -congruent and 2 Trait incongruent conditions) testing their understanding of different kinds of beliefs On the object measure, a location of an object was changed whereas o n t he T rait tasks a behavior of a person changed. In the T rait congruent condition, a person s behavior representing his or her true trait changed temporarily due to an external factor In the Traitincongruent condition, a person s behavior also varied, but no information was provided about the true trait or the reason for the change On all of the T rait trials two perceivers were presented, one who saw one of the behaviors and one who saw the other behavior. Types of beliefs in the two Trait conditions therefore differed. The beliefs examined in the Trait -congruent condition were either a true or a false belief whereas the beliefs examined in the Trait -incongr uent condition were both true beliefs possessing a relative nature. After hearing each belief story, children received questions to test their understanding of different beliefs. On the object measures question s were directed to others beliefs only. On t he Trait measures, children were provided questions for both others beliefs and self belief. Findings showed that there were no differenc es in understanding of others beliefs across the three types of trial: Object, Trait congruent, and Trait -incongruent Older children were overall more successful than 4 year -olds
7 at providing correct judgments and relevant justifications on these measures. Comparisons of performance on the different kinds of beliefs provided on the Trait measure however showed some vari ations. Children performed equivalently on the true belief and false belief questions on the Trait measure but did more poorly on the self questions than on those directed to others beliefs. Children in particular had trouble responding to the Trait incongruent measure, suggesting their difficulty taking into account multiple perspectives Justification responses to self questions also evidenced qualitative change in styles of justifications in c hildren beyond the preschool ages. These variation s suggest the possibility of different kinds of beliefs that can observed in the social situation.
8 CHAPTER 1 INTRODUCTION Reality influences what we think and believe, and the way our minds operate al so shapes representation of reality. Theory of mind, what we understand about minds such as thoughts, emotions, and beliefs, serves as a basis f or successfully recognizing reality. In particular one im portant element for such recognition is appreciation of false beli efs, the idea that beliefs can be mistaken. The links between the reality and our beliefs can be explained by an emphasis on a false belief mechanism. Namely, our belief i s not a direct refl ection of reality but a result of subjectively processed mental representation. False belief has been one of the most popular research top ics in child development for two decades now. Of a variety of research questions that have been examined, one most fr equently asked is the age of mastery as to when children first become able to appreciate false beliefs. Such research is delimited to two specific uses of false belief methodology. One concerns general development of false belief measuring a scale of theor y of mind ability. A structured experimental setting of false belief allows researchers to assess and compare the developmental stages of such understanding in different groups of children. Most of the early part of the literature was devoted to this gener al aspect and there is a great understanding with regard to this knowledge. Researchers in general agree that false belief appreciation is a universal achievement; children between the ages of 3 and 5 in any cultural setting become able to successfully res pond to the two standard false belief paradigms, the unexpected transfer and the unexpected content forms ( for reviews, see Miller, 2000; Wellman, Cross, & Watson, 2001). The other topic for which false belief methodology has been used is relatively new, concerning more specific patterns of belief forms. As described in the previous section, the general aspect of false belief emphasizes a cognitive process that requires a certain level of
9 intellectual ability. The mastery of beliefs can however differ with the diverse contents about which the beliefs are formed. Ascribing a particular belief may dictate the use of certain knowledge that often turns out to be less important for other type s of belief s I n co ntrast to the general aspect of false belief, when and how children understand particular types of false beliefs is relatively unknown and new to the literature. Comparisons of the general and the specific aspect noted in the previous sections bear wi tness to one of most distinctive differences between the trends of false belief research, that is the differences in target s of beliefs employed The unexpected content and the unexpected transfer paradigms concern beliefs about physical facts either a con tent of a container or a location of an object. In contrast, the targets used in the specific aspect of false belief research are directed to beliefs about the social world such as a persons activity or a persons location. These person measures were crea ted in an attempt to have a more socially adapted form of false beliefs allowing researchers to examine possible relations between theory of mind and social development (e.g., Nguyen & Frye, 1999; Symons, McLaughlin, Moore, & Morine, 1997). A series of s tudies with employment of the person-oriented measures have demonstrated some variations in mastery of false beliefs. The first study was by Symons et al. (1997) who utilized a false belief measure with a focus on a location of a caregiver. On the measure a caregiver voluntarily moves to another location without the knowledge of he r child. The child must search for the parent after their separation and therefore when children hear the false belief story, negative emotions such as feeling s of anxiety or sad ness might disrupt their performance. Children between the ages of 3 and 5 had little success on the caregiver measure compared to the unexpected transfer trial. An interesting result was that there was no age -relat ed improvement on this caregiver trial wh ereas the performance on the object measures increased with age.
10 Childrens recognition of emotions negatively correlated with their performance on the caregiver trial, suggesting that it i s the negative emotions causing children to underperform. Another e xplanation for this underachievement is that children lack the ability to understand internality. Either internal or external motives can cause a persons movement, but children had special difficulty reasoning about a persons internally caused movement. They responded well to involuntary movement and object tasks in which the relocation result ed from an external agent. One conclusion offered by Symons et al. (1997) was that the internality in the voluntary movement trial disrupts performance because it is another form of mental state that children have to face in the task. This explanation is more legitimate for older children, in particular 5 year -olds who are in the process of mastering the belief mechanism. They become more sensitive to the situations c aused by another set of mental states and hence find the story more confusing and difficult to understand. Rai and Mitchell (2004) replicated the caregiver measure with a different form of test question which was known to help improve false belief perfor mance in some studies (Siegal & Beattie, 1991; Surian & Leslie, 1999). It however did not strikingly increase 5-year olds responses to the personinternal trials. This finding was also supportive of the claim by Symons et al. false beliefs involving an in ternal agent decrease 5 year -olds chances to succeed. Childrens problem with peoples voluntary action was demonstrated in another study. Nguyen and Frye (1999) developed a task in which a main character built a mistaken belief about a play partners intended activity. Performance of 3 and 5 -year olds on the trial was significantly lower than that on the unexpected transfer trial. This is consistent with what the other Western studies (Rai & Mitchell, 2004; Symons et al., 1997) argued with regard to inferior understanding of false belief s caused by an internal agent.
11 Naito and Koyama (2006) also compared person and object false belief measures but the sample was drawn from Japan, a non -Western countr y The person measure was an unbiased version of the caregiver location story created by Symons et al. (1997). Naito and Koyama changed the caregiver to a neutral person thus removing potential negative feelings caused by the caregiver -child separation. On this revised version of the person -internal measure, Japanese childrens performance did not lag behi nd the unexpected transfer form. Another cultur al study by Ahn and Miller (2009 ) provided an expanded approach from the study by Naito and Koyama. Ahn and Miller utilized the same formats of false beliefs that Naito and Koyama used, in addition to a self -concept measure to examine qualitative differences in self -conception possibly existing between Asian and Western children. Korean children outperformed th e US counterparts on both voluntary and involuntary person trials but the findings failed to replicate what the series of Western studies demonstrated. American childrens performance on the person -internal trial did not significantly differ from that on t he unexpected transfer trial. Summing up, the current false belief literature does not provide a consistent picture with regard to childrens false beliefs about people. As noted, the literature has produced two different results: better understanding of object measures than person and equal understanding of object and person measures. Regardless of what conclusions can be made from these contradictory findings, the current literature needs to provide further developmental evidence about false beliefs with regard to a persons internality by employing a new person task. In this way the discrepancy shown in the literature can be examined beyond the existing person paradigms that have at least one limitation.
12 The limitation of the existing person paradigms concerns the characteristics of the target which do not fully take into account the essence of people or the social world. The voluntarymovement measur e captures some of the important features by adding internal characteristics of a person to the measure. The target however is equivalent to what the standard object paradigms have presented. The target of the voluntary -movement measures is a physical location of a person just as a location of an object is provided by the unexpected trans fer scenario The N ew Trait Measure The present study propose d new person measures to rectify limitations of the previous person false belief methodology. The target of the new measure will be a persons trait, one of the topics common in the social cognition literature (e.g ., Barenboim, 1981; Cain, Heyman, & Walker, 1997; Heyman & Dweck, 1998; Liu, Gelman, & Wellman, 2007; Miller & Aloise, 1989; Rholes & Ruble, 1984; Yuill, 1992). The standard object measures present an object that is characterized from its physical informat ion. In contrast, p ersons behaviors can be explained with two types of internal processes immediate mental states or relatively permanent traits. Of these, one equivalent to the knowledge about physical fact s is the trait conception. A set of mental states provides a situation -based explanation which depends on the temporal status of the mind that can change at times. A trait -based explanation, on the other hand, concerns a more reliable and predictable interpretation of behavior Internal characteris tics of a person, namely his or her traits, are therefore considered as being analogous to physical characteristics of object s Another distinction between beliefs about traits and beliefs about object s concerns the nature of beliefs to be formed in diffe rent contexts. The object belief provides the exact answer for a true and a false belief. In contrast, understanding true and false belief s about peoples traits may be less clear and more difficult. A trait is commonly inferred from specific behaviors whi ch are often governed by an external factor (e.g., unexpected circumstances) rather than the trait
13 itself There is no absolute criterion that can be directly utilized for reasoning about the truth or falseness Furthermore, recognizing someone s traits is not a single cognitive process when different factors are involved in the context. It may require a higher level of cognitive ability to integrat e information from behavior al and situational factor s together. Due to the differences in the two belief fo rms just described in the earlier sections, t he new trait measures created for this study had two different features compared to the Object transfer form. One difference was that the trait measures have two conditions, the T rait incongruent condition and t he T rait -congruent condition The T rait -incongruent condition reflects the relative nature of belief s about the social world by representing two conflict ing forms of trait information. Children for example hear a story in which a main character sees a person behave nicely in one scene and another main character sees the person behave badly in another scene. In contrast to the Trait incongruent condition, t he belief p resented in the T rait congruent condition is more similar to the nature of the belief presented in the Object measure. In the Trait congruent condition, a persons trait is demonstrated in a more consistent way by adding an explanation to the conflict ing s ets of behaviors shown in the T rait incongruent condition. For example a main character knows very well about a person being nice and sees the persons nice behavior every day. In another scene, the person however cannot act nicely due to the occurrence o f an unexpected circumstance. In the last scene, another main character who does not know about the person sees the person s negative behavior. T he re were also procedural differences in the T rait an d the Object measure. The Object transfer form usually offers one test question directed to the false belief of the protagonist. The Trait measure provid es two test questions with regard to either a true or a false belief. The beliefs
14 asked in the Trait -incongruent condition are the true beliefs as formed give n the available evidence. These beliefs however possess a relative nature as well in that the child must realize that two people can have different beliefs about the same target. The test questions asked in the Trait-congruent condition concern either a tr ue or a false belief; t he belief of the first main character is true and the belief of the second main character is false. Furthermore, the Trait measures provided a self question to examine childrens own judgment. The self question has not been examined in the unexpected transfer trial even though the self -other distinction of mind understanding has been a central issue in the relevant literature (e.g., Bartsch, 2002; Gopnik & Astington, 1988; Gopnik & Wellman, 1992; Goldman, 1992; Harris, 1992; Lucariel lo, Donne, Durand, & Yarnell, 2006; Wellman & Gelman, 1992). Inclusion of the self question was to examine childrens subjective reasoning processes beyond the story information presented. The two different conditions of person measures were expected to re veal possible differences in self justification. There were t hree goals in creating the Trait measure. As just noted, investigating possible variations in children s judgments about different types of beliefs was the first goal. The second goal was to more closely examine beliefs that we experience in the social world. Trait judgment is a subjective process in real life. A persons trait is judged by the way the persons behaviors are evaluated by other people. Therefore what we observe about a pers on can be different from what other people do and this is the purpose for which the T rait incongruent condition is created. We however know that someones behaviors can be often governed by an external factor which sometimes operates opposite to his/her tr ue characteristics An example of this condition is the T rait-congruent trial.
15 The third goal was a more general one concerning children s development of trait understanding. The belief forms examined in the Trait measure had resemblance to the specific tr ait forms measured in the trait literature. Previous work in the trait literature has provided systematic comparisons of children s trait understandings using different types of inference ability. The types were behavior -to -behavior inference, behavior to trait inference, and trait to behavior inference. The behavior to -behavior inference has been utilized mainly for the traditional studies in which children are asked to predict other behaviors from behaviors relevant to traits (e.g., Flavell & Miller, 199 8; Miller & Aloise, 1989; Ruble & Dweck, 1995; Yuill, 1992). Research showed that it was only the older group (9 and 10 -year -olds) who could make a correct selection of behaviors consistent with traits. The second form of inference ability utilized in stu dies is trait to -behavior inferences (predicting specific patterns of behaviors from traits explanation). Literature shows th at 4 and 5 -year -olds also had difficulty predicting trait -consistent behaviors from traits (e.g., Liu et al. 2007). The last form of inference ability that has been employed in studies is behavior to trait inference (labeling traits from trait relevant behaviors). Heyman and Gelman (1999) and Liu et al. (2007) demonstrated that 4 -year olds c an infer traits (nice, mean, and shy) from specific behaviors described. Among the three types of inferences, the Trait measure created for this study required children to use the behavior -to -trait inference Because the Trait measures of this study incorporated the experimental procedure utilized in the Object form, it would be interesting to compare childrens performance on the new measures with performance on the traditional trait measures.
16 Mastery of Beliefs Incorporating these two T rait tasks as well as the unexpected transfer trial, the present study was specifically interested in childrens mastery of beliefs The unexpected content trial w as ex cluded from this study due to the procedural disparity fr om the unexpected transfer form All the belief stories presented to children were story book formats. The comparisons between the T rait and the unexpected transfer measures were expect ed to provide information with regard to possible variations in belief mastery. For the comparisons o f performance on object and T rait trials two developmental com ponents of beliefs w ere investigated. The first component concerns age of mastery. This study was interested in examining exactly when children become proficient in the two different types of b eliefs and if there are differences in development as hypothesized in the previous section. The literature demonstrated a robust developmental pattern of judgments on object beliefs during the preschool period. Justifications in children older than 5, howe ver, have been studied in less research Furthermore, there have been no such belief studies to date focusing on preschool childrens false beliefs about underlying traits of people There were also no voluntary movement studies conducted to examine level of performance in children beyon d the age of 5. This study provide d a within -subject investigation of both the object and the trait performance in children with an age range betw een 4 and 7. Adding relatively older children such as 6 and 7 -year -olds to this study was expected to provide a fuller version o f developmental knowledge with regard to both object and person beliefs. T he second important component was criterion for mastery. A justification procedure has not been a common tool to e xplore development of false beliefs in the literature Only a handful of object studie s have provided a justification procedure after the false belief question (Clements & Perner, 1994; Clements, Rustin, & McCallum, 2000; Parker, MacDonald, & Miller, 2007;
17 Ruffman, Slade, & Crowe, 2002; Ruffman, Slade, Rowlandson, Rumsey, & Garnham, 2003; Wimmer & Weichbold, 1994) In the person literature, there was one voluntary-movement study in which only J apanese children participated (Naito & Koyama, 2006) and no pers on study has measured Western childrens explanations. This study employ ed tw o explanation systems to examine childrens justifi cations for belief responses: factual information and added informati on The factual information category has been commonly obs erved i n the previous studies in which situational and mentalistic justifications are investigat ed The situational explanation concerns the search of the protagonist with situational cues presented in the story but it is not directed to the current mental state of the protagonist. A common example of this explanation shown in the unex pected transfer trial would be because thats where she left it. Mentalistic explanation, in contrast, concern s childrens direct mentioning of mental state terms for exampl e, because thats where she thinks it is. Along with the factual information system, a new type of justification, the add ed information category, w as utilized in particular with regard to the Trait trials. The added information was to investigate wheth er children could reason about the target ed trait s beyond the factual level by using knowledge or reasoning tendenc ies they previously possess ed In the case of the T rait incongruent condition, the story itself does not provide sufficient or conclusive information about a persons trait. This m ight result in childrens use of knowledge reflecting their preexisting mindsets. In summary, this study examined development of different kinds of beliefs (self vs. other belief and true vs. false belief) on the Trait measure in children aged from 4 to 7, in addition to the development of object false belief through judgment and justification procedures. Consistent with the lite rature, children s understanding of the object false belief was expected to improve
18 during the preschool years. In contrast, due to the nature of different kinds of beliefs as previously noted, it was expected that there would be some developmental variati ons in performance on the Trait measure. Further expectations were that children would perform equivalently on true belief and false belief and that self understanding would be lower than other understanding. Types of justifications on the Trait measure ac ross the different age groups were also expected to provide evidence of qualitative change in belief development.
19 CHAPTER 2 METHOD Participants There were n ineteen 4 -year olds (mean age: 4 years 5 months, 10 girls and 9 boys), twenty one 5 -ye ar olds (mean age: 5 years 6 months, 9 girls and 12 boys) twenty -seven 6 year olds (mean age: 6 years 4 months, 15 girls and 12 boys), and twelve 7 -year -olds (mean age: 7 years 2 months, 6 girls and 6 boys) recruited from preschool or elementary -school se ttings of a small university town in the southeastern area of the United States. The large majority of the sample was W hite, and three Asian, five Black, eight Hispanic and four middle -eastern children made up the rest of the sample. Design This study was a within -subjects design with in which each child hear d six diffe rent belief scenarios. Two were the object and the o ther four were the trait s cenarios The trait scenarios contain ed two conditions, the T rait congruent and the T rait incongruent condit io n. Performance on the object, the Trait -congruent, and the Trait incongruent condition was c ompared across different age groups to examine the effect of age and type of belief problems. Materials All of the scenarios were constructed on the frame of the u nexpected transfer form with slight variations. The unexpected transfer form in general presents two main characters in the story in which the status of the target (usually an object) for one of the main characters is changed. The unexpected transfer story is usually provided with a sequence of three important scenes: 1) two characters with the target, 2) one character placing the target at a place, and 3) the other main character relocating the target to another place in the absence of the first character. One of the object trial s (one question version) in this study present ed the identical sequence
20 shown in the unexpected transfe r form. Children therefore hear d a story in a three page booklet: 1) two characters with an object (e.g., Michael showing a book to Jenny), 2) one character placing the object in one place (e.g., Michael leaving the book by the bench), and 3) the other character relocating the object to another place (e.g., Jenny placing the book by the rock). The test question was directed to the protagonist s belief about the object (where will Michael look for the book?). The other type of object trial (two questions version) had the same sequence as the unexpected transfer form but with the two test questions which were directed to both of main characters presented in the story (e.g., where will Michael look for the book? a nd where will Jenny look for the book?). The two question version object trial was to provide a similar kind of testing procedure to the Trait trials which will be specifically explained in the next section. The Trait trials also based upon the unexpected transfer form differ ed from the object trial in two aspects. The first aspect concern ed differences in targets. The targ ets of the Trait trials w ere a persons behaviors utilized to infer a trait of the person. Previous studies investigating childrens trait understanding have suggested that young children can understand and predict basic traits such as good, nice, bad, naughty, shy, and selfish (e.g., Bretherton & Beeghly 1982; Cain et al., 1997; Heyman & Gelman, 1999; Liu et al., 2007). Kind/not kind, nice/mean, good/bad, and clean/messy were selected as initial forms of underlying traits for the stories of this study Each story re present ed a set of behaviors that c hildren commonly experience in their everyday life such as sharing crayons with friends, saying hi to people, playing with a dog, and picking up toys. The pilot data conducted to assess the developmental appropriateness of the specific traits selected however indicated that some young children, in particular 4 year -olds, prefer to use superficial terms. They judged the behaviors targeted to clean/messy or kind/not kind with the terms either nice/mean or good/bad. The final forms of the
21 traits w ere therefore nic e/mean and good/bad These traits were selected with consideration of developmental patterns across all ages of children but children s spontaneous answers relevant to the target behaviors w ere encouraged and recorded during the procedure. As previously describe d, the trait behaviors w ere displayed in two different conditions, the T rait -congruent and the T rait incongruent conditions. In the T rait -con gruent trial, children heard a story with three scenes: 1) A main character sees a behavior of a person th at represents the persons trait (e.g., Amy knows Tom well and sees that Tom is sharing crayons with friends everyday), 2) a n unexpected situation leading the person to behave in the opposite way (e.g., Tom cannot share crayons due to the class task that h e has to finish), and 3) t he othe r main character who does not know about the person sees the behavior opposite to his true characteristics (e.g., Elizabeth sees Tom not sharing crayons). The target of the T rait -incongruent trial was als o a behavior of a person, but the trial varie d in the number of scene s present ed. There were o nly two scenes shown to children on the T raitin cong ruent trial without the presence of the middle scene of the T rait -congruent condition. Children, for example, hear d a scenario with two page illustrations: 1) A main character sees a behavior of a person (e.g., Amy sees Tom sharing crayons with friends) and 2) t he other main character sees the opposite behavior of the person (e.g., Elizabeth sees Tom not sharing crayon s with friends). All the stories were presented in a story book form by 8 by 11 inch colored illustrations and the order of the stories was counterbalanced across children. In particular two sets, each of which consisted of three belief stories, were created. The two sets and three belief trials (object, T rai t -congruent, T rait incongruent) in each set were balanced. Such counterbalanc ing created 12 possible orders of six belief trials and these orders were randomly assigned to children.
22 A s described in the previous section, two object and four trait stories were created for the two sets of six false belief trials. The main structure of the se scenarios was identical across the measures except for slight changes in wording in the object trials and the cha nge of the middle scene in the Trait trials In addition, t he sequence of the two scenes in one of the T rait incongruent conditions was also rev ersed in the alternative set to remove a possible order effect resulting from a specific trait presented in a fi xed order The Appendix provides an example of the object and trait scenarios The condition of each trial within the object and the Trait measure was balanced across children. For example the book scenario ( t he one question version of the Object trial) b ecame the two -question version with inclusion of another test question. The sharing crayons scenario be ca me the T rait incongruent trial with the absence of the middle part and the saying hi scenario became the T rait-congruent trial s with the explanation part in the middle of the story. Procedure Children were tested individually in a room in their schools by two female research assistants. One of the research assistants managed testing (presenting stories and asking questions) and the other was responsib le for recording answers provided by the children. After presenting each story, one of the research assistants ask ed children two control questions directed to the current reality of the story (e.g., for an object measure, Where is Jennys book now? and where did Jenny leave the book? for a person measure, what did Elizabeth see Amy do? and what did Tom s ee Amy do?). If the child fail ed to respond correctly to either of the reality questions, the criti cal part of the scenario was reviewed and the qu estion w as asked again. If the child fail ed to answer correctly up to three times the trial was o mitted. After the child succeed ed on both of the reality qu estions, the test questions follow ed The initia l form of test question w as open -ended format (e.g., Where will Michael look for the book? or What does Tom think
23 about Amy?); if there was no response, the child ha d a forced -choice fo llow up with which he or she could make a choice for the answer (e.g., Will Michael look for the book by the bench or by the rock? or Does Tom think Amy is nice or mean?). T he justification questions follow ed each test question in order to examine childrens explanations for their belief judgments (e.g., why do you think Michael will look there? or Why do you think Tom thinks Amy is nice?).
24 CHAPTER 3 RESULTS The central in terests of data analyses were in the compariso ns among childrens performance on different types of belief measures with a developmental e mphasis Repeated measures for dichotomous or ordinal data in particular generalized estimating equation (GEE) analyses (Liang & Zeger,1986; Zeger & Liang, 1986) for the variables of age and belief type were performed throughout the analyses due to two aspects of the data repeatedly measured false belief tr ials and a dichotomous or an ordinal value at each level of false belief. T he GEE analyses were performed using PROC GENMOD in the SAS package, with binomial distribution, logit link, and an exchangeable correlation structure for the dichotomous data, and with multinomial distribution, cumlogit link, and an independent correction structure for the ordinal data. The a ge by type interaction term was also added to each analysis in order to examine the possible interaction between children s age and type of pro blem. This study contained two different kinds of analyses. T he first one was directed to the accuracy of the belief judgment s Children who provided the correct judgment to the test question(s) on each trial (one test question was given on the one -questi on Object version and two test questions were given on the other five belief trials) were considered successful on the trial. If they failed any of test question(s) children did not receive any credit. Such scoring was also applied to childrens judgment s on the self questions. Answers for the self judgments were one of three types ( e.g., nice, mean, and both) A positive trait (nice or good) for the Trait -congruent condition and both positive and negative traits for the Trait incongruent condition were co nsidered as the accurate answers for the self questions. The second analysis examined children s justifications for their belief judgments. Children s adequate explanations were quantified for statistical analyses. In particular, children
25 e arn ed 1 point each for each correct judgment and adequate justification for one trial (0: incorrect judgment, 1: correct judgment with inadequate justification, 2: correct judgment with adequate justification). The s tyle of explanations children used to justif y their judgments was further analyzed in the next step. The style was coded in to categories of factual and added information. Those in the factual type were further categorized into situational and mentalistic explanations S ituational justifications in p articular were for children who answer ed with circumstantial information (e.g., because she left it there, because he left the legos on the floor ). If children mention ed terms directed to protagonists mental states such as think, feel, know, or want ( e.g., because he thinks it is there), the answers w ere categorized into the factual explanation category as mentalistic type. In contrast to the justifications relevant to factual information, justifications categorized into the added information categor y were those using information beyond the story facts presented on the trials. If children s justification s concerned story outlines presented (e.g., he is good because he pick ed up the toys) they were categorized into the factual information type. If the justifications were not based on the stories (e.g., he couldn t pick up the toys because he had a bad day) they were categorized into the added information. Four raters were involved in coding of children s justification of beliefs. The initial coding w as completed independently by t wo raters who were blind to the hypotheses of this study and children s age information. Interrater reliability between the two raters was 93% (110 cases of disagreement over a total of 1603 ratings). Conflict ing ratings were f urther reviewed and resolved by the other two raters who were also blind to children s age information. Analysis of Judgment As previously described, children s spontaneous responses to test and self questions on the Trait measures were recorded. Aside from the conventional traits (nice, mean, good, bad)
26 given in the belief questions, children provided various terms directed to the mental states of the main char acters described in the stories. Four -year -olds appeared to use terms less relevant for describing the specific behaviors presented ( beautiful, happy, and gentle ), although they did use the term messy which is a specific trait term. Although childr en older than 4 also mentioned sad and fun these cases were rare. Older children were more likely to focus on specific trait terms appropriate for characterizing main characters behaviors such as polite, clean, messy, respectable, rough, ru de, and responsible C hildren received the forced -choice question to further clarify some of the terms from which positive or negative traits could not be inferred. On 45 trials out of a total of 474 belief trials, children failed to provide the correct answers to reality questions. Overall younger children made more errors than older children on the Trait trials compared to object trials. In particular, 24 cases were from 4 -year -olds ( 5 on the Object trials, 8 on the Trait incongruent trials, and 11 on the Trait congruent trials), 12 were from 5 -year -olds ( 4 on the Object trials, 5 on the Trait incongruent trials, and 3 on the Trait congruent trials), and 9 were from 6-year -olds ( 2 on the Trait-incongruent trials and 7 on the Trait-congruent trials). These missing trials were omitted from data analyses. A preliminary analysis using the GEE method was performed to e xamine the effect of gender and task order on children s overall false belief performance on the three tasks. Result revealed no significant effect in both cases. The gender and task order variables were therefore omitted from further analyses. Judgment of Other Belief The first analysis was directed to partial correlatio ns among children s performance on the bel ief measures (the Object, the Trait -congruent, and the Tr ait-incongruent trials) with age controlled. There were significant cor relations between the O bject and the Trait congruent ( r
27 (54) = .33, p = .013) and between the Trait -congruent and the Trait incongruent ( r (53) = .29, p = .031). The GEE analysis examined possible relations between childrens age and their accuracy of judgments in different types of belief (the Object, the Trait -congruent, and the Trait incongruent trials ) Descriptive in formation is presented in Table 3 1. There was a significant main effect of age on belief performance, x (3) = 15.10, p = .001. Further comparisons of least square s means showed that 4 -year -olds performance ( M = .40) was significantly lower than that of t he other age groups ( M s = .79, .78, and .84 for 5 6 -, and 7 -year olds respectively), x s > 16.27, p s < .001. Table 3 1 Proportion of c orrect r esponses to b elief j udgments on t est q uestions Object Trait Congruent Trait Incongruent Total 4 year olds 5 year -olds 6 year -olds 7 year -olds Total 30 76 83 87 70 41 77 82 83 72 50 86 69 83 72 40 79 78 84 N either the effect of type of trial nor the age by type interaction emerged from the analysis. Such a result suggests that there were no differences in belief understanding across the three types of measures within each age group of children. Further analyses comparing performance on belief trials within each age group supported this finding. Each age group of children showed equivalent performance across the three measure s. A nalys es of the six trials individually also showed analogous performance within each group. These results suggest that there are no differences between object and trait beliefs and also no differences between true an d false beliefs within the Trait measure. Further analysis was directed to children s false belief understanding on the Trait congruent measure. Again, c hildren s false belief performance on the Trait -congruent measure
28 provided a similar pattern to that o n the Object measure as shown in Table 3 1 Performance of 4 year -olds ( M = 55) on understanding of other people s false beliefs on the Trait -congruent measure was lower than other age groups ( M s = 87, .82, and .8 7 for 5 -, 6 and 7 -year -olds respectively ). Judgment of Self Belief Children s responses to the self questions indicated that there were a total of 10 cases in which children provided no judgments to both the self question and the forced -choice follow up. Younger children, in particular 4 and 5 -year olds, in general compared to the older age groups showed such a tendency. Specifically, 2 cases were from 4 -year olds ( 1 on the Trait congruent and 1 on the Trait incongruent), 6 cases were from 5 year -olds ( 3 on the Trait congruent and 3 on the Trait incongruent), 1 case on the Trait incongruent was from 6-year -olds, and 1 case on the Trait -congruent was from 7 year -old s These cases were omitted from the analyses performed. Table 3 2 Proportion of r esponses to b elief j udgments on s elf q uestions TraitCongruent TraitIncongruent 4 year -olds 5 year -olds 6 year -olds 7 year olds Positive 57 42 65 65 Both 4 25 24 13 Negative 38 33 11 22 P ositiv e 66 56 48 52 Both 3 15 29 22 Negative 31 29 23 26 Note. Proportion of accurate responses is in italics Table 3 2 presents the proportion of responses to self judgments as a function of age and belief type (the Trait -congruent and the Trait incongruent). An analysis performed to examine self judgments confirme d the main effect of task, x (1) = 21.42, p = .001, and age, x (3) =13.57, p = .001. Children were more accurate on the Trait -congruent measure ( M = .57 ) than the Trait incongruent ( M = .18 ). Furthermore, the main effect of age emerged. The differences in least
29 square s means showed that 4 year olds ( M = .29) and 5-year olds ( M = 28) had less success compared to 6 year -olds ( M = .46) and 7 year -olds ( M = .43), x s > 4.31, p s < .036. Another analysis was performed to examine children s correct self responses accompanied by correct false belief respons es on the Trait congruent task. Findings showed a similar pattern to the proportion of children s judgments on self belief provided in Table 3 2. The re was a significant age main effect (x (3 ) = 16. 55, p = .001). Comparing least squares means further revealed that 6 (M = .59) and 7 -year olds ( M = .65) performed better than 4 ( M = 36) and 5 -year olds ( M = 34). Another interesting developmental tendency observed from children s self responses was the age difference in the pattern of judgment. As Table 3 2 indicate s older children were more likely to take into account two perspectives simultaneously compared to the 4year -olds. They tended to answer both positive and negative rather than responding with single perspective (either positive or negative) on the two Trait measures An analysis was performed to confirm this age related characteristic. The result showed that there was age related improvement in the use of multiple perspect ives, x (3) = 9.08, p = .028. From comparisons of least square s means the percentage to simultaneously use two perspectives for overall self judgments in 4-year olds (3%) was significantly lower than the other age groups (5 -year olds = 20 %, 6 year -olds = 26%, 7 year -olds = 17%), x s > 4.04, p s < .044. Analysis of Justification Justification of Other Belief For the analysis of justification, the next sections will be devoted to childrens explanations for the target's belief first, with explanations for the self judgments to follow. Children provided adequate explanations for 95% of their overall correct belief judgments. Such a high rate reflected the fact that the older age group s succeeded on the trials more often than the
30 4 year -olds. In particular, 4 year -olds had 9 cases of inadequate explanations (3 on the Object, 2 on the Trait incongruent, 4 on the Trait -congruent), 5 year -olds had 4 (2 on the Trait incongruent and 2 on the Trait -congruent), 6 year -olds had 3 (1 on the Object and 2 on the Trait co ngruent), and 7 -year olds had 2 (1 on the Trait incongruent and 1 on the Trait -congruent). Table 3 3 summarizes the proportion of children s correct judgments accompanied by adequate justifications as a function of age and belief task. Although the age ef fect fell short of significance, the tendency to provide adequate explanations appeared to increase with age. Older age groups (5 -year -olds: 95%, 6 year -olds: 97%, and 7 -year olds: 97%) in general compared to the 4 year -old s (75%) made fewer errors and a l arger proportion of children in these groups relevantly justified their belief judgments. Table 3 3. Proportion of c orrect j udgments accompanied by a dequate e xplanations Object TraitCongruent TraitIncongruent Total 4 year olds 5 year -olds 6 year -olds 7 year -olds Total 70 100 97 100 96 63 93 94 95 91 86 93 100 95 95 75 95 97 97 When an analysis was performed to examine childrens judgment plus -explanation performance with the quantified scores (0: incorrect judgment, 1: correct judgment with inadequate justification, 2: correct judgment with adequate justification), the age effect was clear. The analysis showed a n age main effect, x (3) = 17.76, p = .002. Further analyses of this effect revealed that older children ( M s = .78, .77, and .83 for 5-year olds, 6 -year -olds, and 7 year -olds respectively) outperformed 4 -year olds ( M = .35), x s > .13.21, p s <.001. As indicated earlier, adequate justifications made for correct judgments to test questions were further coded as factual (situational or mentalistic) and added information. Most of the cases fell into the category of situational explanation, making up 97% of the overall adequat e
31 justifications. Few cases were the mentalistic and the added information type, but the tendency to use those types still appeared to improve with age. In particular, 4 year -olds made only one added information explanation on the Trait incongruent measur e. In contrast, 1 case of mentalistic explanation on the Object measure and 4 cases of added-information explanations (2 on the Trait incongruent measure and 2 on the Trait -congruent measure) were from 5 -year old children. Six -year -olds and 7 -year -olds als o showed a similar explanation style to the 5 -year olds. Two mentalistic explanations (1 on the Object and 1 on the Trait -congruent measure) and 5 cases of added-information explanations (3 on the Trait congruent and 2 on the Trait incongruent) were from 6 year -olds. Finally 7 year -olds made 3 mentalistic explanations (1 on the Object and 2 on the Trait -incongruent) and 2 added information explanations (1 on the Object and 1 on the Trait incongruent). Justification of Self Belief More variations in explan ation style were observed from children s responses to self justifications. Children overall provided added information explanations for 13% of their self judgments. The rest fell into the situational type, and no mentalistic explanation s were made to expl ain self judgments. The proportion of different styles utilized in different age groups of children is summarized in Table 3 4. Again, there was age -related increment in use of the added information style. Three cases (2 on the Trait congruent measure and 1 on the Trait incongruent) were from 4 -year -olds, 2 cases ( on the Trait incongruent) were from 5 year -olds, 4 cases (2 on the Trait -congruent and 2 on the Trait incongruent) were from 6 -year olds, and 8 cases (3 on the Trait-congruent and 5 on the Trait -i ncongruent) were from 7 -year olds.
32 Table 3 4. Proportion of t ype of s elf j ustifications TraitCongruent TraitIncongruent 4 year -olds 5 year -olds 6 year -olds 7 year -olds Situational 98 100 95 86 Mentalistic 0 0 0 0 Added 2 0 5 14 Situational 96 94 96 79 Mentalistic 0 0 0 0 Added 4 6 4 21 An analysis indicated that the age and task effect on style of self justifications fell short of significance, yet there was a trend toward an age main effect (Table 3 4). Differences in percentage across age groups demonstrated that the employment of added information in belief judgment was more usual in older children, markedly 7 year -olds (17%), in general compared to 4 year -olds (6%), 5 -year olds (3%), and 6 -year -old s (4 %).
33 CHAPTER 4 DISCUSSION This study provides the first clear evidence on how young children understand beliefs a bout people s trait s, one of the beliefs constructed about the social world. The Trait measures utilized in this study had several differences from the previous false belief measures. One difference concerns the nature of the targets presented. In the previous work, the target was a physical location of an object or a person; in this study it was a trait of a person. The Trait measures also differed in that they included a self belief question In the previous studies, the Object transfer form or the other social false belief measures did not provide a self question bu t the Trait measure in this study did. Finally in the previous studies the focus was mainly on children s false belief understanding; this study focused initially on a false belief but also extended it to a true and a relative belief. Among 12 questions as ked on the Trait measures o ne of the two test questions on each of the Trait -congruent trial s was directed to false belief understanding. The other questions (one of the two test questions and the self question on each of the Trait measures) concerned tru e beliefs. One of the central research questions in previous work on social false belief was whether children understand social beliefs equivalently to object beliefs. Most researcher s reported that recognizing social belief s is more difficul t than recogn izing object beliefs in Western children (Rai & Mitchell, 2004; Symons et al., 1997). This study did not replicate this outcome. The difference probably results from methodological variations between measures utilized. As described earlier, the characteris tics of the targets in the social false belief measures utilized in previous work and the Trait measures utilized in this study differ. Another explanation for the different results is that the social false belief versions utilized in the Symons and Rai and Mitchell studies have some methodological disparity from the standard Object transfer form as
34 pointed out in the study by Naito and Koyama (2006). Ahn and Miller (2009) demonstrated that the use of the less biased social measure produced an equivalent le vel of performance in a Western sample. Whatever the reasons for the underachievement in sever al Western studies might be, internality does not appear to be the central issue for the lower level of performance. A trait, the target of the Trait measure utilized in this study has more resemblance to internal characteristics of people, compared to a physical location utilized in previous literature. Examination of the existing trait literature also suggests a similar point. F indings indicated that children s understanding of traits from a given context has a similar progression to understanding of false belief. Children s ability to infer a person s internal characteristics develop s around 4 and 5 years of age (e.g., Heyman & Gelman 1999; Liu et al. 2007) the same period when children start to attribute false beliefs to others. A further comparison of findings in the trait literature and the current study also raise s several interesting issue s The trait literature provides an experimental context in which a single perspective (a perceiver or self) is involved. In contrast, the context presented in the Trait false belief measure requires understanding three different perspectives (two perceivers and self) about a person s trait. In general, children s unders tanding of traits and their understanding of beliefs about traits appear to emerge in a similar manner. Similarly to Heyman and Gelman (1999) and Liu et al (2007), this study showed that children s understanding of a be lief about a person s trait develop s during the preschool year s However, using more than two traits simultaneously appears to be more difficult than recognizing one trait. In this study the correct self response on the Trait incongruent measure was to mention both positive and negative trai ts presented in the story. Although there was age related improvement between 4 and 5 -year olds,
35 children across all age groups produced relatively lower rate s of success compared to the other types of false belief measure. In fact, self judgments produ ced overall lower performance compared to other judgments. A general claim in false belief research is that there is no difference in recognizing beliefs of self and other. T his study however showed that children were less likely to succeed at understanding beliefs of self than others. This discrepancy appears to stem from a number of new characteristics the Trait measure had as previously described. The self question utilized in the Trait measure concerns a true belief. Furthermore, the procedure provided in the Trait measure was different from that in the Object measure In the Object content form, the self and other questions have the same right answer. In the Trait measures the self questions require an answer that is never required for the questions about the others. In this way, answering the self questions may requi re an extra reasoning process. In the Object measure, the only process required is to recognize the two beliefs of the main characters. The Trait measure also requires the same mental process to correctly identify the beliefs of the two perceivers. F or sel f judgments provided after presentation of the other judgments another reasoning process such as remembering and integrating the specific situations provided seem s to be necessary. It is also worth noting that there were some consistent findings with previous work on false belief. One of the consistent findings is mastery of false belief development across the various age groups. The previous work on false belief judgment, especially the Object transfer form, has consistently reported that children s false belief understanding improves between the ages of 4 and 5 (Miller, 2000; Wellman et al 2001). T his study provided the equivalent finding, suggesting that false belief development emerges between the ages of 4 and 5. Such mastery was confirmed from the j ustifications data Children older than 4 compared to the 4 -year -olds are
36 more successful at providing an adequate explanation for their correct judgments made on the Object false belief A similar developmental progression for judgment plus -justification was also found in previous studies ( Clements & Perner, 1994; Clements et al., 2000; Parker et al., 2007; Ruffman et al., 2002; Ruffman et al., 2003; Wimmer & Weichbold, 1994) A further result from performance on the Object and the Trait measure provides useful information on specific age differences in belief development. The majority of previous studies have provided developmental evidence for false beliefs in 4 and 5 year -olds. This study examined children from the age of 4 to 7. Most children older tha n 4 years performed well on the Object measure, but the interesting result was from varia tions in performance on the Trait measures. In the Trait measure, children in general performed well on beliefs of others but not on beliefs of self. The dramatic impr ovement in recognizing those beliefs occurred during the preschool year s but success on the self judgments remain ed lower compared to other judgments. Results of a dequate justifications accompanied by correct judgments also confirmed such a developmental tendency. Overall the acceleration of belief understanding, whether it is the Object or the Trait measure, or a true or a false belief occurs between the ages of 4 and 5. Although developmental changes occur between the ages of 4 and 5, the comparisons of older age groups added to the value of this study. Children s understanding of different beliefs continuously improved with skills for correct judgments and adequate explanations. Around age 7, children seemingly start to think more creatively to explai n beliefs they judge. This age difference suggests a possibly new and interesting finding in 7 -year olds. One of the new aspects of this study was to incorporate a new coding system, namely the added information type. This study shows that children s use of addedinformation explanations may start dramatically improving at age 7. This false belief variation suggests that it is after mastering belief
37 understanding that children may start actively applying their personal level of knowledge to the questions a sked especially when the information provided is not adequate (as is true on the Trait incongruent trial) The overarching purpose of this study was to explore variations in children s understanding of beliefs about objects and traits by using a newly cr eated social false belief task, namely the Trait -congruent and the Trait -incongruent measure. Although some r equire further examination most findings of this study extend previous demonstrations in belief research. This study confirmed that an important developmental stage for false belief understanding was the preschool year s Specific findings further suggest possible variations with regard to different belief forms, object and trait beliefs, true and false beliefs, and self and other beliefs across diff erent ages.
38 A PPENDIX BELIEF SCENARIOS One QuestionVersion Object Condition Michael is showing his friend Jenny the new book that he got at the store. They are having so much fun! Then Michaels mom calls him into the house. So Michael puts the book down by the bench and goes into the house. While Michael is inside, Jenny looks at the book, but then she has to go home. As she leaves, she puts the book down by the rock. When Michael is finished talking with his mom he comes back o ut to look at his book. 1st reality question: Where is the book now? 2nd r eality question: Where did Michael last see the book? Test question: Where will Michael look for the book? Forced choice question: Will Michael look by the bench or will he look by the rock? Justification question: Why will he look there? Two Question Version Object Condition Susie is playing with her doll in her backyard. Its her favorite doll and she is having lots of fun! Susie plays with her doll for a long time. Then she gets u p to go to the bathroom. Before she goes inside she leaves her doll in the sandbox. While Susie is inside, her sister Ashley comes outside and sees the doll. She picks it up and plays with it. Then their mother calls Ashley to come in. Ashley sets the doll by the tree. 1st reality q uestion: Where did Ashley leave the doll? 2nd r eality q uestion: Where did Susie leave the doll? 1st t est q uestion: Where will Susie look for the doll? Forced c hoice question : Will she look in the sandbox or by the tree? Justification: Why will she look there?
39 2nd t est q uestion: Where will Ashley look for the doll? Forced c hoice question : Will she look in the sandbox or by the tree? Justification: Why will she look there? 1st Trait -Congruent Condition This is Tom. Tom and Amy are best friends. Tom always sees Amy sharing her crayons with other friends. Every day Tom sees Amy sharing her crayons. Her friends look happy when Amy shares them. However, one day, Amy cannot share her crayons with her friend because her teacher to ld her to finish her picture. This is Elizabeth. Elizabeth does not know about Amy because she just started school. While Elizabeth is looking for friends to play with, she sees Amy not sharing crayons with her friend. Her friend looks sad. 1st reality q ue stion: What did Elizabeth see Amy do? 2nd r eality q uestion: What did Tom see Amy do? 1st t est q uestion: What does Tom think about Amy? Forced c hoice q uestion: Does he think Amy is nice or mean ? Justification question: Why does he think that Amy is nice / mea n ? 2nd t est q uestion: What does Elizabeth think about Amy? Forced c hoice q uestion: Does she think Amy is nice or mean ? Justification question: Why does she think that Amy is nice / mean ? Self q uestion: What do you think about Amy? Forced c hoice q uestion: Do you think Amy is nice or mean ? Justification question: Why do you think that Amy is nice / mean ? 2nd Trait -Congruent Condition This is Karen. Karen and Brian are best friends. Karen always sees Brian cleaning up his toys at school. Every day she sees Brian picking up his toys. Their classroom looks clean when Brian
40 picks them up. However, one day, Brian cannot clean up his toys because his mom comes to pick him up early. Brian has to hurry and cannot clean up his legos. This is Mark. Mark does not know about Brian because he just started school. Mark walks into the classroom to play with the legos. When Mark gets to the classroom, he sees Brian leaving his toys on the floor. The room looks messy. 1st reality q uestion: What did Mark see Brian do? 2nd reality q uestion: What did Karen see Brian do? 1st test q uestion: What does Karen think about Brian? Forced choice q uestion: Does she think Brian is good or bad? Justification: Why does she think that Brian is good/bad? 2nd test q uestion: What does Mark think about Brian? Forced choice q uestion: Does he think Brian in good or bad? Justification: Why does he think that Brian is good/bad? Self q uestion: What do you think about Brian? Forced choice q uestion: Do you think Brian is good or bad? Justification: Why do you think that Brian is good/bad ? Trait -Incongruent Condition This is Paul. When Paul is on a walk one day, he sees Sarah saying hi to her friend. Her frien d looks happy. This is Jessica. The next day when Jessica is walking down the street, she se es Sarah walking by her friend without saying hi. Her friend looks sad. 1st reality q uestion: What did Jessica see Sarah do? 2nd r eality q uestion: What did Paul see Sarah do? 1st t est q uestion: What does Paul think about Sarah? Forced c hoice q uestion: Does he think Sarah is nice or mean?
41 Justification question: Why does he think that Sarah is nice/mean? 2nd t est q uestion: What does Jessica think about Sarah? Forced c hoice q uestion: Does she think Sarah is nice or mean? Justification question: Why does she think that Sarah is nice/mean? Self q uestion: What do you think about Sarah? Forced c hoice q uestion: Do you think Sarah is nice or mean? Justification question: Why do you think that Sarah is nice/mean? Reversed Trait -Incongruent Condition This is Billy. One day, when Billy is looking for his friend outside, he sees Mr. Miller yelling at his dog on the side of the road. His dog looks sad. This is Rachel. The next day while Rachel is walking to her house, she sees M r. Miller playing with his dog. His dog lo oks happy. 1st reality q uestion: What did Rachel see Mr. Miller do? 2nd r eality q uestion: What did Billy see Mr. Miller do? 1st t est q uestion: What does Billy think about Mr. Miller? Forced c hoice q uestion: Does he think Mr. Miller is good or bad? Justifi cation question: Why does he think that Mr. Miller is good/bad? 2nd t est q uestion: What does Rachel think about Mr. Miller? Forced c hoice q uestion: Does she think Mr. Miller is good or bad? Justification question: Why does she think that Mr. Miller is goo d/bad? Self q uestion: What do you think about Mr. Miller? Forced c hoice q uestion: Do you think Mr. Miller is good or bad? Justification question: Why do you think that Mr. Miller is good/bad?
42 LIST OF REFERENCES Ahn, S., & Miller, S. A. (2009 ). Theory of mind and self -concepts: A comparison of Korean and American children Manuscript submitted for publication. Barenboim, C. (1981). The development of person perception in childhood and adolescence: From behavioral comparisons to psychological constructs to psychological comparisons. Child Development 52, 129 144. Bartsch, K. (2002). The role of experience in childrens developing folk epistemology: Review and analysis from the theory theory perspective. New Ideas in Psychology, 20, 145161. Bretherton, L. & Beeghly, M. (1982). Talking about internal states: The acquisition of an explicit theory of mind. De velopmental P sychology 18, 906 921. Cain, K. M., Heyman, G. D., & Walker, M. E. (1997). Preschoolers ability to make dispositional predictions within and across domains. Social D evelopment 6 53 75. Clements, W. A., & Perner, J. (1994). Implicit understanding of belief. Cognitive Development, 9, 377395. Clements, W. A., Rustin, C. L., & McCallum, S. (2000). Promoting the transition from implicit to explicit understanding: A training study of false belief. Developmental Science, 3, 8192. Flavell, J. H. (1999). Cognitive development: children's knowledge about the mind. Annual Rev iew of Psychol ogy 50, 21 45. Flavell J. H., & Miller, P. H. (1998). So cial cognition. In D. Kuhn & R. S. Siegler (Eds ), Handbook of child psychology: Vol. 2. Cognition, perception, and language (pp. 851898). New York: Willey. Goldman, A. I. (1992). In defense of the simulation theory. Mind and Language 7 104119. Gop nik, A., & Astington, J. W. (1988). Children's understanding of representational change and its relation to the understanding of false belief and the appearance reality distinction. Child Development 59, 26 37. Harris, P. L. (1992). From simulation to folk psychology: The case for development. Mind and Language 7 120144. Heyman, G. D., & Gelman, S. A. (1999). The use of trait labels on making psychological inferences. Child Development, 70, 604 619. Liang, K. Y. & Zeger, S. L. (1986) Longitudinal data analysis using generalized linear models Biometrika 73, 1322.
43 Liu, D., Gelman, S. A., & Wellman, H. M. (2007) Components of young childrens trait understanding: Behavior to trait inferences and trait to -behavior predictions. Child Development, 78, 1543 1558. Lucariello, J., Donne, M. L., Durand, T., & Yarnell, L. ( 2006). Social and intrapersonal theories of mind: I interact therefore I am. In A. Antonietti, O. L. Sempio, & A. Marchetti (Eds.), Theory of m ind and l anguage in d evelopmental c ontexts (pp. 149172). New York: Springer. Miller P. H., & Aloise, R. R. (1989). Young childrens understanding of the psychological causes of behavior: A review. Child Development 60, 257285. Miller, S. A. (2000). Childrens understanding of preexisting differences in knowledge and belief. Developmental Review, 20 227 282. Naito, M., & Koyama, K. (2006). The development of false -belief understanding in Japanese children: Delay and difference? International Journal of Behavioral Development 30, 290 304. Nguyen, L., & Frye, D. (1999). Children's Theory of Mind: Understanding of Desire, Belief and Emotion with Social Referents. Social Development, 8, 70 92. Parker, J. R., MacDonald, C. A., & Miller, S. A. (2007). Jo hn Thinks That Mary Feels False belief across the physical and affective domains. Journal of Genetic Psychology, 168, 4362. Rai, R., & Mitchell, P. (2004). Five -year old childrens difficulty with false belief when the sought entity is a person. Journal of Experimental Child Psychology, 89, 112 126. Rholes, W. S., & Ruble, D. N. (1984). Childrens understanding of dispositional characteristics of others. Child Development, 55 550 560. Ruble D. N., & Dweck, C. S. (1995). Self -conceptions, person conceptions, and their development. In N. Eisenberg (Ed.), Review of personality and social psychology: Vol. 15. Social development (pp. 109 139). Thousand Oaks, CA: Sage. Ruffman, T., Slade, L., & Crowe, E. (2002). The relation between childrens and mot hers mental state language and theoryof -mind understanding. Child Development, 73, 734751. Ruffman, T., Slade, L., Rowlandson, K., Rumsey, C., & Garnham, A. (2003). How language relates to belief, desire, and emotion understanding. Cognitive Development 18, 139158. Siegal, M., & Beattie, K. (1991). Where to look first for childrens knowledge of false belief. Cognition, 38, 1 12.
44 Surian, L., & Leslie, A. M. (1999).Competence and performance in false belief understanding: A comparison of autistic and no rmal 3 -year -old children. British Journal of Developmental Psychology 17, 141 155. Symons, D. K., McLaughlin, E., Moore, C., & Mori n e, S. (1997). Integrating relationship constructs and emotional experience into false belief tasks in preschool children. J ournal of Experimental Child Psychology, 67, 423 447. Wellman, H. M., & Gelman, S. A. (1992). Cognitive development: Foundational theories of core domains. Annual Review of Psychology 43, 337375. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta anal ysis of theory-of -mind development: the truth about false belief. Child Dev elopment 72 655 684. Wimmer, H., & Weichbold, V. (1994). Childrens theory of mind: Fodors heuristics or understanding informational causation. Cognition, 53, 4557. Yuill, N. ( 1992). Childrens conception of personality traits. Human Development 35, 265279. Zeger S. L. & Liang K. Y. ( 1986). Longitudinal data analysis for discrete and conti nuous outcomes. Biometrics 42, 121 30.
45 BIOGRAPHICAL SKETCH Sunghee Ahn received he r Bachelor of Education at Duksung Women s University and her Bachelor of Science at Yonsei University in South Korea Upon completion, s he entered th e p s ychology program at the University of Florida in August 2003 and received her Master of Science in May 2006. She graduate d with her Ph.D. from the University of Florida in the summer of 2009 H er research interests include children s development of social cognition, in particular theory of mind understanding.