Abstract
Countless characteristics lists have been made to describe gifted children. But what evidence exists to support them? If such lists are to be useful, they must be appropriately contextualized and grounded in empirical support. Lacking these, they cannot be useful and many existing lists are severely lacking in both of these things. In this article, I first provide background on characteristics lists and their uses. Second, I outline six limitations of current lists. Third, I introduce a formal nomenclature for determining what constitutes a characteristic of gifted students. Finally, I propose two possible paths forward. First, stop creating or using characteristics lists. Alternatively, if characteristics lists are to be created and consumed, they need to better align the field’s actions with its aspirations. Without sufficient empirical support, characteristics lists will not help schools and can exacerbate both inequity and distrust in research. Calling something a characteristic is a privilege that must be empirically earned.
Plain Language Summary
Many characteristics lists of gifted children exist online, on school websites, and in textbooks. In this article, I first provide background on these lists and their uses. Second, I outline six problems with current lists. Third, I introduce a system for deciding what should be considered a characteristic of gifted students. Finally, I recommend two possible options. One is to stop using characteristics lists. The other is a set of steps for how characteristics lists can be created and used. These steps will help align lists with standard research expectations. Without evidence to support them, characteristics lists will not help schools. They may make things worse by creating unfair outcomes and causing people to lose trust in research.
Millions of students are assessed each year across the globe to be identified as academically gifted and talented. How this is done varies. Depending on the country, gifted identification can be determined by federal, state, and/or local rules and definitions. Regardless, we know it is often done poorly, leaving many students in learning contexts ill-suited to help them. This is demonstrated by massive under-identification of students from some demographic groups, particularly students from racial minority families, students with limited English proficiency, and students with disabilities (e.g., Grissom & Redding, 2015; Peters et al., 2018). Incorporating characteristics lists of gifted students into the identification process has been proposed as a tool to help support both identification and understanding of gifted children (e.g., Rimm et al., 2018; Webb et al., 2007). If the goal is to identify students who would benefit from services being provided, do characteristics lists help? Or are they an empty promise?
With this article, I seek to catalyze a conversation about lists of characteristics of gifted students with two goals. First, so that characteristics lists can be consumed by informed users. Second, so that the research community can more accurately describe the empirical support (or lack thereof) of current characteristics lists and their applications. I argue that characteristics lists currently lack sufficient empirical support to be useful and trusted tools.
In this article, I define how the term characteristic has been used and introduce current characteristics lists and their purported uses. Next, I outline limitations of current characteristics lists and their items. Third, I introduce a formal nomenclature for determining what constitutes a characteristic of gifted students. Finally, I propose two potential paths forward. One is to stop using characteristics lists. As an alternative, I introduce some next steps for how characteristics lists can be created and consumed that will help better align the field’s actions with its research values to improve its ability to help gifted students and their families. Relying on tools that don’t work perpetuates problems and wastes resources.
What Is a Characteristic?
In this section, I describe how the term characteristic has been used. Outside of gifted education, educational psychologists have described characteristics generally as “. . . positive and negative aspects of a student, including their personal qualities, attributes, abilities, temperament and nature” (Allen et al., 2018, p. 4). General examples of positive characteristics include self-efficacy, conscientiousness, and coping skills; negative characteristics include anxiety and depression. Features such as age, gender, and race/ethnicity might be considered demographic characteristics. This is also how the U.S. Department of Education (n.d.) used the term characteristics in its overview of building evidence for educator effectiveness. In this view, the term characteristic appears interchangeable with a generic descriptor.
Within gifted education, the term characteristic is not typically used as a generic descriptor. Instead, there appear to be two consistent facets of characteristics. They describe what is “usual” for gifted children as well as what makes them “unique” from other children (e.g., Rimm et al., 2018, p. 23). In this way, characteristics of gifted students are framed as typical in gifted students and rare in typical students. And a list of these items is proposed to be useful in distinguishing the two groups from each other. To avoid confusion with how the term characteristic is used in gifted education, I adopt the use of the generic term “feature” when talking about any descriptor (e.g., age, eye color, personality type). Thus, a feature might be considered a candidate characteristic. Only features that are usual and unique among gifted students would be considered characteristics.
Other descriptors like trait and behavior may also be used when discussing characteristics lists and their items (e.g., “traits and characteristics”; Callahan, 2018, p. 153; “behavioral characteristics”; Renzulli et al., 1971, 2009). Regardless, there is a conceptual consistency in what is being discussed. Characteristics describe what is usual within gifted students and what makes them unique from typical students.
The use of the term characteristic about gifted children goes back at least a century to Terman’s (1925) first volume of his Genetic Study of Genius: Mental and Physical Traits of a Thousand Gifted Children. Although using the term traits as the titular descriptor, the text in the book appears to use the term characteristic interchangeably (e.g., “It is obvious that conclusions regarding the characteristics of gifted children in general will be valid only to the extent to which the latter requirement [representative of all gifted children] has been met”; [p. 19] and “What are the outstanding characteristics of this group of gifted children?” [p. 633]).
Sample Characteristics Lists
Characteristics lists abound. Describing every list that appears online is beyond the scope here. Rather, I focus on lists that appear in textbooks and on the webpages of influential organizations. I have purposefully selected examples that are from respected sources to minimize concerns about surmounting the extremely low bar of finding something wrong on the internet. Characteristics lists or reference to them can be found on/in:
government agency websites (e.g., Alberta Education, 2011; Colorado Department of Education, n.d.; New South Wales Department of Education and Training, 2004)
school district websites (e.g., Denver Public Schools, n.d.; Montgomery County Public schools, n.d.);
advocacy groups (e.g., American Mensa, n.d.; California Association for the Gifted, n.d.; Davidson Institute, n.d.; National Association for Gifted Children [NAGC, n.d.-a, n.d.-b, 2019, 2020], Pennsylvania Association for Gifted Education, n.d.);
popular press books (e.g., Silverman, 2012; Webb et al., 2007);
gifted education textbooks (e.g., Callahan & Hertberg-Davis, 2018; Clark, 2008; Rimm et al., 2018).
Perhaps the most widely used/cited list of gifted characteristics is from Clark’s (1979/2008) textbook Growing up Gifted, first published in 1979 with a seventh edition published in 2008. Groups in several states in the United States (e.g., California Association for the Gifted, n.d.; Colorado Department of Education, n.d.) have documents based on Clark’s list. Her list has also inspired the Ontario Association for Bright Children’s (n.d.) guidance on understanding bright children, a resource guide for teachers in the New Brunswick Educational Services Division (2007), a handbook for teachers in the Newfoundland and Labrador Department of Education (2013), as well as a “fact sheet” by the Canadian Psychological Association (2024). Her list is also featured on the New South Wales Department of Education and Training website in Australia.
Clark’s (2008) textbook contains four lists of characteristics (cognitive function, affective, physical/sensing, and intuitive). Each list contains three columns: “differentiating characteristics,” “examples of related needs,” and “possible concomitant problems” (p. 74). (A fifth list of common characteristics of highly gifted individuals [p. 83] contains only a list of descriptors without the latter two columns). Items in her cognitive function characteristics list include “advanced comprehension,” “flexible thought processes,” and “an evaluative approach toward self and others” (pp. 74–75). For “advanced comprehension,” the related need was “to be given access to challenging curriculum and intellectual peers,” while the possible concomitant problem was “poor interpersonal relationships with less able children of the same age; adults consider a gifted child ‘sassy’ or a ‘smart aleck’; a dislike of repetition of already understood concepts” (p. 74).
In the seventh edition of her textbook, Clark (2008) credited her lists as being “an extension” (p. 73) of Seagoe’s (1974) work that appears in a report from a California school district (Martinson, 1974; all mentions of Seagoe and Martinson were removed in the eighth [and final] edition of Clark’s textbook, although the characteristics lists remain largely unchanged). Seagoe’s list consists of 16 characteristics that each has a concomitant problem (e.g., the characteristic “keen power of observation; naive receptivity; sense of the significant; willingness to examine the unusual” has the concomitant problem of “possible gullibility,” p. 28).
What is the empirical basis for such an influential characteristics list? The sole citation for the source of Seagoe’s (1974) list is a prior report to the California state legislature by the same editor (Simpson & Martinson, 1961), for which Seagoe served on the advisory committee. This earlier report described a study of students in California who had been identified as gifted by scoring at least 130 on an IQ test by 1957 (when the study began). However, students were only selected to take the IQ test based on prior assessment and teacher evaluations. 1 No other research studies are cited by either Clark or Seagoe.
Themes Across Characteristics Lists
Other lists cite Clark (2008) as a source for their lists (e.g., Callahan, 2018; Rimm et al., 2018; Webb et al., 2007). As such, it should not be surprising that several unifying themes exist across many characteristics lists. I list four such themes in this section to give readers a greater sense of familiarity with characteristics lists of gifted students.
Caveat That Not All Gifted Students Exhibit All Gifted Characteristics
Several caveats are often given when characteristics are described. For example, list-makers (e.g., Alberta Education, 2011; Callahan, 2018; Rimm et al., 2018; Webb et al., 2007) note that gifted children are not all the same, that not all gifted children exhibit all characteristics, and that lists like theirs are common throughout the field. Rimm et al. (2018) noted that “Most of the descriptions that follow are ‘usual’ characteristics, traits that have appeared and reappeared in studies of gifted children and adults. All traits will not and cannot apply to each and every gifted and talented student” (p. 23). Callahan listed five general underlying principles about characteristics that included, “Principle 4: Not all gifted students in a given domain exhibit all of the characteristics of giftedness in that domain. And not all gifted students will exhibit the traits that characterize them as gifted all the time” (p. 154). Perhaps most famously, Renzulli summarized this caveat as giftedness “is developed in certain people (not all people), at certain times (not all the time), and under certain circumstances (not in all circumstances)” (Renzulli, 2004, p. xxvii).
Acknowledgment That Both Positive and Negative Gifted Characteristics Exist
When thinking about characteristics of the gifted, positive aspects of giftedness may spring to mind. However, lists may also describe negative characteristics. Different lists handle this in different ways. Some lists may group all items together, such as “rapid learner” and “desire to organize things and people through complex games or other schemas” (Webb et al., 2007, pp. 11–12). Other lists may include a description of how each item may come with problematic components (e.g., Callahan, 2018; Clark, 2008; Seagoe, 1974). This description can include consequences of the characteristic (e.g., the item of “power of concentration” can have a consequence of “resistance to interruption”; Seagoe, 1974, p. 20) or social consequences (e.g., the item “high level of language development” can have a consequence of “perception as a ‘show off’ by children of the same age”; Clark, 2008, p. 74). Other lists include items that are specifically labeled as negative characteristics (e.g., “perfectionism, which can be extreme” or “depression”; Rimm et al., 2018, p. 24; “impatient with slowness of others”; Webb et al., 2007, p. 26). Silverman and Golon (2008) go so far as to assert that, “Traits that may be viewed as dysfunctional—intensity, sensitivity, perfectionism—need to be seen as typical manifestations of this population” (p. 219).
General vs. Specific Types of Gifted Characteristics
Gifted characteristics lists can vary in scope. Some lists include items across a variety of domains (e.g., Rimm et al., 2018, p. 24; Webb et al., 2007, p. 11). Other lists focus on specific domains, such as separate lists for the cognitive, creative, language, or music domains (Callahan, 2018; Renzulli et al., 2009). Clark (2008) included lists specifically on cognitive characteristics, affective characteristics, intuitive characteristics, and physical/sensing characteristics, as well as a list that focuses on the highly gifted.
Implicit Dichotomization of Continuous Variables
Existing lists also regularly treat features that vary on a continuum (e.g., “The ability to process information quickly”; Johns Hopkins Center for Talented Youth, 2022, para. 5; “Thinking is abstract, complex, logical, and insightful”; Webb et al., 2007, p. 11) as a dichotomized yes or no. Although it is possible for features to be dichotomous (e.g., can correctly spell the word platykurtic; knows the capital of Alberta), these are often reframes of continuous features (e.g., spelling performance; geography knowledge). Although many items convey continuous relative differences (e.g., “Unusual alertness as early as infancy”; Webb et al., 2007, p. 11), this appears to be relatively rare.
Use of Characteristics Lists
Once a characteristics list has been created and shared, there is essentially no cost to further disseminate or use it. In this section, I list five ways that both list creators and list posters frame how they believe such lists should/can be used.
Explain “Who the Gifted Are”
At times, lists are framed as tools to help consumers better understand who gifted children are. For example, a post by the Davidson Institute (n.d.) asserts, “At its core, giftedness is a brain-based difference that contributes to our vibrant and neurodiverse world. This neurological difference means that profoundly gifted students experience a different intellectual, academic, and social-emotional development trajectory than neurotypical individuals” (para. 2). Similarly, Winner (2014) stated that “Gifted children have three atypical characteristics: they are precocious, they learn differently from typical children (marching to their own drummer), and they are intensively motivated to learn (showing a rage to master)” (p. 297). From this perspective, to know the characteristics is to know the child. This perspective may align clearly within the gifted child paradigm (Dai & Chen, 2013).
Support Parent Understanding
Another proposed use of characteristics lists is to support parent understanding of gifted children. The Davidson Institute (n.d.) states, “Looking for gifted traits in children can provide information for parents, educators, and students themselves to decide whether they want to pursue intelligence testing, acceleration, or simply have a better understanding of who these children are” (para. 1). Similarly, Webb et al. (2007) encouraged parents to consider the “behaviors” (p. 12) included in their list of common characteristics of gifted children. The authors encourage this because they believe that parents may overlook—or even resist—signs of giftedness in their child. On the contrary, Rimm et al. (2018) note that parents may even go so far as to too easily accept “undesirable characteristics in the name of giftedness” (p. 39) such as accepting arrogance and talking too much out of a belief that they “come with the territory of giftedness” (p. 31). Meanwhile, Smutney (2011) asserted that parents see characteristics better than others, “But parents see much more in their children—the exceptional ability, yes, but also their heightened sensitivities, intuitive understanding, empathy far beyond their years; also, their untraditional ways of learning. Taken together, these characteristics can present special challenges in school” (p. 37). Despite the differing views on what parents do or do not see in their children, all these perspectives share the belief that familiarity with characteristics can help support parental understanding.
Part of the Identification Process
Similarly, as stated in the above example from the Davidson Institute, a third way characteristics lists are framed/used is as formal or informal identification tools. Formally, characteristics lists or teacher rating scales can be part of a school’s identification process. Informally, they can be used to introduce teachers and parents to the world of gifted education. For example, Rimm and colleagues (2018) stated that, “Identifying characteristics of gifted students is important because it helps teachers and parents recognize and understand gifted children” (p. 38). Similarly, Silverman et al. (1986) concluded that their list helps parents and that, “Earlier identification by parents, supported through professional testing, could lead to appropriate early intervention” (pp. 35–36). When used this way, lists could direct people to action. This action could be as simple as pursuing more information. Or it could be as formal as too few items checked means a student will not be considered for a service.
Learning Needs
A fourth framing used for characteristics lists has been as descriptors of the learning needs of gifted students. For example, a chapter describing gifted students from Alberta Education (2011) pairs characteristics with specific learning needs (e.g., “avid reader” and “access to diverse reading materials” [p. 96]). Clark’s examples of related needs serve a similar purpose (e.g., “ability to predict; interest in future” paired with “to be provided with opportunities for exploration of ‘what if’ questions and activities of probability and prediction”); (p. 79). In addition, in a position paper, the California Association for the Gifted (n.d.) stated, “To provide appropriate education for gifted children, it is important to understand the characteristics of giftedness” (p. 3). All of these are examples of how characteristics lists are framed as helping meet the learning needs of gifted children.
Advocacy in Support of Gifted Education Broadly
Finally, a fifth framing use for characteristics lists has been to help advocate for the existence of gifted education. For example, the California Association for the Gifted (n.d.) stated, “The California Association for the Gifted believes that achievement tends to be higher, and self-efficacy and self-esteem more healthy when parents and educators understand these characteristics and provide appropriate environments for gifted students” (p. 5). Similarly, the use of characteristics is common throughout NAGC documents. For example, in the “when to test” section of its website, NAGC (n.d.-a) recommended that “For younger children, alternative measures of giftedness include characteristic checklists . . . .” In the “identification” section, NAGC (n.d.-b) lists “gifted characteristic checklists” under the Nominations section as well as including several specific characteristic checklists under subjective measures that can be used for identification. Characteristics also come up in NAGC’s TIP sheets that aim to help parents and practitioners. For example, the Serving the Whole Gifted Child TIP sheet states, “Empowering the whole gifted child requires that we look beyond test scores and consider gifted traits, characteristics, and behaviors” (NAGC, 2020, p. 1). The Talking to Your Child about Giftedness TIP sheet includes a suggested conversation starter of “Does your child exhibit characteristics common with gifted children? Discuss traits and behaviors associated with gifted individuals, and how these characteristics may affect interactions with others” (NAGC, 2019, p. 2). These examples are not meant to be a systematic assessment of the use of characteristics lists by advocacy groups like NAGC. Rather, they demonstrate the regular use of a framing that builds on three assertions: (1) characteristics of gifted students exist, (2) we know what those characteristics are, and (3) that knowing about them can/should catalyze action.
Limitations of Characteristics Lists
Despite having a breadth of suggested uses, characteristics lists have several limitations. These limitations may make them lack sufficient quality to accomplish any of the aforementioned suggested uses. Some of these limitations have been briefly noted previously (e.g., Ford & Whiting, 2008; Rinn & Majority, 2018). In this section, I describe six limitations in detail.
Lack of Empirical Evidence
One limitation is a lack of empirical support for the lists supporting understanding the difference between a child who is gifted versus one who is not. Without empirical support, lists run the risk of inhibiting successful identification and weakening any claims of empirical basis for the field.
Currently, there are no formal requirements or guidelines for creating characteristics lists. Without empirical evaluation, it is difficult to differentiate a list written with nothing more than good intentions from a list with a strong empirical basis that reliably and accurately accomplishes its intended use. Existing lists often provide an anecdote such as a story of a child having a sense of humor to support items (e.g., Rimm et al., 2018; Webb et al., 2007). Such anecdotes may vividly illustrate what such an occurrence may look like in a child. But they do not establish evidence of whether such occurrences are the exception or the norm.
The inclusion of citations can add a veneer of credibility but does not guarantee it. For example, when describing her characteristics list, Silverman (2014) stated that “The characteristics in the scale also have been supported by other experimental and clinical studies, and in the professional literature” (p. 3). However, upon inspection, many citations are not primary research; they are handbooks, summaries of research, or marginally related studies. The first item of her list, “Good problem solving/reasoning abilities” (p. 3) is illustrative. Four references are cited: Davidson, 1986; Keating & Bobbitt, 1978; Parkinson, 1990; Sternberg, 1986. The first is not an empirical article presenting data; it is a chapter in an edited volume on conceptions of giftedness. The second citation reports three experiments with comparisons irrelevant to giftedness and problem solving. The comparisons were on reaction times, sorting cards (comparing “AA” to “Aa”), and a memory task of numbers with one, three, or five digits. These are not examples of solving problems. In fact, the study does not mention giftedness or reasoning ability. If anything, the study reports how students of different problem-solving levels differ while performing simple tasks. The third citation is from a newsletter edited by Silverman that I could not find. A contemporaneous review of the newsletter portrays it as targeting parents and practitioners and “an excellent introduction to the psychology and education of the gifted for those in education fields or for anyone with an interest in these children (Silverman recommends sharing the newsletter with your pediatrician)” (Clinkenbeard, 1989, p. 125). The final citation is a chapter discussing a theory; it reports no new data. Thus, an item that appears to have four independent sources of support has no clear direct empirical support provided.
This type of referencing does not appear to be an outlier. Other write-ups by Silverman (e.g., Silverman & Golon, 2008) about other iterations of her list share similar citation patterns. For example, the five citations provided by Silverman and Golon (2008) consist of two links to webpages of the Gifted Development Center (run by Silverman), two handbook chapters by Silverman, and a book for educators and counselors that mentions checklists generally. Such window dressing is not limited to one example. Similarly, textbooks that contain chapters on characteristics of gifted students (e.g., Callahan, 2018; Rimm et al., 2018) also contain similar citation patterns. They contain a long string of citations that include unpublished manuscripts, links to websites, review articles, case studies, and other handbooks. 2
The inclusion of citations likely demonstrates a belief that having empirical support is beneficial for characteristics lists. However, when the reference provided does not directly link readers to the primary source of empirical support, the situation exemplifies what I call a “handbooks all the way down” problem (cf. the infinite regress problem of “Turtles All the Way Down,” 2025). If references supporting characteristics lists consist of citations of handbooks that in turn cite older handbooks, that cite even older handbooks, there is no empirical foundation provided. Without empirical foundation, the assertions approach the Argument From Authority (2025) logical fallacy or the armchair best thinking that Callahan and Moon (2007) warned against when evaluating the research literature. They go so far as to assert that speculative/anecdotal research, “. . . may serve as a starting point, but it does not provide strong and generalizable evidence that can be used for decision making” (p. 306). That is, an assertion is not the same thing as an empirical finding. Instead, Callahan and Moon argued that “We must challenge researchers to rise to the occasion and create and report studies that yield valid empirical or evidenced-based information” (p. 316). Has the field of gifted education done this? Is there credible evidence supporting lists of characteristics of the gifted? The citation behavior I describe above suggests that many lists lack credible empirical evidence.
I do not mean to dismiss the value of review articles and handbooks. They can provide enormous value when describing overall trends. However, they add a layer of potential distortion between the reader and data. The American Psychological Association (APA) states that a “basic principle” of academic writing is to “Cite primary sources when possible, and cite secondary sources sparingly” (APA, n.d.-a) and that secondary sources should be cited, “when the original work is out of print, unavailable, or available only in a language that you do not understand” (APA, n.d.-b). To me, this means that primary studies are recommended when making a list of references to serve as the foundation for the empirical support of a characteristics list. Without direct citations to primary research, empirical support becomes clouded or (worse) exaggerated.
Lack of Alignment With Definitions and Domains
A second limitation of characteristics lists is their lack of alignment to specific definitions of giftedness. Without such alignment, lists run the risk of mis-describing students who may not benefit from the specific service that is going to be provided. Calls for aligning research studies with specific definitions of giftedness are growing (e.g., Rinn, 2024a, 2024b). For example, some definitions (e.g., Subotnik et al., 2011) highlight giftedness as a developmental trajectory, indicating that giftedness will appear differently at different developmental stages. In this view, a characteristic at one developmental stage may appear different than further along the developmental trajectory.
As Callahan (2018) has noted, “. . . any consideration of the characteristics of gifted and talented students must be a multifaceted discussion to incorporate this range of conceptions and definitions of giftedness” (p. 153). Because some definitions of giftedness are broader than others, this opens the door for different sets of characteristics. For example, creativity is part of the definition and gifted identification process in some states (e.g., Georgia). In these states, seeing high rates of creativity in gifted students would not be surprising; it is mandatory given the definition of giftedness in those states. In this case, creativity is not a characteristic of giftedness; it is literally a component of the definition of giftedness. This would be like calling excessive height a characteristic of being tall. It is explicitly a part of the definition. But for other states that do not have creativity as part of their definition, seeing high creativity in their gifted students would not be circular logic. Thus, different definitions can lead to inclusion (or exclusion) of not just different students but also different characteristics.
This need for alignment is akin to how a list of physical characteristics of sprinters differs from one for long-distance runners or even swimmers who race in sprints instead of running. They describe athletes who differ from each other in meaningful ways. None of these lists would be “wrong” while having different value depending on the similarity between how the list was developed and how the list was being used. If the list was called “characteristics of athletes” but only sprinters were observed to create the list, it might be useful for track coaches while being a lousy tool for volleyball coaches. Both sets of coaches are looking for good athletes but have different definitions of what constitutes a good athlete for their sport. The same is true for different definitions and domains of giftedness; the differing definitions lead to different types of students being identified. These different types of students may have different characteristics from each other. This lack of alignment to specific definitions and specific domains means generic characteristics lists (or lists used with different definitions) may describe different groups of students than who would be described by a specific definition of giftedness.
Lack of Alignment With Identification and Selection Practices
A third limitation of characteristics lists is their lack of alignment with specific identification practices. Even within the same definition/domain of giftedness, different identification/selection criteria (e.g., which assessments, cutoffs, or norms are used, as well as how the assessments are combined) identify different total amounts and different representations of students as gifted (Long et al., 2024). If different practices lead to different students being identified, then that will (also) have potential effects on any characteristics they may (or may not) share. Thus, any characteristics list developed relying on specific identification criteria may not generalize. If a characteristics list was developed to match identification practices, such specificity can facilitate appropriate alignment. If the list was not specifically aligned to a specific identification process, then it could introduce bias and error into that gifted identification process.
The need for alignment with identification practices may be similar to how auditions for a marching band differ from auditions for a concert band. For a marching band, how a student plays their instrument while physically moving becomes highly salient. Less so for a concert band. Thus, despite both seeking to select top musical performers, the different contexts require different criteria, which (may) lead to the selection of different students. The same can be true for different types of gifted programs. A competitive debate team will need to use different assessments than a program for creative writing or advanced literature. Without alignment to selection practices, a characteristics list could mis-describe students who would flourish in the service.
Lack of Conceptual Clarity
A fourth limitation concerns a lack of conceptual clarity. McBee and Makel (2019) showed how verbal definitions of terms (like giftedness) leave ample ambiguity for subsequent action and interpretation and can lack internal consistency. Characteristics lists may demonstrate such problems through jingle/jangle fallacies. Jingle fallacies occur when a single term is used to describe multiple different concepts and jangle fallacies occur when multiple terms are used to describe a single concept. Jingle/jangle problems have long been associated with testing and identifying gifted students (e.g., Kelley, 1927; Lohman, 2005; Lubinski, 2004). With this limitation in mind, I propose that if no reliable mechanism exists—or no reliable evidence exists—to differentiate features, they cannot be considered disparate characteristics. For example, is there a reliable and valid method to differentiate “high level of language development” from “high level of verbal ability” (Clark, 2008, p. 74)? If not, that is a potential jangle problem of listing one characteristic as two. Similarly, do the characteristics of “unusual intensity” and “unusual emotional depth and intensity” (Clark, 2008, pp. 75–76) represent unique characteristics? Or are they minimally different examples of the same underlying construct exemplifying a jangle fallacy?
Researchers may be accustomed to related items that are grouped as a subscale. Nevertheless, if characteristics lists do not come with instructions clearly connecting related items, such repetition gives extra weight to some underlying construct.
Similarly, a jingle fallacy may appear if/when a singular descriptor (e.g., “cynicism”; New South Wales Department of Education and Training, 2004, p. 8) is used to describe what may be an amalgamation of multiple features. For example, experiences of a student having desires/needs that do not align with their chronological age (latent feature 1) and spending time in an environment that insufficiently addresses those desires/needs (latent feature 2) may result in cynical beliefs (observed feature). In this case, cynicism may only be observed when two latent features interact.
Self-Fulfilling Prophecy and Bias
A fifth limitation of characteristics lists is their potential to unintentionally exclude students. As pointed out by Ford and Whiting (2008), if some list items are associated with particular cultures, characteristics lists could be biased against students from different cultural backgrounds. This is particularly problematic if characteristics lists are created by looking at who has previously been identified as gifted. Thus, rather than facilitating the identification of gifted students, lists could inadvertently skew who is identified and miscommunicate information about gifted students. For example, if we observed the entire population of U.S. Presidents for the first 250 years of U.S. history, one could say that “male” is a characteristic of Presidents because 100% of them have been men. One might also be able to say that being White was largely a characteristic. Based on observation, both of these features would appear true despite neither being a legal necessity (for much of the 250-year history at least). Following such a list of “characteristics” of U.S. Presidents would remove all non-men and non-White people from consideration. Siegle and Powell (2004) recognized this problem when they wrote, “as long as we use popular lists of characteristics to identify students, we risk failing to identify those who do not fit the listed characteristics” (p. 28).
If we recognize that there are problems in current identification practices (e.g., broad under-identification, bias in identification, bias in which schools identify), then relying on observation of features that identified students share sets the field on a selection bias self-fulfilling prophecy doom loop. In research, this is called collider bias or Berkson’s paradox (2025). Collider bias can be particularly problematic in observational studies (which are highly common in gifted education) and clinical samples. The essential fix is proper sampling of the full population and avoiding non-random samples. This point has been made elsewhere in gifted education (e.g., Pfeiffer & Foley-Nicpon, 2018; Pfeiffer & Jarosewich, 2007) and can have massive influence on outcomes.
For example, Lavrijsen and Verschueren (2023) compared mental health data of identified gifted students with that of students who were not identified as gifted but scored highly on cognitive ability assessments. As discussed above, how giftedness was identified is relevant. The authors state that in Flanders—where the study was conducted: there is no systematic ability testing in place; rather, the identification of gifted typically occurs as part of individual counselling trajectories for children experiencing academic or psychosocial difficulties . . . children with pre-existing psychological difficulties may be overrepresented in the labeled group. (p. 11)
Thus, gifted students may not represent students with high cognitive ability generally; instead, their selection has been biased by the presence of other features that are not intended to be relevant to gifted identification but become so due to logistical decisions. Their results reflect this bias. Adolescents formally identified as gifted reported lower global self-esteem and higher levels of emotional problems, worry, and hyperactivity and inattention than their peers not labeled as gifted. Parents of gifted students reported that their children had more emotional problems and conduct problems than parents of children with average ability. Based solely on these results, gifted students appear quite troubled. However, the community sample shows quite different results. Adolescents with high or very high cognitive ability reported higher levels of global self-esteem and lower levels of conduct problems compared to adolescents with average cognitive ability. Adolescents with high cognitive ability also reported less hyperactivity and inattention than adolescents with average cognitive ability. In other words, in the community example, high-performing students generally had profiles of high levels of functioning and low levels of mental health problems. If they had relied solely on identified gifted students, mental health problems would appear more frequent, not less. With this limitation in mind, when any list of characteristics is created based on an observational study, the features included in that list will reflect all biases that were part of how giftedness had been defined and identified, thus extending a cycle of bias.
Lack of Awareness of Prevalence Rates
Finally, existing characteristics lists fail to report prevalence rates of characteristics. As I argue in more detail below, to call a feature a “characteristic,” we must know the prevalence of that feature in both gifted and non-gifted students. For example, “having two eyes” is surely quite common in gifted students. But because the prevalence rate is likely indistinguishable from non-gifted students, it is not a particularly useful descriptor of gifted students. And yet the research that serves as the basis of Seagoe’s list (the closest thing to an empirical foundation of many lists) did not collect prevalence rates from typical students.
Simultaneously, a feature that is not particularly common among gifted students but is extremely rare in typical students could still be useful. If a feature appears in only 1 in 10,000 students but appears in 1 in 10 gifted students, the presence of this feature can be quite informative. But can a feature be considered a characteristic of gifted students if 90% of them don’t demonstrate it? Regardless, without knowing prevalence rates in both gifted and non-gifted samples, it is difficult for consumers of lists to make inferences about the presence or absence of features.
In this section, I have articulated six limitations of the present state of knowledge about characteristics of gifted students and the lists that purport to describe them. Without knowing the pervasiveness of the problems listed above, parents and teachers could place an unearned amount of trust in characteristics lists. Moreover, when lists are introduced to the public sphere, removing them can be difficult. Other fields have developed Brandolini’s Law (2024), which is an adage on it taking an order of magnitude more effort to refute incorrect information than it took to produce it. I hope that gifted education can avoid this level of effort in avoiding the use of characteristics lists that feature the above limitations.
In addition, it is important to note that at times authors of lists have subsequently backed away from the lists they created. For example, one famous list sought to differentiate gifted students from those who were “merely” bright (Szabo, 1989). Peters (2016) made a case for how the distinction being made is a distraction, arguing that both groups have learning needs that aren’t being met in the regular classroom. Szabo herself eventually went even further. According to journalist Jay Mathews (2018), she made “the list as just a tool for discussion at a conference. She never intended it for circulation. ‘It somehow took on a life of its own over the years and I dread even looking at it!’” (para. 10). Unfortunately, this admission appearing in a major newspaper does not remove the list from use. Versions (often mis- or un-attributed) of Szabo’s list remain posted on numerous websites of educational and advocacy groups (e.g., Beaverton School District, n.d.; Buffalo Hanover Montrose Schools, n.d.; Colorado Association for the Gifted, n.d.; Las Brisas Elementary School, n.d.; Taibbi, 2012; Washougal School District, n.d.).
Introduction of a New Nomenclature for Characteristics
In this section, I introduce a new nomenclature for characterizing features as characteristics. If characteristics lists are to be used, there needs to be clarity around what constitutes a characteristic. This framing will help categorize and describe features and communicate information about students. I have developed two rules that a feature must comply with to be considered a characteristic:
Rule 1. A feature must be typical (many gifted students 3 have the feature; if the feature is atypical, it is not a characteristic).
Rule 2. A feature must distinguish one group from another (gifted students have the feature but non-gifted students do not; if the feature is not distinguishing, it is not a characteristic).
Thus, to be a characteristic of gifted students, a feature must be typical of gifted students (Rule 1) and distinguish them from non-gifted students (Rule 2). To illustrate the importance of typicality, I use an example other than giftedness that has well-established population prevalence rates: autism. According to the Centers for Disease Control, in the United States, the autism rate in boys (1 in 42) is 4.5 times higher than in girls (1 in 189), making the prevalence of autism substantially higher in boys than girls (Christensen et al., 2016). However, autism is not considered a characteristic of boys because its prevalence is still low overall, violating Rule 1. Whether it also violates Rule 2 is less clear. Out of every 100 individuals with autism, approximately 82 are boys and 18 are girls. A sample frequency tree is shown in Figure 1A and provides more precise information on calculating prevalence rates. Recall the example of having two eyes from above to illustrate the relevance of whether something is a distinguishing feature. Without knowing the prevalence rate of the feature in gifted students as well as in non-gifted students, it is unclear whether the frequency with which a feature appears in a given population complies with the above rules.
Sample Descriptive and Diagnostic Natural Frequency Trees.
Categorizing When a Feature Is a Characteristic
As shown in Figure 2, these rules lead to four possible categorizations. A feature is a Characteristic if it is typical in one group (making it typical) but not the other (making it distinguishing). A feature is Uncharacteristic if it is atypical in the population of interest (violates Rule 1) but is prevalent in others. A feature is a Coinciding Feature (e.g., having two eyes) if it is typical in both groups (violates Rule 2). Finally, a feature is a Shared Rarity if it is neither typical nor more prevalent in a group (e.g., having only one eye; violates Rules 1 and 2). By adopting this nomenclature, we can more clearly communicate about all features as well as which features are characteristics of gifted students.
Nomenclature for Determining Whether a Feature Is a Characteristic.
Although the two rules are both expressed as dichotomous (yes or no), both are continua (i.e., proportions of populations). Plotting prevalence rates of features as continuous variables avoids a false dichotomization that characteristics lists may mistakenly communicate (e.g., depicting groups or individuals as either always or never demonstrating a feature). In addition, treating prevalence rates as continua communicates more precise information than dichotomization or verbal descriptions. If prevalence rates put a feature close to the origin of Figure 2 (as depicted by the diamond), the presence of that feature is less informative than if it is further from the origin. For example, if 52% of gifted students demonstrate a feature and 48% of non-gifted students also demonstrate that feature, its presence is not particularly informative. However, if 92% of gifted students demonstrate a feature and only 4% of non-gifted students demonstrate that feature (as depicted by the circle), its presence is far more informative. This is also how rare features can still be informative. If 1% of gifted students have a feature whereas one in a million non-gifted students have it (as depicted by the triangle), the presence of that shared rarity can still be quite informative despite it not being a characteristic of either group.
Descriptive Versus Diagnostic Lists
Characteristics lists have an additional limitation that is not mentioned above: There are multiple types of characteristics lists. Matthew T. McBee has noted (personal communication, 2019) that—depending on how a characteristics list was created and what it was created for—these different types of lists can describe two distinct groups of students. First, there are descriptive lists that report characteristics that are common among students who have been identified as gifted. The formal notation for this would be p(feature | identified gifted); the probability of a feature being present in students who have been identified as gifted. A different type is a diagnostic list that reports p(identified gifted | feature), which refers to the probability of a student being gifted given the presence of a feature. Diagnostic lists report whether students who exhibit a particular feature (e.g., asks deep probing questions) are likely to be gifted, whereas descriptive lists report features that are common among those already identified as gifted. Many consumers (and possibly creators) of characteristics lists may overlook this distinction.
Although descriptive and diagnostic lists may appear similar, their implications mean drastically different things. Figure 1 illustrates this via two natural frequency trees. Using data from Christensen et al. (2016), this example shows the distinction between the probabilities of A—a boy having autism (238/10,000 or 1 in 42) and B—if a child has autism, the probability of that child being a boy (238/291 or about 8 in 10). Although similar pieces of information are used to calculate both probabilities, these are clearly very different rates and provide different pieces of information. In gifted education, if a research team wanted to create a descriptive list, they would compare students who had been identified as gifted with those who had not. On the contrary, if the team wanted to create a diagnostic list (that could be useful for identification purposes), they would have to start with a general population of students (e.g., Pfeiffer & Jarosewich, 2007). Recognizing this distinction is important because—as described above—many characteristics lists are recommended for diagnostic purposes. If the list was created descriptively (i.e., describing already identified gifted students), it would be inappropriate for diagnostic purposes. Without vigilance, it can be easy for list consumers to mistakenly confuse things like underrepresentation caused by bias in the identification process with useful information.
Next Steps
Trying to correct errors on the internet is an established joke of a never-ending futile effort (Munro, n.d.). Recognizing this, a more fruitful task than finding errors in characteristics lists may be to determine whether any list can be ignored—or whether lists should be used at all. Assessing lists based on their ability to avoid/overcome the limitations listed above may provide value to consumers of lists. Just as one would not use a plane that cannot fly, a characteristics list that suffers from too many limitations is likely not useful. As such, one potential path forward is to cease the use of characteristics lists.
Alternatively, guidelines may also be useful for list-makers and consumers. In this section, I propose three broad actions that the gifted education research community can take to begin to assess and address the above limitations. Following these steps will help characteristics lists have a better chance to achieve their intended uses as well as to evaluate existing characteristics lists. My recommendations consolidate and build upon those made previously (e.g., Callahan & Moon, 2007; Martin et al., 2010; Pfeiffer & Foley-Nicpon, 2018; Rinn, 2024a, 2024b). Following these recommendations may not assure usefulness or quality. Rather, I propose them as more of a minimum threshold for consideration of using a characteristic list.
Gather Empirical Evidence Including Prevalence Rates
Grounding characteristics lists in systematic empirical support will help determine the value of the list and distinguish it from lists that lack evidence to support their claims. Relying solely on anecdotes instead puts the field in what Callahan and Moon (2007) labeled “arm-chair best thinking” (p. 305). Relying solely—or even predominantly—on convenience samples, clinical samples, extremely small samples, and qualitative research does not assess prevalence rates and thus is not helpful in establishing which features are characteristics. In fact, as demonstrated by Lavrijsen and Verschueren (2023), such non-systematic samples may create misleading beliefs about gifted students.
To earn the privilege of calling a feature a characteristic, we must first discover prevalence rates of that feature in both gifted and non-gifted populations. This is the only way to determine what features are typical in gifted students and distinguish them from non-gifted students. To acquire this knowledge requires population-level epidemiological studies. As pointed out by Pfeiffer and Foley-Nicpon (2018) in their discussion about twice-exceptionality, population-level epidemiological studies are required to make any realistic estimate about prevalence rates. Without realistic estimates of prevalence rates, how can any feature be determined to be a characteristic?
In addition, having clearly defined features that can be reliably measured (and differentiated from each other) will improve the usefulness of characteristics lists by avoiding jingle/jangle fallacies in such lists. Vague items that may list one underlying feature multiple times or that may describe multiple distinct features in a single item will be more clearly revealed (and revamped) from lists that have empirical evidence. Again, empirical evidence is needed to assess item reliability and validity.
Connect Lists With Specific Definitions, Domains, and Identification Practices
Without a connection to a specific definition of giftedness, it is unclear whether the list will be useful, too generic to provide value in a specific situation, or whether it could even lead consumers astray. Similarly, connecting characteristics lists to specific domains also has potential for increasing their usefulness. Just as a sprinter in running differs from a sprinter in swimming, characteristics of students who would benefit from an accelerated math service may differ from characteristics of students who would benefit from an in-class language enrichment service. Use of a single list to describe two distinct populations may simply not be feasible, much less useful.
Rinn (2024a, 2024b) made a similar argument when writing about conducting research on psychosocial skills. She emphasized the need to focus on aligning with specific definitions and operationalizations of terms (e.g., “research should clearly specify what they mean by gifted in their sample . . . clearly define the constructs they are measuring in their study, from both a theoretical and conceptual lens and also from a measurement lens”; Rinn, 2024b, p. 37). From this perspective, results from general prevalence data do not establish a universal list of “real” characteristics of gifted students. Rather, characteristics vary based on the definition of giftedness adopted as well as the tools and cutoffs used.
Peters and colleagues (2023) made a similar case when discussing the assessment of identification systems. They argued, “[a]n effective identification system is one that measures the same skills and dispositions necessary for success in a particular service” (p. 140). This is important because weak alignment leads to missing students who would benefit from the service and/or mis-selecting students who would not benefit from the service. Two specific factors they mention are alignment across domain and intensity of the service. Consider the opposite, where only a single characteristics list was used for all definitions of “gifted” or across all domains. It would ignore all potential differences in which type of student is being sought for what purpose. This would be analogous to using a single characteristics list for all definitions of “athlete.” Our ability to identify promising distance swimmers, gymnasts, and hockey goalies would suffer. Relying on a generic characteristics list should be assumed to have massive validity limitations.
Rather than assume characteristics are definition—and domain—general, their presence and prevalence connected to specific definitions and operationalizations would help make them useful—and assess whether they are useful. Namely, if the prevalence of specific features within specific definitions and specific domains is known, then they might be useful for either diagnostic or descriptive lists (as illustrated in Figure 1A and B). To accomplish this, definitions of giftedness must have sufficient clarity that practitioners could reliably act upon them (see McBee et al., 2016; McBee & Makel, 2019).
This does not mean that the field needs to adopt a single universal definition of giftedness. Whether that would be beneficial or harmful is beyond the scope of this article. Rather, it means that different definitions—and different identification/selection practices—will likely lead to the need for different characteristics lists.
Kill Zombie Ideas
Zombies are not living but do not die (Romero, 1968). Zombie ideas lack evidence to support them (and may have even been refuted) but remain common in the world and culture (e.g., Feldman Barrett, 2019; Krugman, 2013). Researchers in psychology have asserted that systematic incentives play a role in the perpetuation of undead ideas that lack evidence (e.g., Ferguson & Heene, 2012; Fried, 2024). Given that even retraction of an article is not associated with a long-term change in its citations (Candal-Pedreira et al., 2020), it can be quite difficult to remove an idea from popular discourse (i.e., Brandolini’s Law). Substantial psychological research suggests that debunking an inaccurate framing is difficult; it requires demonstrating why the initial belief was misguided and replacing it with a detailed correct framing (e.g., Chan et al., 2017). Characteristics lists that lack empirical support but continue to live on the internet and in school settings are zombie ideas.
Existing characteristics lists may appear to align with casual observation. This false familiarity may make them persuasive and persistent. But relying solely on informal observation leads to worldviews that include a flat earth and the spontaneous generation of life from nonliving matter. And—as demonstrated in the Flanders example above (Lavrijsen & Verschueren, 2023)—observations of identified gifted students may mislead about the actual population of interest. Without moving beyond observation and arm-chair best thinking, we cannot know which aspects of characteristics lists survive scientific scrutiny. My goal with this article is to initiate a conversation that will magnify the focus on which ideas about giftedness and characteristics have sufficient evidence to merit policy, which require additional research, and which need their heads cut off.
To kill zombie ideas, existing lists must be evaluated on their empirical evidence. To prevent the rise of new zombie ideas, any new list must avoid the above limitations. The path forward described in Table 1 may serve as a useful guide when evaluating existing characteristics lists. The more questions with clear positive answers, the more likely the list will provide value to consumers. The more questions with unclear answers, the more likely the list needs additional assessment prior to being used or more caution when it is being used.
Characteristics List Limitations and a Proposed Path Forward.
Future Research
In this article, I have outlined how characteristics lists are used, their limitations, and proposed a path forward that (I believe) would better align actions with empirically supported aspiration. Whether characteristics lists should be used is beyond the scope of this article. Rather, I argue that they need to meet a minimal quality threshold to have even a chance at being useful. However, my proposal requires investigation to assess whether it accomplishes what it seeks. Namely, does the proposed path forward lead to more efficiently and effectively identifying and serving gifted and talented students? If not, should they continue to be used?
Useful future research could include assessing whether using characteristics lists helps improve identification/selection for a gifted program or service. It could also include seeking systematic understanding of the empirical support serving as the foundation of empirical lists, how lists are viewed by relevant stakeholders, and the effects of their use on students across various contexts could all help improve understanding and effectiveness of characteristics lists. Future research could include investigations such as: Whether the proposed path leads to the creation of more empirically aligned characteristics lists or a systematic investigation of how existing characteristics lists cite references. How well do existing lists align with the proposed path forward? How can policymakers and practitioners make informed decisions about using and interpreting characteristics lists? Systematic assessment of how policymakers and practitioners view or use such lists. Are they generally trusted and relied upon? Or are they generally viewed with skepticism? Are they treated like a horoscope or a medical diagnosis? Despite characteristics lists being created, suggested, and used across the field, it appears that we know little about their empirical support, effects, or how they are perceived by relevant stakeholders. If a list is going to be a part of real-world policy or practice, understanding its origin and effects is a worthwhile endeavor.
Implications
It is a dereliction of duty and a violation of the public trust if the research community misrepresents a thinly supported set of assertions as though they are well grounded in empirical support. It is the responsibility of the research community to rigorously evaluate what is known and what still needs further investigation. If there is a rich and well-sourced body of empirical research supporting characteristics lists, the research community needs to highlight this strength while making the benefits of such lists more transparent to parents and practitioners. If there is not a well-sourced body of support, the research community needs to act before it advises. Or it needs to advise with clearer caveats. Absent data, the research community risks its reputation and the value that it provides to society.
Adopting the perspectives and path forward shared in this article will require researchers, reviewers, and editors, as well as educators and administrators to adapt. Some may be defensive. Being told “we’ve been doing it wrong” is hard. But it is the job of professionals to adapt as new information is acquired. Long-term behavior change upon acquisition of new information is useful. It’s called learning. And excelling at learning is a feature that I believe is a characteristic of the field of gifted education professionals, not just its students.
Footnotes
Author’s Note
This manuscript benefited from the thoughtful feedback from Matthew J. Hays, Scott J. Peters, and several attendees of the 2024 Wallace and NAGC conferences. Mara Shurgot provided the wonderful graphic design support.
Ethical Considerations
This research did not require IRB approval because it did not involve human subjects.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The writing of this manuscript was done with partial financial support from an Esther Katz Rosen Fund Grant from the American Psychological Foundation (2017).
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Open Practices
The data analyzed in this study are not available for the purpose of reproducing the results. The code or protocol used to generate the findings reported in the article is not available for purposes of reproducing the results or replicating the study. There are no other newly created, unique materials used to conduct the research.
Artificial Intelligence Use
The authors confirm that no generative AI tools were used in the development of this article.