Finding High Potential among Culturally, Linguistically and Economically
Diverse Students: Two New Scales for Equitable Identification

Maggie Smith-Peterson, Minneapolis Public Schools
Kelly Stewart, Minneapolis Public Schools
Karen Westberg, University of St. Thomas


In the 2007 report, Overlooked Gems: A National Perspective on Low-Income Promising Learners (Stambaugh & VanTassel-Baska), demographer Harold Hodgkinson explained that within a decade, no single ethnic group in the U.S. will predominate among public school students. In other words, schools in the U.S. are becoming more racially, culturally and linguistically diverse every year. Given this fact, it is alarming to note that many public school districts across the country underrepresent students of color, English learners (EL), and students qualifying for free or reduced lunch plans among their identified gifted population (Callahan et al., 2014). Furthermore, gifted and advanced learners who are culturally, linguistically, and economically diverse (CLED) consistently perform below their white, middle- and upper-class peers (Plucker et al., 2010). The use of standardized tests to identify gifted CLED students will clearly fail to produce an equitable result. In a recent national survey, fifty-one percent of elementary programs reported having a plan in place to develop talent among traditionally underidentified groups (Callahan et al., 2014). Yet, when it comes to useful tools to accurately and efficiently identify high potential CLED students, the field is at a loss: very few tools have been created that specifically target the identification of gifted CLED students, and among those that do exist, even fewer have the sufficient research to prove their efficacy.

The use of teacher rating scales to identify gifted learners is a long-standing practice. Whereas much evidence shows that teacher bias limits the participation of CLED students in gifted programs (Bruch, 1975; Deslonde, 1977; de Wet & Gubbins, 2010; Ford & Grantham, 2003), other studies have shown that teacher rating scales can improve accurate identification of talent, especially when paired with quality training (Frasier et al., 1995; Hoge & Cudmore, 1986; Frank, 2007; Swanson, 2006). While a few tools, such as the HOPE Scale (Gentry & Peters, 2010) have demonstrated both reliability and validity evidence in identifying low income students, we saw a need to develop teacher rating scales that specifically addressed aspects of language and culture. Additionally, the Scales for Rating Behavioral Characteristics of Superior Students (otherwise currently known as the Renzulli Scales) address more universal characteristics of giftedness, and therefore, we sought to create teacher rating scales for CLED students that were not redundant with these more generalizable items. These new scales are called the High Potential Culturally and Linguistically Diverse Scale and the High Potential Culturally and Economically Diverse Scale.

Instrument Development

The items on the High Potential CLED Scales synthesize many of the behavioral characteristics cited in empirical and theoretical literature focused on gifted students from these populations (see Tables 1 and 2). A few items are verbatim from the literature, while others summarize behavioral characteristics cited in multiple studies focused on CLED gifted learners. After conducting a thorough review of the literature focused on behavioral characteristics of CLED gifted learners, we organized the resulting items into two scales: the first, the High Potential Culturally and Linguistically Diverse (CLD) Scale includes items focused on aspects of language and culture, and pertains to high potential among English Language Learners; the second, High Potential Culturally and Economically Diverse (CED) Scale focuses on characteristics related to class and economic disadvantage among high potential students. Both scales included items that describe aspects of culture, such as behaviors related to acculturation and the development of a cultural identity. In our original study, each of the scales included 18 items. We arranged the items with a 6-point frequency response scale, which is common in many teacher rating scales for gifted identification, such as the Renzulli Scales and the HOPE Scale.

Instrument development takes much time and effort. In order to test the psychometric properties of any new tool, researchers must gather many samples from a wide swath of the population. In our case, we asked teachers across the country to fill out the High Potential CLED Scales on many different kinds of learners in their classroom and collected information on students’ economic status, participation in English language learner services, and academic performance to ensure we had a truly heterogeneous group of students in our sample. We were then able to match individual teacher-completed scales to either the High Potential CLD Scale for students who were culturally and linguistically diverse, or the High Potential CED Scale for students who were economically disadvantaged. We examined the results by performing a variety of statistical procedures with various cross sections of our data, including exploratory and confirmatory factor analysis, cross-correlation analysis, and various criterion validation procedures. After performing these and other statistical procedures, each scale has been refined to include only ten of the original eighteen items.

After more than three years of instrument development research, our results indicate we have developed scales that can be used to accurately support equitable identification practices. These results have prompted us to share the scales with others in the field seeking to increase identification of CLED students for advanced learners and gifted services. Both of the scales have yielded strong reliability (Cronbach’s alpha = 0.87 for CLD scale and Cronbach’s alpha = 0.95 for CED scale). The scales have demonstrated good model fit, indicating that each scale measures a unified concept while still capturing the various aspects within it (e.g., the CED scale measures both interpersonal and intrapersonal aspects of behaviors typified by CED students with high potential). Importantly, by correlating completed scales from our sample with Cognitive Abilities Test Form 7 scores, we have been able to demonstrate that the items on the scales relate to strong performance on a measures of high potential (the subscales and composite scale scores on the CogAT 7), and therefore distinguish between CLED high potential students and their peers (for a more technical description of our methods and results, please see our presentation at NAGC in November).

Practical Considerations

Institutions using teacher rating scales for identification of students for gifted programs have generally implemented them in one of two ways: either as a tool for nominating students for identification screening, or as part of a multiple measures identification process. In both cases, the first and most important step educational leaders should take is to train teachers in the use of the High Potential CLED Scales. As mentioned, research shows that educators who have not been trained to recognize the characteristics of gifted students who are CLED tend to default to a more mainstream (and therefore culturally limiting) conception of giftedness. Thus, training classroom teachers in their content prior to implementing the High Potential CLED Scales is essential. Teachers can work together to generate both classroom examples and non-examples of the behavior each item describes. They can also be guided to relate concepts from the literature about high potential CLED learners with the items on the scales. Although a formal teacher training manual has yet to be developed for these scales, opportunities to generate and relate classroom examples to the items on the scales will increase inter-rater reliability and the efficacy of the scales for identifying CLED students with high potential. Furthermore, because the items on the scales have been culled together from the literature, a good deal of reference material is available for teachers and administrators to review and apply to professional development opportunities (see citations, Tables 1 and 2).

Assessors can use the High Potential CLED Scales as part of a multiple measures identification protocol by calculating a total score and creating local norms for each scale. Individual scores can be calculated for each scale by adding the numerical responses for each item into a total score, and these total scores can be used to calculate a local norm for the district, school, or institution. Thus, students can each receive a percentile rank relative to their local peers. By determining an acceptable range for identification of above average potential (such as, above the 7th stanine, or 77th percentile), assessors can efficiently determine CLED students who might benefit from services in conjunction with other assessment data. In our research process, we used specific criteria for each scale: students who qualified for free or reduced-price lunch were eligible to receive the CED scale, whereas students who qualified for English language learner services were eligible to receive the CLD scale. In the state of Minnesota, classroom teachers are not given access to students’ economic information. Therefore, an advantage of using the High Potential CLED Scales as part of the multiple measures identification protocol includes the ability to disaggregate student data at the district level to determine economic status.

As a tool for talent spotting or nomination of high potential CLED students, the scales might be implemented at the beginning of the school year as a tracking device: after having been trained in the CLED characteristics and corresponding items on the scales, teachers can regularly record the frequency of CLED student behaviors using the scales. This practice may be especially useful in the identification of English language learners because teachers both know who these students are and often struggle to observe high potential due to language barriers. Although this is an area for continued research for us, we have already seen interesting results when EL teachers work alone or in collaboration with classroom teachers to complete the CLD scale, including a more sophisticated interpretation of the items on the scale.

Finally, a well-designed identification protocol should always match the services being provided to identified students. Schools, districts and other institutions interested in identifying more CLED students for gifted programming must design options that provide language scaffolds, culturally relevant curriculum, and other supports related to the needs of CLED students (NAGC, 2012). Fortunately, districts and researchers have begun to innovate replicable models that practitioners can turn to when designing strong programming options for high potential CLED students, including the Young Scholars Program, the School-wide Enrichment Model, and many of the Javits-funded projects that have been created over the past twenty years. Furthermore, gifted educators can collaborate with EL teachers, equity and diversity specialists, bilingual education staff and others to combine practices from each field when designing programming.

Our hope is that the High Potential CLED Scales can contribute to efforts across the country to include diverse students more equitably among identified populations. Throughout this process, we were struck by the need for more quantitative studies related to CLED learners. Although equitable identification has been an important topic in the field of gifted education for several decades, it is surprising to note that much of our understanding of the characteristics of CLED talented learners are based on anecdotal evidence. This is partly due to the complexity of such a demographic, wherein specific cultural groups present very different behaviors based on social norms and other factors (Esquierdo & Arreguin-Anderson, 2012; Irby & Lara-Alecio, 1996; Ryu, 2004). Further research on the High Potential CLED Scales would include more focused analysis on different cultural groups with an aim to investigate if items are generalizable to the broad category of CLED despite cultural differences. Additionally, differences in teachers’ cultural and linguistic backgrounds, as well as training, years of service, and other demographic data should be investigated to further develop and make recommendations about applications of the scales. Other areas of inquiry include collaborative efforts between EL teachers and classroom teachers to complete the scales, as well as various methods of teacher training on the scales. In the world of instrument development the research process can often span many years, sometimes decades, as tools are continually revised and investigated in myriad contexts. At this time, we are pleased to present a tool with strong validity and reliability support for general use in the identification of CLED students for gifted and advanced programming.

*Special thanks to Melanie Crawford, Director of Talent Development and Advanced Academics for Minneapolis Public Schools, who supported, encouraged, and inspired this work from the beginning.