Abstract
Research has demonstrated mixed results regarding differences in social and emotional characteristics between gifted and typical populations. The purpose of this secondary analysis of data from the Early Childhood Longitudinal Study: Birth Cohort (ECLS-B) is to investigate the affective characteristics of early mathematics and literacy ability among preschool children using a logistic regression analysis. Specifically, parents and childcare providers were asked a series of questions relating to socially maladaptive behaviors, concentration, empathy, worry, and friendship of a nationally representative sample of children born in 2002. The results of the study demonstrate that childcare providers and parents of preschool children (M age = 53 months) have different patterns of observations of social and emotional characteristics of children, and that concentration and socially maladaptive behaviors were significant predictors of early giftedness in literacy. There were no social or emotional predictors of early giftedness in mathematics found in this study.
Parents and childcare providers provide insights into the social and emotional development of the children under their care. When taken collectively, these observations can illuminate patterns of development across groups. Few studies, however, have investigated differences in social and emotional development between gifted and typical children in early childhood prior to beginning formal schooling. This is an important perspective, as social interactions have an impact on academic success (e.g., Bergen, 2002). In this study, the researcher used a nationally representative database (Early Childhood Longitudinal Study: Birth Cohort [ECLS-B]; Snow et al., 2009) of children in the United States to analyze the social and emotional characteristics that are related to mathematics and reading achievement in preschool children. By gaining an understanding of the affective components of development, researchers will be able to formulate effective curriculum and interventions for gifted children in preschool.
Specifically, the relation between parents’ and childcare providers’ observations of preschool children’s social and emotional characteristics and early childhood giftedness, as measured by mathematics and reading ability using the ECLS-B database, was investigated in this study. In addition, the correlations between parents’ and childcare providers’ observations of child characteristics were examined. The ECLS-B is a nationally representative, longitudinal database including children from birth to their first year of kindergarten, collected by the National Center for Educational Statistics (NCES), part of the Institute for Educational Sciences.
Background
Social and Emotional Characteristics of Gifted Children
The field of gifted education has long been concerned about the social and emotional development of gifted children (e.g., Robinson, 2002). The trend in the empirical research is that gifted adolescents are as well-adjusted, or more, when compared with typical adolescents (Jones, 2013), with students identified by general ability (Mueller, 2009), domain-specific aptitude (Vialle, Heaven, & Ciarrochi, 2007), and academic achievement (Sayler & Brookshire, 1993). As a part of a larger compilation of research regarding academic acceleration, Robinson (2004) concluded that gifted children who were accelerated did not tend to have greater social and emotional difficulties, but individual differences were important to adjustment to new educational contexts. In reviewing the existing research, Reis and Renzulli (2004) reported that higher ability students are well-adjusted as compared with their peers, but some issues may develop when there is a mismatch between gifted students’ abilities and educational environments. In addition, Bain and Bell (2004) found that teachers of fourth through sixth graders did not report any differences in peer relationships between high-achieving gifted and nongifted children. However, many researchers emphasize that although gifted students are often well-adjusted, global measures may mask differences in patterns of emotional well-being among gifted students identified by their respective school systems (e.g., Blackburn & Erickson, 1986; Freeman, 1997; Sowa & May, 1997). Specifically, gifted adolescents may face differences in expectations, stress, awareness of social and societal difficulties, social acceptance, asynchronous development, and emotional sensitivity, as proposed by researchers, such as Altman (1983); Neihart, Reis, Robinson, and Moon (2002); and Hébert (2011). There is a considerable body of research concerning these topics among older students; however, this article will focus on the research specific to early childhood.
Social and emotional characteristics of early childhood giftedness
Despite the interest in the social and emotional development of gifted students, relatively few studies have investigated the social and emotional differences in early childhood. Some studies have found that there are no differences in social development for gifted and nongifted children in early elementary school with identification based on school criteria with multiple measures (Curby, Rudasill, Rimm-Kaufman, & Konald, 2008) and in a preschool university-based laboratory school with identification based on intelligence test scores (Lupkowski, 1989). However, Curby et al. (2008), as a part of the same study, found that gifted elementary children have higher levels of task orientation. Studies also have found that parent and childcare provider observations of intellectually gifted preschool children’s characteristics were aligned with standardized measures, but these studies focus more on developmental and cognitive characteristics than affective (Hodge & Kemp, 2000; Ogurlu & Cetinkaya, 2012). The few studies that have observed young children’s play behaviors demonstrate that intellectually gifted children tend to engage in higher levels of social play at a laboratory school (Lupkowski, 1989). In addition, Barnett and Fiscella (1985) found that intellectually gifted preschool children had higher levels of cognitive play when compared with matched-aged peers of typical abilities. Finally, Abroms and Gollin (1980) found that social and intellectual giftedness were unrelated among 3-year-olds. I believe that the paucity of this research, particularly including subject-area aptitude, indicates a need for an investigation of social and emotional characteristics of young children using a large data set that is nationally representative.
In another line of research, social skills and behavior have been widely documented among young children as being significant predictors of early school success (e.g., Agostin & Bain, 1997). For example, Bramlett, Scott, and Rowell (2000) demonstrated that teacher and parent ratings of affective characteristics (e.g., adaptability and persistence) at the beginning of the year were significantly linked to first-grade achievement at the end of the year. Based on current educational practices and accountability mechanisms, Denham (2006) recommended increased social and emotional measures to highlight the importance of these constructs in academic development throughout elementary school. However, this research has not focused on how social and emotional development interacts with high abilities in young children.
Definition of Giftedness
Not surprisingly in a field as diverse as gifted education, many researchers have conceptualized the construct of giftedness in various ways (Sternberg & Davidson, 2005). The National Association for Gifted Children (NAGC; 2010) defined giftedness in the following way: Gifted individuals are those who demonstrate outstanding levels of aptitude (defined as an exceptional ability to reason and learn) or competence (documented performance or achievement in top 10% or rarer) in one or more domains. Domains include any structured area of activity with its own symbol system (e.g., mathematics, music, language) and/or set of sensorimotor skills (e.g., painting, dance, sports). (para. 1)
Further refining this definition, in this study, I narrowed the domains to early literacy and mathematical competence (rather than ability) in the top 5% of the nationally representative sample. Although this does not include broader definitions of giftedness (e.g., Renzulli & Reis, 1997; Sternberg, 1999), the measures of early literacy and mathematical achievement were chosen because they represent salient measures of early school success and are relevant indicators for children entering school and elementary gifted programs. Thus, for the purposes of this study, early giftedness is defined in terms of early competence in a distinct set of skills rather than more general measures of cognitive abilities or broader definitions of gifted behaviors, and is limited to advanced academics. This approach to identification, based on achievement and demonstration of academic need (rather than global measures of intelligence), has been advocated by researchers in the field of gifted education in recent years (e.g., Peters, Matthews, McBee, & McCoach, 2013). It is important to note that there are no data from the database about the eventual placement in gifted programs of children identified in this study as gifted in mathematics or literacy, and thus, the term gifted in this study is not directly linked to a child receiving services through schools. I have focused this study on domain-specific abilities in young children because this construct is more salient in early school contexts and educational need.
Perceptions of Children From Parents and Teachers
Although there are no studies directly comparing parent and teacher/childcare provider reports of the social-emotional characteristics of young gifted children, the perceptions of these two groups are integral to the nomination process for gifted services once the child begins school (e.g., Johnsen, 2011). Both parents (Ciha, Harris, Hoffman, & Potter, 1974; Jolly & Matthews, 2012) and teachers (e.g., Gagné, 1994) are viewed as having valid recommendations of children in need of gifted services. However, researchers have shown that teacher recommendations and perceptions are also influenced by race and socioeconomic status (SES; McBee, 2006), gender and subject-area interest (Siegle, Moore, Mann, & Wilson, 2010), and social and personality characteristics (Alviderez & Weinstein, 1999; Endepohls-Ulpe & Ruf, 2005). Furthermore, Persson (1998) demonstrated that teachers in Sweden ranked gifted children as more emotionally mature and more willing to help others. Given these findings, and the importance of teacher and parent nominations in the identification process, the investigation of these sources (e.g., parents and childcare providers) of ratings of the affective characteristics of young children is a part of this study.
The correlations between these sources are investigated, as previous researchers have shown low correlations between these two groups in rating the behavior of children (e.g., De Los Reyes & Kazdin, 2005). The moderate to low correlations between parents and teachers have been reported among children with emotional problems (Achenbach, McConaughy, & Howell, 1987), children with serious emotional disturbances (Greenbaum, Decrick, Prange, & Friedman, 1994), children suffering from depression (Reynolds, Anderson, & Bartell, 1985), and children with behavioral problems (Kolko & Kazdin, 1993). In addition, Sankar-DeLeeuw (2002) found that parents and teachers of young gifted children had differing views of early childhood giftedness. In the current study, I investigated whether this pattern of correlations is true for parents and teachers of children in the general population.
Early Mathematical Giftedness
Early talent in mathematics has long been studied in the field of gifted education (Stanley, 1996); however, these studies typically begin identifying students well past entering school. The robust findings from longitudinal studies of mathematically gifted children indicate high levels of academic success and career attainment for students identified in middle school, particularly in areas of science, technology, engineering, and mathematics (STEM; Benbow, 2012; Wai, Lubinski, Benbow, & Steiger, 2010). Studies of young children indicate that mathematical precocity is unrelated to working memory (Okamoto, Curtis, Jabagchourian, & Weckbacher, 2006) or creativity (Baran, Erdogan, & Cakmak, 2011). These children did exhibit higher levels of numerical conceptual structure (Okamoto et al., 2006). Observable behaviors of mathematically gifted young children included choosing to play or work with numbers, finding patterns, planning steps to solve problems, grouping objects, assembling puzzles, and grouping/regrouping objects in many ways (Lapp & St. John, 2009). In addressing elementary-aged children with mathematical promise, Sheffield (2003) discussed mathematical frames of mind, formalization and generalization, creativity, curiosity, and perseverance. Thus, her conception of mathematical giftedness in childhood includes affective characteristics and dispositions. Based on these characteristics, scholars have recognized the need for and importance of early identification and rigorous and challenging curriculum for young mathematically gifted children (Koshy, Ernest, & Casey, 2009). However, little research has been conducted investigating the affective characteristics of this population of young children, which may have educational implications, as indicated by Agostin and Bain (1997) on the general population of children.
Measuring early mathematical giftedness
For the purposes of this study, children were classified as mathematically gifted if they scored in the 95th percentile of the nationally representative (weighted) sample on the ECLS-B Cognitive Direct Child Assessment: Preschool Math. Items for this assessment were selected by NCES to measure school readiness in mathematics and to align to measures included in the Early Childhood Longitudinal Study: Kindergarten Cohort (ECLS-K) data set (a national longitudinal study of children from kindergarten through middle school; Tourangeau, Nord, Le, Sorongon, & Najarian, 2009) and the developmental cognitive measure in previous waves of data collection in the ECLS-B (Snow et al., 2009). The measures of academic mathematical tasks are appropriate for this study’s definition of mathematical giftedness because they relate both to developmentally appropriate mathematical tasks and to the types of abilities that are likely to be recognized as advanced in future school settings. The specific constructs that were measured included number sense, geometry, counting, operations, pattern understanding, and measurement (Snow et al., 2009). These were assessed through the use of manipulatives (i.e., color counters) and, for the most advanced students, more abstract arithmetic (i.e., number sentences). Thus, this assessment used developmentally appropriate measures of school-related mathematical skills and provided a way to identify mathematically gifted children from this large data set.
Early Language and Literacy Giftedness
One of the most salient characteristics of giftedness in early elementary school is advanced language abilities (e.g., Gross, 1999), particularly early and advanced reading skills. Precocious readers experience early academic success in the areas of word decoding (Foster & Miller, 2007), as well as reading comprehension, spelling, and fluency (Tafa & Manolitsis, 2008). Thus, students entering school with strong reading and pre-reading skills are more likely to be identified by teachers as advanced or gifted. However, research does indicate that early reading is not significantly related to overall intelligence scores but is more likely related to the ability to coordinate many processes and skills together to read fluently (Jackson, 1988). Precocious readers do tend to have higher levels of visual memory (Olson, Evans, & Keckler, 2006). Taken as a group, precocious readers tend to come from families that value education, are highly responsive to questions, and frequently participate in shared reading of storybooks with high-quality interactions (Crain-Thoreson & Dale, 1992; Olson et al., 2006). However, there is little research concerning the affective differences of young children with high levels of literacy skills, which may help to inform the planning of successful educational programs for these learners.
Measuring early language and literacy giftedness
For the purposes of this study, early literacy giftedness is defined by the top 5% of scores on the ECLS-B Cognitive Direct Child Assessment: Preschool Literacy Assessment. This assessment was meant to measure school readiness, and thus included many items from the ECLS-K data set (Tourangeau et al., 2009), as well as developmentally appropriate items from other language assessments used in large-scale studies (Snow et al., 2009). The NCES-trained staff gave this assessment only to children who passed the initial screening of competence in the English language; non-English speakers are not included in the analysis (Snow et al., 2009). The assessment was aligned to national literacy frameworks and included phonological awareness, letter sound knowledge, letter recognition, print conventions, and word recognition (Snow et al., 2009). High performance on this assessment indicates strong pre-reading and early literacy skills that would predict early academic success in elementary schools and represent early competence in the domain of literacy, which is defined in the current study as early literacy giftedness.
Comparison of Affective Characteristics of Children Gifted in Content-Specific Areas
Several studies have demonstrated differences between children gifted in various subject areas (e.g., Dauber & Benbow, 1990). Analyses of the Study of Mathematically Precocious Youth (SMPY), a longitudinal study of gifted adolescents begun in 1971, indicated that verbally gifted adolescents tend to perceive themselves as less able in social contexts than mathematically gifted adolescents (Brody & Benbow, 1986; Dauber & Benbow, 1990). Even as early as 1945, Cattell found advanced verbal abilities associated with nervous emotionality, whereas mathematical ability was associated with dominance. Additional researchers have documented that adolescents tend to define their academic self-concept in verbal and quantitative domains (e.g., Marsh & Shavelson, 1985), which provides validity to the constructs used in this study as domains of giftedness.
Due to the dearth of research regarding giftedness and advanced achievement in early childhood, there is a need for a systematic study of a large database to identify whether there are social or emotional predictors of early childhood giftedness in mathematics and reading. As social interactions and emotional development play a key role in the development of academic abilities, it is important to investigate whether children identified as gifted in these subject-area domains have characteristics different from typical children, and whether parent and teacher perceptions of these characteristics differ.
Purpose of the Study
Given the relative paucity of research regarding young gifted children, particularly regarding subject-specific abilities and affective characteristics, this study was undertaken to utilize a large, nationally representative data set to further describe this population. The correlation of ratings of parents and teachers can also help inform the identification process and help focus on valid sources of data regarding the affective characteristics of young children. Finally, because researchers have shown that social and emotional development is linked to early school success (e.g., Agostin & Bain, 1997), this topic can be useful in planning programs for the gifted. As research concerning the affective characteristics of gifted adolescents becomes more nuanced, the results of this study of early childhood can help inform the research.
The purpose of this study is to analyze the affective characteristics of young children with advanced academic achievement. The research questions guiding this study are as follows:
Method
To answer the four research questions, I conducted this study in three parts. To answer the first question, a factor analysis on parent and childcare provider interview items regarding the social and emotional traits of the children was conducted using the entire, unweighted sample of children who had non-familial childcare outside their home. Then, answering the second question, a bivariate correlation between the emerging factors was conducted. Finally, to answer the last two questions, two logistic regression analyses, with giftedness in mathematics and literacy as outcome variables, were conducted with defined samples, adjusted to reflect national demographics using weights provided by NCES.
Data Sources
The ECLS:B is a federally funded project that collects a variety of data about social, health, and educational measures of young children across the United States. Data from the third wave of the ECLS:B (Snow et al., 2009) were used in this study. This sample was drawn from children born in the United States in the year 2002, and the third wave of data collection consisted of those children in the sample at approximately 4 years of age. Data from only the third wave of data collection were used for this study, as it represents the age of interest, prior to the onset of formal schooling, but developed enough to demonstrate the evidence of early mathematical and literacy skills. The sample includes children from every state and region in the United States, and includes children from rural, suburban, and urban communities. The sample used for the logistical regression analyses includes the gifted group of children scoring in the 95th percentile on the achievement tests (mathematics or literacy) and the comparison group of children scoring in the 50th to 55th percentile on the same tests. The total unweighted sample size was 650 for the literacy analysis and 550 for the mathematics sample; these have been rounded to the nearest 50, per NCES guidelines (Snow et al., 2009). The ECLS:B data were gathered with a two-stage cluster-based system, in which certain populations of children (i.e., American Indian, low birth weight infants, and twins) were oversampled (Snow et al., 2009). Thus, the use of a weighted sample is necessary for data analysis to reflect a nationally representative sample and was used for the regression analysis in this study. The weights used for the analyses were provided by the NCES (see Snow et al., 2009) based on the sources of the data (i.e., direct child, parent interview, and childcare provider interview) and the wave of data collection (i.e., third wave). These weights were applied for the logistic regression analyses.
Because the gifted and comparison groups were defined separately for the mathematics and literacy analyses, two sample groups were used in this study (see Table 1 for the demographic information for each sample). In the mathematics sample, there were slightly more males (52.0%), and in the literacy sample, slightly more females (57.8%). In both samples, the majority of children were White (mathematics, 66.6%; literacy, 64.2%), with some African American (mathematics, 13.3%; literacy, 11.6%), Hispanic (mathematics, 7.7%; literacy, 10.6%), Asian (mathematics, 5.2%; literacy, 6.2%), and multiracial children (mathematics, 6.0%; literacy, 7.1%). The majority of children in both samples attended center-based care other than Head Start (mathematics, 74.0%; literacy, 84.2%), with comparatively fewer attending home-based care (mathematics, 15.1%; literacy, 9.4%) or Head Start (mathematics, 10.9%; literacy, 6.4%). As the analysis of data for this study relied on reports of affective characteristics from both parents and childcare providers, children who did not have non-familial childcare arrangements outside the home on a regular basis were not included in this study. The average age for the mathematics analysis was 53.42 (SD = 3.89) months and 53.22 (SD = 3.71) months for the literacy analysis (see Table 2). The unweighted number of participants in the mathematics sample was 550 (rounded to the nearest 50 per NCES guidelines), with approximately equal numbers of gifted (unweighted n = 300, rounded to the nearest 50; n = 59.4%) and typical children (unweighted n = 250, rounded to the nearest 50; n = 40.6%). Similarly, the literacy sample had an unweighted sample size of 650 (rounded to the nearest 50 per NCES guidelines) with 55.4% gifted (unweighted n = 350) and 44.6% typical children (unweighted n = 300). Approximately 17.9% of the gifted sample (unweighted n = 200, rounded to the nearest 50) were identified as gifted in both mathematics and literacy samples, this is approximately 3% of the total sample. For the mathematics analysis, the mean mathematics score (M = 41.33, SD = 9.69) was higher than the literacy score (M = 37.06, SD = 13.43). The opposite pattern held for the literacy analysis, with a mean literacy score (M = 39.80, SD = 14.20) slightly higher than the mathematics score (M = 39.30, SD = 9.70). Finally, the mean socioeconomic score was slightly higher for the mathematics analysis (M = 0.61, SD = 0.77) than the literacy analysis (M = 0.56, SD = 0.80). All descriptive statistics for the two samples can be seen in Table 2.
Demographic Information From the ECLS-B.
Note. ECLS-BC = Early Childhood Longitudinal Study–Birth Cohort; NCES = National Center for Educational Statistics.
Unweighted sample sizes are rounded to the nearest 50, per NCES guidelines.
Descriptive Statistics of Sample for Logistic Regression Analyses.
The dependent variables included in the models were child’s age in months, gender, family SES, race, childcare arrangements, and affective factors based on parent and childcare provider interviews. The descriptive statistics for each of these variables are presented in Table 2. Family SES is a composite variable in the ECLS:B database that included father’s education, mother’s education, father’s occupation, mother’s occupation, and household income (for more detailed information about how this variable was calculated, see Snow et al., 2009). Information about race, family SES, and childcare arrangements were obtained from parent interviews at the 4-year-old wave of data collection.
Affective characteristics
Items used for the factor analysis of affective characteristics came from interviews with mothers (or in a small percentage of cases, resident fathers) and childcare providers. These interviews occurred in the third wave of data collection for the ECLS-B study, and the parent interviews occurred in the same session as the Direct Child Assessments. The items asked parents and childcare providers about affective characteristics and social behaviors of the child and were drawn from the Preschool and Kindergarten Behavioral Scales–Second Edition (Merrell, 2003) and the Social Skills Rating System (Gresham & Elliott, 1990). Items were answered using Likert-type responses, ranging from never (1) to very often (6). Individual items were combined to create factors, representing elements of affective and social development.
ECLS-B Cognitive Direct Child Assessment: Preschool Math
This measure of school readiness and academic ability in mathematics for preschool children was used to identify mathematically gifted children and typical children for the current study. This assessment was developed by the NCES using items from various validated assessments (Snow et al., 2009). The assessment consisted of two parts. The first part, consisting of core items (n = 28 items), was used to route students who scored very low (32% of the sample) or very high (11% of the ECLS-B sample) on the core items for additional testing (Snow et al., 2009). The children in these two groups continued for the second part of the assessment, which consisted of additional lower level (n = 10) or higher level (n = 8) items (Snow et al., 2009). The assessment, including both core and supplemental components, was administered using item-skipping rules, to ensure that no child was requested to answer items regarding lower level skills when he or she had already demonstrated mastery of more advanced concepts (Snow et al., 2009). Thus, no child was administered the entire battery of 28 items. Final scores were determined using item response theory (IRT) to account for skipped items and difficulty levels of items (Snow et al., 2009). For more information regarding the creation of these assessments, see Snow et al. (2009). Even for the highest ability students, a ceiling effect is unlikely for this instrument because less than 0.25% of students answered more than 34 of the 36 items correctly (Snow et al., 2009). Items in the core assessment included number sense, counting (including color manipulatives), geometry, and pattern understanding (Snow et al., 2009). The advanced second stage included additional items concerning operations, such as simple word problems and number sentences (with and without manipulatives). The reliability, as measured by coefficient alpha, was .86 for the core battery and .62 for the supplementary-high form (Snow et al., 2009). The mean score for the weighted sample in the current study was 30.45 (SD = 9.51).
ECLS-B Cognitive Direct Child Assessment: Preschool Literacy
Similarly, giftedness in literacy, as well as the comparison group of typical children, was identified using scores on the literacy ability test given in the third wave of data collection. This assessment consisted of a single form of 37 items across the following domains: letter recognition (n = 8 items), letter sounds (n = 6 items), early reading—recognition of simple words (n = 4 items), phonological awareness (n = 10 items), knowledge of print conventions (n = 8 items), and matching word (n = 1 item; Snow et al., 2009). The test design included rules for item skipping to ensure that students were asked a minimum number of items to demonstrate mastery of the domains, and IRT was used to calculate the final scores. There is little evidence of a ceiling effect on this assessment, as less than 1% of the ECLS-B sample answered more than 33 of the 37 items correctly. The reliability of this assessment over the entire ECLS-B sample was .81 as measured by the coefficient alpha (Snow et al., 2009). The mean of the sample for the current study was 39.80 (SD = 14.20).
Factor Analysis
The factor analysis was conducted to reduce the number of variables to be included in the regression analysis and to make the parent and childcare provider observations more comparable. The factor analysis included all children in the sample who attended non-familial childcare outside the home (including center- and home-based care and Head Start programs), and the unweighted sample size was 5,950 (rounded to the nearest 50 per NCES guidelines; Snow et al., 2009). The parent and childcare provider items were included in two separate factor analyses. The factor analysis used principal components analysis with a direct oblimin rotation and resulted in similar five-factor solutions for each of the analyses, based on the Scree plot and eigenvalues for each solution. Items were retained or eliminated from the final factors based on their pattern coefficients on each factor. Items with an absolute value of the coefficients less than 0.5 on any factor or 0.25 on more than 1 factor were eliminated. After the factor analysis was run, reliability estimates for each factor were calculated.
After the factors were established, the bivariate correlations between parent observations and childcare provider observations were determined. This demonstrated the differences between the affective characteristics of children in home and school settings and context for those observations.
Determining Sample
The next step in the study design was to determine and specify the sample for further analyses. The appropriate weight was applied to the raw data set, resulting in a nationally representative data set. This step is necessary because the NCES originally oversampled small populations such as low birth weight babies, Native American children, and twins (Snow et al., 2009). Because further analysis utilized the ECLS-B Cognitive Direct Child Assessment: Preschool Literacy Assessment, children who did not demonstrate competency in the English language (i.e., could not answer the practice items in English) were not included in the study. In addition, the subsequent analyses used both parent and childcare provider surveys, and children who did not have regular non-familial childcare were not included in the analysis.
To define the gifted group in the areas of mathematics and literacy, the cutoff scores on both direct child assessments (mathematics and literacy) for the 95th percentile were calculated. To define the comparison (typical) groups, scores in the 50th to 55th percentiles were also determined. Weighted cases were then coded to identify cases in the gifted and comparison (typical) groups in both mathematics and literacy. The comparison group was chosen from the middle of the distribution to eliminate possible confounding of data if children close to the cutoff scores were included in the analyses. Thus, the regression analyses compare children in the 95th percentile of scores with children with scores in the middle of the distribution (50th–55th).
Logistical Regression Analyses
After applying the appropriate weights to the data set to more closely represent national demographics and identifying the comparison and gifted groups, two logistical regression models were analyzed to predict, from the variables included in the model, that a child would demonstrate giftedness in mathematics and literacy ability, respectively. Thus, giftedness in literacy or mathematics was the outcome variable of these analyses. Giftedness in mathematics and reading ability were defined as scoring in the 95th percentile on the mathematics ability test or literacy ability test, respectively. These regression analyses included both the affective factors and demographic and background variables about the children. Specifically, both parents’ and childcare providers’ observations of children’s maladaptive behaviors, empathy, concentration, worry, and friendships were included. Demographic variables included gender, age, family SES, race, and childcare arrangements. Child’s age in months and family SES were continuous variables, and all affective factors were treated as continuous variables. The other variables in the equation were nominal. For gender, female served as the reference variable; for race, White served as the reference variable; and for childcare arrangements, home-based care served as the reference variable.
Results
Factor Analysis
The factor analysis revealed a five-factor solution for both parent and caregiver perceptions of child affective characteristics. These factors included socially maladaptive, empathy, concentration, worry, and friendship. However, the factor structure of each of these differed slightly between parents and caregivers. The results of these analyses can be found in Tables 3 and 4.
Structure Matrix for Parents’ Observations of Affective Characteristics.
Note. Italics indicate item was eliminated from final factors.
Item was reverse scored.
The coefficient was less than .001.
Structure Matrix for Childcare Providers’ Observations of Affective Characteristics.
Note. Italics indicate item was eliminated from final factors.
Item was reverse scored.
The coefficient was less than .001.
Socially Maladaptive factor
The Socially Maladaptive factor measured parent and childcare providers’ reports of the frequency of children to engage in socially maladaptive behaviors, such as “temper tantrums,” “annoys other children,” and is “physically aggressive” (see Tables 3 and 4 for a complete list of items from both surveys). The means for the Socially Maladaptive factor for parents and teachers were 2.31 (SD = 0.62) and 2.06 (SD = 0.81), respectively (see Table 5), on a 6-point scale, indicating a fairly low reporting of socially maladaptive behavior overall among the sample. The reliability, as measured by Cronbach’s alpha, was .766 with five items for the parent survey and .822 with four items for the childcare providers.
Reliability Estimates for Affective Characteristics.
Reliability alpha was not calculated for the Worry factor as it consisted of one item.
Empathy factor
The Empathy factor measured parents’ and childcare providers’ reports of the frequency of the child to demonstrate care towards peers. Sample items included, “Child tries to understand others” and “Child comforts other children” (see Tables 3 and 4 for a complete list of items on both surveys). On a 6-point scale, the parents rated children higher (M = 3.67, SD = 0.71) than the childcare providers did (M = 3.45, SD = 0.89; see Table 5). For the parent survey, the internal consistency, as measured by Cronbach’s alpha, was .696 for three items and for the childcare providers, it was .815 for the same three items (see Table 5).
Concentration factor
The Concentration factor measured parents’ and childcare providers’ reporting of the child’s ability to pay attention to tasks and adult directions and task commitment. Sample items include, “Child pays attention well” and “Child has difficulty concentrating” (reverse scored; see Tables 3 and 4 for a complete list of items). The mean parent-reported score was 3.65 (SD = 0.66), and for childcare provider, it was 3.77 (SD = 0.74) on a 6-point scale (see Table 5). The reliability, as measured by Cronbach’s alpha, was .729 with three items on the parent scale and .782 with four items on the childcare provider scale (see Table 5).
Worry factor
The Worry factor measured the parents’ and childcare providers’ reports of the child’s tendency to worry. This was measured by one item on both scales: “Child worries about things” (see Tables 1 and 2). The mean for parents was slightly higher (M = 2.30, SD = 0.66) than for childcare providers (M = 2.13, SD = 0.96) on a 6-point scale, but both represented a relatively low incidence of worry reported among this sample (see Table 5). No reliability was calculated for this factor as it consisted of one item.
Friendship factor
The Friendship factor measured parents’ and childcare providers’ report of the ability of the child to initiate and maintain social relationships with peers. Sample items include, “Child is liked by others,” “Child is invited to play by other children,” and “Child is accepted by other children” (see Tables 3 and 4 for a complete list of items). Overall, parents (M = 4.17, SD = 0.96) and childcare providers (M = 4.33, SD = 0.67) reported high levels on this factor (see Table 5). The internal consistency, as measured by Cronbach’s alpha, for the parent survey was .674 with three items, and for the childcare provider survey, it was .730 with two items (see Table 5).
Correlation Analysis
A bivariate correlation was conducted on the affective factors included in the study, including both parents and childcare providers (see Table 6). The majority of the bivariate correlations were statistically significant (p < .001; see Table 6); however, due to a large sample size (unweighted n = 5,950; rounded to the nearest 50 per NCES guidelines), the likelihood of Type I error is greatly increased (Shadish, Cook, & Campbell, 2001). Therefore, the most useful interpretation of these results is through an analysis of the strength of the correlations, indicating the meaningfulness of the results. For the parent factors, concentration was positively related to Empathy (r = .352) and negatively related to Socially Maladaptive (r = −.501). Empathy (r = −.225) and Friendship (r = −.274) were also negatively related to Socially Maladaptive. Friendship was positively related to Empathy (r = .423) and Concentration (r = .332). Similar patterns among the childcare providers were also seen in the data. The Socially Maladaptive factor was negatively related to Empathy (r = −.257), Concentration (r = −.442), and Friendship (r = −.264). In addition, Concentration was positively related to Empathy (r = .273).
Correlations for Affective Characteristics.
Note. Correlations in bold indicate the correlations between parent and childcare provider observations of similar factors.
p < .001.
The correlations between parents’ and childcare providers’ reports on similar factors were also examined (see Table 6). This revealed relatively low levels of correlations between items that were theoretically measuring the same constructs from two differing perspectives. The Socially Maladaptive factor had the largest correlation (r = .337) between the parents’ and childcare providers’ reports. The Friendship factor was actually negatively correlated (r = −.176). The other factors had low correlations (Empathy, r = .192; Concentration, r = .277; Worry, r = .103). Thus, parents and childcare providers give varying reports of a child’s affective and social characteristics and behaviors.
Logistical Regression: Mathematics
The results of the logistical regression predicting early giftedness in mathematics indicated that age (B = .170), SES (B = 1.42), and race (Not Ascertained, B = 4.365; Pacific Islander, B = −15.287; American Indian, B = −15.242) were significant (see Table 7). After accounting for all of the demographic variables in the model, childcare arrangements and the affective factors were not significant predictors of early childhood giftedness in mathematics. Specifically, older children and children from more affluent backgrounds, as well as children from White, African American, Hispanic, and multiracial backgrounds, were more likely than children from Pacific Islander or American Indian backgrounds to score in the 95th percentile on the ECLS-B Cognitive Direct Child Assessment: Preschool Mathematics as compared with children scoring in the 50th to 55th percentile on the same assessment.
Results of Logistic Regression Predicting Early Giftedness in Mathematics.
White is the reference variable.
Home-based care is the reference variable.
p < .05. **p < .001.
Logistical Regression: Literacy
The results of the logistical regression predicting early giftedness in literacy demonstrated that SES (B = .883), race (Not Ascertained, B = −16.080; Pacific Islander, B = −15.757; American Indian, B = −16.265; Asian, B = 1.118), and childcare arrangements (Center-Based, B = 1.067) were significant predictors (see Table 8). After accounting for all of the other variables in the model, age (B = .053) was not a significant predictor of early childhood giftedness in reading (scoring in the 95th percentile on the ECLS-B Cognitive Direct Child Assessment: Preschool Literacy), as compared with children scoring in the 50th to 55th percentile on the same assessment. Specifically, children from more affluent backgrounds and in center-based care, as well as children from White, African American, Hispanic, and multiracial backgrounds, than Pacific Islanders, American Indians, or children for whom race was not ascertained were more likely to be gifted in the area of literacy. Asian children were more likely than all other ethnic groups to score in the gifted range on the literacy assessment. There were three affective characteristics that significantly predicted early giftedness in reading: parent observations of Concentration (B = .589) and childcare provider observations of Concentration (B = .901) and Socially Maladaptive (B = .593).
Results of Logistic Regression Predicting Giftedness in Literacy (95th Percentile in Achievement).
White is the reference variable.
Home-based care is the reference variable.
p < .05. **p < .001.
Discussion
The results of this study demonstrate four primary findings. Answering the first research question, the factor analysis demonstrated that the factor structure of affective characteristics reported by parents and of those reported by childcare providers were similar. Answering the second question, there were low correlations between parent and teacher observations of children’s affective characteristics. To answer the third research question, age, SES, and race predicted early giftedness in math, but no affective characteristics were significant predictors. However, answering the final research question, SES, childcare settings, and race predicted early giftedness in literacy, but there were affective predictors as well.
Parents’ and Teachers’ Perceptions of Affective Characteristics
The correlations between childcare provider and parent reports of children’s affective characteristics indicate that the context plays an important role in which behaviors are noted or exhibited. The low correlations between parents and childcare providers demonstrate that children do not exhibit the same behaviors in home and school settings, or that childcare providers and parents interpret and perceive these behaviors in different ways. As researchers continue to investigate social and emotional factors in cognitive development and giftedness in early childhood, the results of this study underscore the importance of gathering information from a variety of sources, and the importance of parent and school partnerships for gifted children, particularly in the nomination process (Jolly & Matthews, 2012; Weber & Stanley, 2012).
Affective Characteristics and Giftedness in Mathematics
The results of this study indicate that SES and race are the most important predictors for early precocity in mathematics and literacy. As minority group children (Pacific Islander and American Indian) were significantly less likely to score in the 95th percentile on either the mathematics or literacy assessments, as compared with White children, even after accounting for SES, the talent of many of the country’s youngest children is yet to be developed, a point also highlighted by the NAGC (2011). Programs developed to enhance early childhood education, such as Head Start and other public and private preschool programs, should focus on these demographics of children to provide the resources and environments that would enhance the development of advanced mathematical and literacy skills. Future research should focus on identifying best practices in these environments.
Childcare environments were only a significant predictor of early giftedness in literacy, and not in mathematics, specifically that children in center-based care were more likely to demonstrate early talent in literacy skills than children in home-based or Head Start programs, after accounting for SES and other variables in the model. Because the focus of many early childhood programs is language development in preparation for early literacy, it follows that students in more structured programs at childcare centers would be more likely to demonstrate early traits of giftedness in literacy. Parents should also be informed of the advantages of center-based care for the development of school readiness skills. As Head Start programs tend to serve children who are at greater risk, there may be fewer resources to meet the needs of highly talented children. Advocates for gifted education focusing on early childhood should focus their efforts on providing greater resources for public programs serving children at the most risk and developing guidelines for effective practices.
Finally, much research in the area of gifted and talented education has focused on the social and emotional characteristics of gifted children (e.g., Neihart et al., 2002). Using the logistical regression analysis in this study, it appears that there is little evidence to support the theory that there are differing affective characteristics for preschool children gifted in mathematics. Specifically, children who showed early precocity in mathematics did not display differences in socially maladaptive behaviors, friendships, concentration, empathy, or worry as reported by parents or childcare providers as compared with typical children. This indicates that the findings of Dauber and Benbow (1990) and Brody and Benbow (1986) can be extended to early childhood for mathematically advanced children. If social or emotional differences among this population are detected in later stages of development, these most likely emerge as a result of environmental (e.g., peer groups, school contexts) or developmental (e.g., adolescence) factors, such as those proposed by Reis and Renzulli (2004).
Affective Characteristics and Giftedness in Literacy
The affective predictors for early giftedness in literacy demonstrate a pattern different from mathematics. Specifically, the parents and childcare providers’ perception of a child’s ability to concentrate was a significant predictor of early giftedness in literacy. This finding supports the earlier findings of Bramlett et al. (2000), who also demonstrated persistence as a predictor of school readiness. The perception that giftedness literacy is a learned trait, whereas the perception that mathematics is more innate, as well as actual differences in the acquisition of talent in these areas, may account for the differences in predictors of these two areas of giftedness. In addition, a teacher’s perception of a child’s tendency for socially maladaptive behaviors was also a positive predictor of early giftedness in literacy. Although the children exhibiting giftedness in literacy had a greater ability to stay attentive to tasks and pay attention to the teacher (Concentration factor), they are also more likely to throw temper tantrums and annoy other children (Socially Maladaptive factor) in childcare environments. It could be concluded that children with advanced skills in literacy are more prone to exhibit these maladaptive behaviors in school contexts than at home. One explanation of this pattern of behaviors may be that children with advanced literacy skills become bored with a typical preschool curriculum and the emphasis on basic and foundational literacy skills, and act out in socially maladaptive ways to express their frustrations. Other explanations may include actual differences in the social development of children who are gifted in literacy. This hypothesis is supported by the differences found by Dauber and Benbow (1990) and Brody and Benbow (1986) in adolescence. Further research should focus on this topic in more depth, looking for causes for these differences, and identifying effective interventions for young gifted children.
Implications
There are several implications in this study for researchers, practitioners, and policy makers. First, for researchers in the field of gifted education, the findings concerning the social and emotional characteristics of young children are particularly relevant. The results of this study indicate that for a large, nationally representative sample, there are few affective differences for preschool children who are gifted in literacy and mathematics. Overall, young gifted children show no differences in friendships, worry, or empathy from typically developing children. Differences for children gifted in literacy focus on higher concentration and more socially maladaptive behaviors. Thus, findings regarding few social and emotional differences in older children (e.g., Bain & Bell, 2004; Robinson, 2004) can be expanded to young gifted children.
The difference in patterns of affective characteristics between the two areas of giftedness (i.e., literacy and mathematics) may have various explanations. Mathematical giftedness had no affective characteristics as significant predictors, whereas giftedness in literacy had several, so this may be due to differences in the domains. It could be hypothesized that early giftedness in mathematics is associated with more conceptual and abstract thinking (Sheffield, 2003), whereas early literacy is more reflective of the acquisition of specific skills (Jackson, 1988), and thus, early mathematics ability may be more reflective of later gifted behaviors and follows the pattern of findings from research regarding gifted adolescents (e.g., Jones, 2013). The increased maladaptive behaviors of early advanced readers in classrooms may reflect negative actions in a context focused on reading skills already mastered by this group. Early childhood classrooms are typically focused more heavily on early literacy than on mathematics (Taylor, 2014), so this may also explain the difference in patterns of reported characteristics. Future research should further investigate the differences between the affective characteristics of young children with these domain-specific areas of giftedness.
For practitioners, the differences among gifted children in literacy have important implications. The differences in anger may indicate that these children need more appropriate curriculum to match their abilities. More research should be conducted to investigate the causes of these differences and to investigate whether interventions on social skills are effective for young gifted children. The discrepancy of children identified as gifted in reading from Head Start programs also indicates a need for this program to develop strategies to accelerate the literacy development of at-risk children with high potential. Future research should also determine whether interventions in early childhood have an effect on the identification of children.
Finally, the discrepancies of identification of children from low-SES backgrounds and from Pacific Islander and American Indian backgrounds for giftedness in mathematics and reading have implications for policy makers to provide early intervention programs for children at risk of not developing their full potential. Programs targeting children from poverty should include opportunities to develop talent in literacy and mathematics of young children. In addition, the higher likelihood of children in center-based care to demonstrate giftedness in reading indicates the benefit of early academic experiences for young children. These programs should also target minority-race and low-SES populations (NAGC, 2011), but future research should be directed to see whether specific interventions are effective for high-ability children. Policy makers should be aware of the benefits of such programs and provide these opportunities for a larger population of children.
Limitations
The limitations of this study involve the restrictions involved in using a secondary data source. Primarily, the sample used for this study excludes non-English speakers, as they were unable to complete the literacy and mathematics assessments. In addition, because the analysis utilized data from both parents and childcare providers, it does not include children who did not receive childcare outside the home in the third wave of data collection. The definition of giftedness in this study only incorporates early childhood achievement in mathematical and literacy domains. This may not include many gifted children based on intelligence scores, creativity, or other domains. In addition, there is no information in the database documenting how many of the children identified in this study as gifted are eventually placed in a gifted program. Finally, the regression analyses in this study indicate which characteristics and groups are more likely to score in the top percentiles of early childhood academic achievement, but they cannot determine causation. Thus, no generalizations of causality can be made from these results.
Final Thoughts
In conclusion, the results of this study have identified the predictors of early childhood giftedness from a large, nationally representative sample. In particular, there were few social and emotional differences for young, academically gifted children, but age, SES, ethnicity, and childcare environment were predictors of giftedness in preschool. Children gifted in reading tended to be identified as having greater concentration by parents and childcare providers and greater tendency toward socially maladaptive behaviors by childcare providers. Policies should be developed to increase opportunities for children from poverty to have accelerated curriculum in literacy and mathematics and provide more opportunities for talent development.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.