Cheryll M. Adams
Ball State University
Carolyn M. Callahan
The University of Virginia
While many educators have emphasized the need to identify giftedness in young children, there is seldom a concerted effort to identify primary level children for gifted programs (Clark, 1988; Kitano, 1989; Rubenzer, 1979; Shaklee, 1992; Whitmore, 1986, 1988). One often-cited reason for not acting to identify young children is the inadequacy of identification procedures to evaluate and assess giftedness currently in use in most school systems. The National Association of the Education of Young Children (NAEYC, 1988) has adopted a position statement on Developmentally Appropriate Practices in Early Childhood Programs Serving Children from Birth through Age 8, which expresses concern about the use of standardized testing for placing young children in special programs and the practice of making decisions based on a single score or measure. The position of the NAEYC is based on agreements that instruments used for such selection are not reliable and valid when used with very young children. Further, teachers are often unable to recognize signs of giftedness in young children and continue to select only students who are high achievers in the classroom (Roedell, 1985; Whitmore, 1982).
Another problem facing educators that cuts across identification of gifted students of all ages stems from the failure of traditional assessment instruments to identify gifted students from the population of economically disadvantaged, limited English proficient, and minority children. Educators have been making recommendations for change to address these issues for two decades and agree that direct observations are useful in identification of disadvantaged and culturally diverse learners. Yet, little has been done to validate new forms of assessment. Clearly, there is a need to identify other reliable and valid methods to assess giftedness in young children, particularly those who are culturally different or economically disadvantaged.
Howard Gardner (1983) expands the definition and assessment of intelligence to include seven separate intellectual domains: linguistic, logical-mathematical, musical, spatial, bodily-kinesthetic, interpersonal, and intrapersonal. The major thrust of Gardner’s theory is that individuals tend to have strengths in specific cognitive functions. According to his theory, individuals are capable of exceptional development in any one or a combination of these seven discrete intelligences. Gardner (1989) further cautions that “intelligences must always be conceptualized and assessed in terms of their cultural manifestation in specific domains of endeavor” (p. 6). For example, to assess spatial skills a child might be given a small kitchen appliance or tool from his or her environment to take apart and put back together. One NRC/GT Collaborative School District in Maryland, the Montgomery County Schools, was awarded a Javits grant to pilot an application of Gardner’s theory. The project staff of The Early Childhood Gifted Model Program has developed a Checklist for Identifying Learning Strengths based on the theory of multiple intelligences, a means of searching for the talents of culturally diverse, economically disadvantaged gifted students. Classroom teachers have been trained to use particular tasks to elicit behaviors relating to the specific intelligences and to use the checklist to identify gifted young children for the program. The checklist consists of seven sections, each corresponding to one of the seven intelligences identified by Gardner. Each section is comprised of seven to eleven statements describing ways that intelligence may be manifested in the child. For example, under the verbal-linguistic heading are statements such as, “Enjoys word play;” “Expresses ideas easily, either orally or in writing;” and “Is a good storyteller or writer.” Students high in visual-spatial ability may exhibit characteristics such as, “Chooses to express ideas through visual media;” “Takes things apart and puts them back together again;” or “Can organize and group objects.” The observer gives each domain an overall rating of one (“You have not observed these behaviors”) to four (“You almost always or always observed them”). A five indicates “No opportunity to observe these behaviors” (during data analysis, these scores were dropped). The observer may also check any of the descriptors that may be particularly strong indicators for the child. An overall rating is obtained for each intelligence. There is also a section for the observer to add comments that might help another teacher plan for the child.
The NRC/GT staff has been collaborating with the staff of the Early Childhood Gifted Model Program in establishing the psychometric properties of the checklist. First, a reliability study was undertaken to establish intrarater reliability and stability for the checklist. In Round One all 365 students in kindergarten through second grade in the schools participating in a pilot study were rated by teachers who had received training in the use of the scales. One month later the names of 10 students were randomly selected from each classroom. These students were rated again by the rater who had observed them previously. One hundred thirty-six students were included in this process.
When the same teacher rated the same child after a one-month interval, the intrarater reliability for kindergarten students were moderately high (ranging from .713 on the logical-mathematical scale to .782 on the spatial scale). Correlations across the two ratings for first grade scores ranged from .496 (music) to .775 (interpersonal). At the second grade level, intrarater reliability ranged from .681 (bodily-kinesthetic) to .811 (linguistic).
These intrarater reliabilities are not high enough to warrant placement decisions about individual children on the basis of the checklist scores alone, but they are reasonable for considering modification of instruction in conjunction with other data a teacher has about the child’s achievement. The reliabilities are also sufficiently high to warrant further investigation. We, therefore, looked to see if the seven domains were independent. As expected, and as preliminary evidence of construct validity, scores across domains were not highly correlated with each other. Each domain appeared to be measuring attributes that were unique.
Currently, we are analyzing additional data to establish inter-rater reliability as well as the relationship between this instrument and other measures of intelligence.
The results of the study support Gardner’s assertion that the domains appear to be discrete. At this time, teachers in the project are using the results to focus activities for the children by differentiating the curriculum according to an individual child’s identified strengths.
For further information about the checklist contact: Dr. Waveline Starnes, Montgomery County Public Schools, 850 Hungerford Dr., Rockville, MD 20850