The behavioral rating scale of presented self-esteem for young children 1

      The concept of global self-esteem has been regarded by the field as a psychological commodity, linked to adaptive functioning and positive mental health [1]. Beginning at about age eight, children can verbalize the concept of global self-esteem, that is, they can self-report that they “like themselves overall as a person,” “that they are satisfied with who they are,” “that they like the way they are leading their lives, in general.” However, young children, ages 4 to 7, do not yet have this ability. Cognitive limitations during this period extend to the inability of young children to create a concept of their overall worth as a person, namely a representation of their global self-esteem that can be verbalized [1-3]. Such a self-representation req uires a higher-order integration of domain-specific self-evaluations that have first been differentiated.

      Young children do begin to describe themselves in terms of concrete cognitive abilities, physical abilities, how they behave, how they look, and friendships they have formed [1]. Our own research has resulted in the development of an instrument, The Pictorial Scale of Perceived Competence and Social Acceptance [4], which assesses young children’s self-perceptions in the specific domains of cognitive, physical, and social competence. However, unlike our instruments for older children and adolescents that contain items to assess global self-worth or self-esteem (e.g., “I like myself as a person”) no such items appear on our pictorial measure because they hold no meaning for young children. Moreover, the specific domains that are included are not clearly differentiated from one another, as revealed through factor-analytic procedures [4] nor are they integrated into a higher-order concept of their self-esteem.

      The literature bears out these observations. Those studies that purport to study “self-esteem” in young children are diverse, but most agree that the direct self-report of one’s worth as a person is not within the repertoire of the young child. However, studies continue to claim that they assess self-worth or self-esteem in young children employing alternative measures. Some [5,6] utilize “implicit” measures of self-esteem which reduce to a child merely describing the self as good, happy, fine, etc. versus bad, mean, or mad. Others claiming to study the development of self-esteem in young children [7] identify what they consider to be theoretically-derived dimensions of selfesteem (i.e., character, responsibility, academic competence, athletic skills, and physical appearance) assessed through selfreport items of these dimensions but do not include items to tap global self-esteem.

      Certain investigators turn to evaluations by adults in the lives of young children to assess their self-esteem [8] (see Gruber & McNinch), who have created items rated by parents that they feel reflect parental perceptions of their child’s self-esteem in the home environment). Still others have turned to teachers’ ratings of items that they merely assume assess children’s self-esteem. For example, Maxwell and Chmielewski [9] utilize a formidable 64-item instrument where teachers rate each young child in their sample. However, no psychometric data are presented nor were the items subjected to a factor analysis to determine if an interpretable number of dimensions might have been revealed. In addition, a meta-analysis [10] of numerous diverse studies assessing self-esteem, self-perceptions, and self-evaluations focusses only on the reliability of the various instruments and does not address issues of item content or the validity of these tools. Many of these investigators conclude that the field needs thoughtful measures of self-esteem in young children although they do not specify the form that such instruments should take nor the issues that should be considered.

      Thus, the present effort represents three studies that begin with the observation that young children (prior to age eight) cannot yet conceptualize the construct of global self-esteem and therefore cannot yet verbalize it. Does this mean that the young child does not possess self-esteem? We think not, based on our extensive observations and interactions with young children. Before such a global concept can be cognitive constructed, we have observed that very young children appear to experience high or low self-esteem, they exude or present it in behavioral manifestations that are observable by adults such as teachers. The first study represents open-ended interviews with experienced teachers of young children asking them to describe those behaviors that they feel are manifest by high as well as low self-esteem children. In Study 2, in order to select those items that best describe high and low self-esteem children, a Q-sort was performed with a different sample of experienced teachers. In Study 3, we took the 15 items that best discriminated high from low self-esteem in young children and crafted them into a teacher rating scale that constituted a four-point series of items, scored 4 through 1. This constitutes the Behavioral Rating Scale of Presented Self-Esteem in Young Children, which was the ultimate goal of this research.

Behaviorally-presented self-esteem in young children

      The purpose of this article is to report on the construction of such a measure, based on the fact that we, as well as teachers, can observe the behavioral manifestations that come to define self-esteem in young children. Moreover, these behaviors can be rated by teachers, day-care workers, or other adults familiar with a given group of children on a questionnaire that will reveal individual differences in self-esteem across that group of young children. Thus, in approaching the construct of “behaviorallypresented self-esteem” we first invoked the aid of teachers of young children who had considerable experience with this age group. Haltiwanger found that early childhood educators frequently make reference to children’s self-esteem and that this is a very prevalent and meaningful concept that distinguishes children from one another.

Study 1: Open-ended interviews with experienced teachers

      As a first step in the development of a teacher questionnaire to tap behaviorally-presented self-esteem in young children, Haltiwanger conducted open-ended interviews with 20 teachers who had considerable experience (15 years on average) with young children in order to generate an item pool from which we would eventually select those items that best discriminated high- from low-self-esteem children. Teachers in this first study were asked to describe those behaviors that characterize the prototype of the high self-esteem child and those that characterize the prototype of the low self-esteem child. However, equally important were their descriptions of those behaviors that they felt did not allow them to discriminate between the two self-esteem groups. Teachers had definite opinions about behaviors that were both relevant and irrelevant to this construct.

      From these interviews, we culled 84 behavioral descriptors, brief phrases representing behaviors that teachers felt described the prototypical high self-esteem and the prototypical low self-esteem child, behaviors that strongly discriminated between these two types of children. Also important to this first study was teachers’ identification of behaviors that did not discriminate the two groups, behaviors that were not relevant to the definition of behaviorally presented self-esteem.

Study 2: The development of a Q-sort procedure

      We next employed a traditional Q-sort procedure in which Haltiwanger [11] asked a separate group of teachers (from 7 different educational settings serving young children) to sort these 84 items into nine categories (see Appendix). Teachers placed the descriptors, transferred to cards, on a larger depiction of this Figure. On the far left they indicated which behaviors were most like the high self-esteem and in the adjacent columns they placed descriptors that were less and less like the high self-esteem child. In the tallest column in the middle they placed those behavior that were irrelevant, neither like or unlike the high or low self-esteem child. At the extreme right of the diagram they selected behaviors that were most like the low self-esteem children. The adjacent columns moving toward the center column represent behaviors less and less like the low self-esteem child.

      After these sorts were completed by teachers, we focused on analyzing those behaviors in the most extreme categories that define both the prototypical high self-esteem child and the prototypical low self-esteem child. Reliability analyses indicated very substantial agreement (81%) among teachers.

Content analysis of the Q-sort results

      Behaviors that discriminate between high self-esteem and low self-esteem children. There were two primary categories of items that defined the high self-esteem child in the eyes of experienced teachers.

      a. Active displays of confidence, curiosity, initiative, and independence. Examples of this category include: Trusts his or her own ideas, approaches challenge with confidence, initiates activities confidently, takes initiative, sets goals independently, is curious, explores and questions, is eager to try doing new things. Two other behaviors seemed to convey the more general manifestation of these attributes: describes self in positive terms and shows pride in his or her work.

      b. Adaptive reaction to change or stress. Examples of this category include: Able to adjust to changes, comfortable with transitions, tolerates frustration and perseveres, and social-emotional expressions such as smiling.

      Corresponding categories describing the low self-esteem child represent the converse of these two sets of items, emerged:

      a. Failure to display confidence, curiosity, initiative, independence. Examples include: Doesn’t trust his or her own ideas, lacks confidence to initiate, lacks confidence to approach challenge, is not curious, does not explore, hangs back, only watches others, withdraws and sits apart, and describes self in negative terms.

      b. Difficulty in reacting or adapting to change or stress. Examples include: Gives up easily when frustrated, reacts to stress with immature behavior, reacts inappropriately to mistakes, face shows negative feelings.

      This content analysis is particularly illuminating given what it reveals about the nature of self-esteem as seen through the collective eyes of experienced teachers. It suggests two primary dimensions, one active and one more reactive. The active dimension represents a style of approaching or confronting challenging situations rather than the display of skills per se. That is, the high self-esteem child manifests confidence, curiosity, and interest in his/her world, whereas the low self-esteem child avoids challenge, novelty, and exploration. The reactive dimension involves the response of the child to change, frustration, or stress. The high self-esteem child reacts more adaptively, whereas the low self-esteem child reacts with immature, inappropriate, or avoidant behaviors.

      What behaviors do not discriminate? Of particular interest were the behaviors that do not seem to discriminate between high and low self-esteem children, according to teachers. Most noteworthy, if not striking, was the fact that competence per se is not a correlate of overall self-esteem in young children, for example, cognitive or physical skills were deemed irrelevant. Rather, displays of confidence were markers. Behavioral examples were moving forward to show personal initiative and approaching challenging tasks with confidence. It would thus appear that confidence, as a behavioral style, is not synonymous with competence, at least at this age level. This is illuminating since it suggests that the origins of a sense of confidence during early childhood do not necessarily reside in the display of skills, more objectively defined. Rather, confidence is manifested in taking initiative, preferring challenge and making decisions about activities to explore. During later childhood, the link between confidence in the self and one’s level of competence apparently becomes stronger. In early childhood, the developmental path to high self-esteem will be facilitated by adult support (e.g., from parents and teachers), including sensitivity, and contingent responsiveness. In particular, specific support for exploration, mastery, curiosity, and personal initiative will all promote a sense of confidence that in will eventually contribute to the development of skills reflecting competence.

      Thus, in middle childhood, competence will become a much more critical factor contributing to self-esteem, as our own work has demonstrated in numerous studies [1]. We would argue, in bridging these two developmental periods, that socialization practices that reward displays of confidence will lead the young child to engage in behaviors that would allow him/her to begin to develop skills and competencies that will subsequently become a defining predictor of self-esteem. This is a major and significant finding of these preliminary studies, namely that confidence and not competence seems to represent the defining features of behaviorally-presented self-esteem which, we hypothesize may predict subsequent competence in the middle years of childhood and beyond. That is, the confident young child will begin to expand his or her world through exploration and curiosity. This will undoubtedly promote skill learning, mastery, knowledge, cognitive competence, and physical skills. If such skills are considered to be important, as William James argued in 1898, they will eventually lead to high verbalizable self-esteem in the middle childhood years and beyond [1].

Study 3: The development of a behavioral rating scale of presented self-esteem

      The next step toward this goal was to take the most commonly identified characteristics identified by teachers that defined high self-esteem and low self-esteem in young children in the Q-sort. We only selected items where there was major agreement among teachers as to the most common features of high self-esteem and low self-esteem children. This process led to our categorizing attributes and then examining the frequencies for each of the attributes that teachers had identified. Our goal was to develop an economical set of items that experienced observers of children, be they teachers or others very familiar with a group of children, could easily rate. After a few iterations, we developed a scale in which 15 items, which pitted high-self-esteem descriptors against low self-esteem descriptors, were created, based upon the most common distinctions that had been reliably identified by teachers. Noteworthy is the fact that in the construction of this final instrument we did not include those items which did not define either the prototypical high- or low-self-esteem child. These non-discriminating behaviors were important to identify, in conceptualizing what did and did not constitute behaviorally-presented self-esteem in young children, from teachers’ perspectives. However, we reasoned that items that did not discriminate the two groups should not be included in the actual measure of those behaviors that did define high and low self-esteem.

      It is recommended that this rating scale be filled out by children’s main teacher who has had sufficient time in the classroom to have observed a variety of behaviors including those with other children in that class. The items on the scale were generated by teachers and occur within the classroom. Certain items involve interactions with other children. A different adult (e.g., a camp counselor or adult monitoring children in a play group) observing the children in a group setting may also be appropriate.

Final version of the behavioral rating scale of presented self-esteem in young children

      We developed this 15-item questionnaire including two clusters of items that were somewhat conceptually different, although we did not anticipate that they would define statistically separate factors. We expected that they would be highly related as revealed in the correlations we subsequently conducted. However, the first category included more items (9 versus 6) suggesting that they carry more weight in defining presented self-esteem in young children in the eyes of teachers.

      The first category reflected active displays of confidence, curiosity, exploration, initiative, and independent goal setting which defined high self-esteem. Low self-esteem was defined as the absence of these characteristics, namely, the failure to display confidence, curiosity, exploration, initiative, and independent goal setting. Each of these opposing attributes were pitted against one another in a given item (as will be described in the next section of the manual). There were nine items in this first category on the rating scale itself, items 1, 3, 4, 5, 8, 9, 12, 13, 15.

      The second category of items assessed reactions to change or stress, often in social situations. Positive reactions indicative of high self-esteem include the ability to assert one’s point of view if opposed by other children, getting involved in challenging social situations, tolerating frustration, and expressing positive facial expressions, even in the face of stress. Within a given item, the opposite of these characteristics were captured as reflections of low self-esteem. These included behaviors such as inability to express one’s point of view, avoiding challenging social situations by hanging back, giving up easily if frustrated, and expressing negative facial expressions. There were six items in this second category, items 2, 6, 7, 10, 11, 14.

      The fact that there are a few more items (nine) in the first category representing Confidence/Initiative (C/I) than those six representing Reactions to Stress (R/S) reflects the findings of the initial teacher Q-sort. Teachers felt that there were more items in the first category defining self-esteem than in the second category. However, the strength of items in each category was comparable.

Question format

      The question format was modelled after “the structured alternative format” devised by Harter [12,1] in her development of a life-span battery of Self-Perception Profiles. The purpose of this format was to discourage respondents, in this case teachers, from giving socially-desirable responses. As the examples below reveal, the respondent is presented with two statements, one on the left and one on the right, for each item. The adult respondent is asked to indicate which of the two statements is more descriptive of a given child. After making this decision, the teacher (or other adult rater) goes to that side of the item and checks whether that description is “Very Much Like this Child” or only “Sort of Like this Child.” Thus, the adult rater selects only one response per item, sometimes it will be on the left, sometimes on the right, because the items are counter balanced in terms of whether the positive or negative description occurs first. The sample ratings below reflect the high self-esteem child.

      Scoring: Items are scored from 4 to 1 where the higher scores reflect higher Self- Esteem. Those items which present the higher self-esteem description on the left side of the statement are scored 4, 3, 2, 1, for example, the first and fourth sample items above.

      Counterbalanced items that present the low self-esteem description on the left are scored 1, 2, 3, 4, for example, the second and third sample items above (see the scoring key for the entire scale in the Appendix).

      Across the 15 items, the scored are summed and then divided by 15. Thus, each child will have a score between 4, designating the highest self-esteem, and 1, representing the lowest self-esteem. We are suggesting the following interpretive guidelines. Mean scores from 3.5 to 4 reflect very high self-esteem. Those from 2.5 to 3.4 reflect moderate self-esteem. Those from 1.5 to 2.4 reflect low self-esteem. Finally those from 1 to 1.4 reflect exceeding low self-esteem that should be treated as cause for concern.

Study 3: Administration of the behavioral rating scale of presented self-esteem scale for young children

      New samples of teachers of young children were recruited in area schools. In samples of 5 different teachers rating their pupils in classes of approximately 30 children each, the internal consistencies ranged from .85 to .91 for the 15-item scale.

      The mean scores across the five different classrooms ranged from 2.99 to 3.10, above the midpoint of the scale, 2.5, where scores which could range from 1 to 4. Within each classroom, good variability was observed as indicated by standard deviations of from .67 to .80. Thus, young children do vary in their behaviorally-presented self-esteem as judged by their teachers.

      These findings were supported in an independent study by Verschueren, Marcoen, & Buyck [13] who reported a mean teacher rating of 2.99 in a sample of five-year-olds, with excellent variability between 1.4 and 4.0 across individual children. Internal consistency reliability (Cronbach’s alpha) for the 15- item scale was reported to be .95.

      We have not recommended factor-analytic procedures given that no distinct a priori factors were built into this particular instrument. Nor was there any conceptual framework suggesting that different factors might emerge. This was a purely empirical effort to determine whether, in the judgments of teachers of young children, there were some common behaviors observed by teachers that reflected their views of the manifestations of self-esteem. Thus, we have calculated a single score across the 15 items.

      Conceptually, we did identify two clusters of items. The larger set of items reflected initiative- or confidence-related themes (approach vs. avoidance of age-appropriate mastery tasks in the classroom). A smaller cluster of items tapped more socioemotional behaviors that addressed how the child coped with common challenging social situations or frustrations. However, we did not anticipate nor did we find that these two clusters defined separate factors, in fact, they were highly correlated. Verschueren et al. [13] have reported similar findings, namely that a one-factor solution is the most appropriate.

      Discriminant validity: In one sample, teachers were asked to rate their children on a one-item question that assessed the child’s overall self-esteem. A second one-item question asked teachers to rate the child’s intelligence. Findings revealed a correlation of .65 for the behaviorally-presented self-esteem ratings and the summary assessment of the child’s self-esteem. In contrast, the behaviorally-presented ratings correlation with the summary rating of the child’s intelligence was only 39.

      These findings bolster the contention, based on the Q-sort results, that behaviorally-presented self-esteem does not manifest itself in cognitive skills per se, at this age level. As we concluded, presented self-esteem is more highly linked to displays of confidence, more so than competence. Moreover, we know from considerable research (see Harter) [1] that beginning in middle-childhood and beyond, perceived cognitive competence, if deemed important to the participant, is highly predictive of perceived verbalizable global self-esteem. One can only conjecture that the young child who displays initiative, curiosity, exploration, and confidence is laying the groundwork to become a competent learner in middle childhood and beyond which will, in turn, predict self-esteem.

      Our approach to the assessment of behaviorally-manifest, presented self-esteem in young children began with our assumption that experienced teachers, in a society focused on the importance of self-esteem, would develop a sensitivity to observable behaviors in young children that might reflect developmentally-early manifestations of this construct. Our own earlier research has revealed that at about age eight, children develop a concept of self-esteem, their overall worth as a person, that they can verbalize [1]. Prior to this age, such a verbalizable concept is not in their self-repertoire. Young children can give vivid descriptions of their virtuosity in specific domains, for example, their cognitive and physical competence. These are typically exaggerated given that they cannot yet make the distinction between their real and their ideal self-concepts. However, they cannot yet cognitively formulate a higher-order conceptualization of their overall worth as a person, our societal definition of global self-esteem.

      Yet do children display some experiential sense of their selfesteem, their worth as a person, and what might that look like? The two co-authors, with a great deal of independent experience in many classrooms of young children, came to the conclusion that such children do exude a rudimentary sense of self-esteem that can be observed in their behavior. Witness the young child purposely marching across the classroom toward the new, challenging puzzle in the play area. Observe the child at “showand-tell time,” confidently describing how the penguin toy he or she brought represents a rather strange bird who has a hard life, in reality. Such displays of mastery and confidence do not go unnoticed by experienced teachers who choose to foster such attributes in their pupils.

      Thus, our beginning point was to trust the observations of teachers, to allow them to educate us about the manifestations of self-esteem in young children. We began with interviews of teachers, which Haltiwanger conducted, compiling a large battery of suggestions for what defined the prototypically high selfesteem child and the prototypically low self-esteem child. However, an important feature of this first phase was to include, in the interview, what behaviors were irrelevant to self-esteem, an insight that became very critical to our research question.

      Based on these open-ended responses, we culled 84 brief descriptions that represented a range of observable behaviors, with varying relevance to high and low self-esteem. We subjected these to a Q-sort procedure with new samples of teachers which pared down the list of behaviors to 15 that seemed to reliably distinguish between high and low self-esteem children.

      We converted these behaviors into a questionnaire format where those familiar with a group of children could rate each of them on these 15 behaviors. These ratings came to define individual differences in “presented self-esteem” based on teachers’ judgments, which possessed face validity, discriminant validity, and high reliabilities. The content of this instrument revealed that dimension of confidence, initiative, curiosity, exploration, adaptive reactions to frustration and stress defined teachers’ concepts of the early precursors of behaviorally-manifest selfesteem in young children. High self-esteem children manifest these characteristics. Low self-esteem children lack these attributes, in fact, they show the opposite.

      Of particular interest, and perhaps somewhat counterintuitive, were the behaviors that teachers identified in the Q-sort to be irrelevant to self-esteem. Most notable were cognitive and physical skills. That is, behaviors that reflected competence were not defining features of self-esteem, in the judgment of teachers, at this age level. Rather, behaviors that demonstrated confidence were much more relevant to self-esteem, from the teachers’ perspective.

      What are the implications of these findings for educators and parents? They suggest that stimulating a child’s curiosity, rewarding their interest in exploring topics that excite them, supporting mastery goals, encouraging initiative, and helping them to express what they learn will have enormous benefits. Perhaps more so than skill drills at this age level. The reasoning, here, is that children who develop these attributes will eventually evolve into competent learners in areas which interest them. The history of education has dictated skill domains that are deemed critical to master, the initial “three Rs”, for example, reading, writing, and arithmetic. We are not arguing that such basic skills should not be taught. However, they should complement the motivation to master subjects of interests, to foster curiosity and exploration and the initiative to pursue topics of interest that command children’s attention. Unfortunately, our American school system seems to be going in the opposite direction, with its emphasis on standardized tests to prove not only a student’s competence but the quality of teachers, who are now forced to teach to the test in order not only to prove their own worth but justify their pay and their occupational future (see Harter) [3].

      The findings of the present research, designed to devise a questionnaire for teachers or other adults working with young children, to rate their self-esteem has two other implications that were not anticipated at the inception of this effort. As noted in the introduction, we had no particular theoretical framework guiding this research. Rather, we trusted experienced teachers to give us their opinions about the observable manifestations of self-esteem in young children. The first insight was to closely examine the content of the items that teachers felt defined self-esteem. With no predilections in mind, it struck us that the behaviors that teachers identified were very consistent with attachment theory [14,15]. That is, the behaviors that define a secure attachment style were precisely those that teachers identified as representative of high self-esteem children. These included exploration, initiative, confidence, curiosity, and adaptive social relationships. In contrast, attributes that define the anxious and avoidant attachment styles became apparent as descriptors of low self-esteem children. We find this interpretation intriguing.

      However, there is one final caveat, a second morsel of food for thought. The construct of self-esteem has lost some of its luster in the last two decades (see the summary of arguments in Harter) [3]. As a nation, we have been accused of attempts to enhance self-esteem unrealistically [16]. For example, California attempted to elevate its citizens’ self-esteem on a state-wide level, without success. Schools have followed a playbook of interventions to enhance self-esteem, convincing students that they are special when they are not. Leary [17] has articulated the many ways that self-enhancement is a curse, describing the various strategies that adults engage in, including self-serving biases that are designed to lead people to believe that they are much more talented than they have a right to believe. A carefully constructed study has now demonstrated that narcissism in on the rise [18].

      How are these observations to be reconciled with the findings of our study of young children’s self-esteem? Arguments that self-esteem in contemporary society is inflated, that selfenhancement strategies are more in evidence, and that interventions to enhance feelings of self-worth are prevalent apply to self-reported or verbalized self-esteem. The “I am awesome” phenomenon may well need to be tempered in our youth. Recently, a high-school principal did just that in his commencement address to seniors and their proud parents. He quite simply said to and about students, “You are not that special, contrary to what you may think!” The students were horrified and the parents were outraged. Many egos, in both generations, were bruised. His remaining arguments were undoubtedly lost on the audience. But he preached about realism and the value of hard work and effort in order to obtain a special status in our society. One has to earn entitlement, he emphasized. It is not conferred by a mere high school diploma. Needless to say, the post- graduation parties lacked the expected cheer.

      Indiscriminant praise at any age will reinforce unrealistic views of self that will eventually lead to inflated self-esteem when that concept emerges in a child’s vocabulary. With young children, we too often automatically mouth the phrase “Good job!” with feigned enthusiasm, when the “job,” quite frankly, has not been that good! And, as Damon [16] has pointed out, children know it. Effusive, unearned, praise does not serve our children well. However, reinforcing them for mastery efforts, curiosity, exploration, and initiative, guiding them toward ageappropriate skills, showing pride in truly good performance will ultimately result in self-perceptions that are positive and realistic. Thus, the very behaviors that teachers have thoughtfully identified as the early manifestations of self-esteem in young children are precisely what we should be supporting.