What is Character Education?: Development of a Prototype Free

What is character education, and can it be differentiated from other types of educational programs that attempt to encourage what could be called “growth as a person?” This is clearly an important and fundamental question in the field of character education, but it is rarely directly addressed in the literature. However, most literature on the topic discusses character education without being clear about what that means.

Lapsley and Narvaez (2006) effectively discussed why this lack of definitional clarity is problematic. The effectiveness of character education programs is typically evaluated via improvements in youth behavior. Without a good definition of what constitutes a character education program, though, any program intended to increase prosocial behaviors or resistance to negative behaviors could be treated as an example of character education. Lapsley and Narvaez provided examples of cases where reviews of the effectiveness of character education programs were overly inclusive of programs with little or no focus on the concept of character. It is therefore impossible to gauge the value of character education programs accurately without a clear understanding of what makes a program focusing on personal growth a character education program. At a more practical level, I have found that educators and others interested in developing character education programs are often unclear about what that actually means.

A review of the literature on character education and of existing character education programs reveals substantial variability in what is meant by that term. For example, some of the major figures in the field of character education have identified intentional moral development as the essential element. Berkowitz, Althof, and Bier (2012, p. 72) defined character education as “the intentional attempt in schools to foster the development of students’ psychological characteristics that motivate and enable them to act in ethical, democratic, and socially effective and productive ways.” Ber-kowitz and Bier (2005) provided examples of others who have shared this emphasis on the moral and prosocial (see also Nucci & Nar-vaez, 2008).

Other theorists, as well as developers of programs widely considered examples of character education, have rejected, minimized, or expanded on this perspective (Rivers, 2004). For example, the character education program implemented by the KIPP network of charter schools focuses on seven character strengths— zest, grit, self-control, social intelligence, gratitude, optimism, and curiosity (Tough, 2011)—none of which is primarily moral or communal in nature with the possible exception of gratitude. Lickona and Davidson (2005; see also Character.org, 2014) suggested character education programs focus on the enhancement of both moral and performance functioning, while Berkowitz, Bier, and McCauley (in press) identified four key targets: moral, performance, intellectual, and civic character. Still others have argued that character education programs should allow each participating youth to decide what elements of character to focus on for themselves (Bates-Krakoff, McGrath, Graves, & Ochs, 2017; Linkins, Niemiec, Gillham, & Mayer-son, 2014).

This diversity of perspectives raises concerns about whether a comprehensive definition of character education is even possible. This article explores one approach to resolving the problem, by locating the problem in how we conceptualize what it means to define something rather than in how we conceptualize character education. The next section will outline the concept of a prototype, and describe how this concept provides an approach to definition that will prove helpful in the context of character education.

Beginning with Plato, philosophers and scientists assumed that defining a lexical category meant identifying the necessary and sufficient conditions for inclusion in that category. This approach to defining concepts assumes the existence of natural kinds, or objective dividing lines in nature. Applying this approach to the definition of character education provided by Berkowitz et al. (2012) would suggest a character education program is marked by (1) location in a school; (2) intentions of fostering certain psychological characteristics; and (3) a focus on characteristics that enhance ethical and social functioning. Programs that do not demonstrate all these features presumably would fall outside the domain of character education, though the necessity of all three conditions is left implicit in their definition.

Wittgenstein (1953/1958) is credited with first making the case that categories can be formed based on family resemblances among members of the category rather than on essential features. Research in cognitive psychology has more recently suggested that humans naturally create categories based on family resemblances rather than necessary and sufficient conditions. This process involves the development of a prototype, an ideal for the category, and then allocation of objects to the category based on consistency with that ideal (Rosch, 1973; also see Vygotsky, 1962, chap. 5). Our reliance on prototypes to understand categories explains why we see a cat as more mammalian than a whale or platypus, for example.

Traditional definition and definition by prototype can be contrasted in several ways. The former is formal and logical, clearly distinguishes members from nonmembers of the category, and sets firm boundary conditions for category membership. Definition based on a prototype is heuristic and practical, encourages the perception of degrees of membership in the category, and sets fuzzy boundary conditions that can leave it ambiguous whether certain objects belong in the category.

Rosch’s (1973) work has inspired a substantial literature that attempts to define the prototype underlying various common language terms. For example, Kinsella, Ritchie, and Igou (2015) conducted a series of studies intended to understand what we mean when we call someone a hero. The concept of the prototype has also been used in taxonomic efforts when necessary and sufficient conditions for classification are unclear, debatable, or nonexistent. A good illustration of this practice is the current diagnostic system for mental disorders. For example, the diagnosis of major depressive disorder is based on nine symptoms characteristic of depression, no one of which is necessary for the diagnosis (American Psychiatric Association, 2013). In contrast to studies on natural language prototypes, this prototype had to be developed rather than discovered, and was created by a study group that derived their nine central features from technical literature on the nature of depression.

The diagnostic system for mental disorders offers a useful example of how the concept of prototypes can be applied to the definition of character education. The development of a prototype for character education would involve the identification of central features for character education programs based on research about character education, with no implication that any are necessary or sufficient for such a program. The remainder of this article is devoted to the development of a proposal for such a prototype based on a review of programs that ostensibly provide training in character, and of literature on character education as well as the related concepts of virtue and moral education. The goal is to provide a starting point for dialog among experts about what would constitute an optimal prototype for character education.

Table 1 introduces seven central features for a prototype of a character education program. These were identified through a review of prior literature that provides definitions for character education, and review of the curricula for several character education programs. The first two features are structural, while the last five focus on content. None of the features focuses on program process, because at present there is no literature consistently suggesting any didactic practices are uniquely featured in character education programs. However, as will be discussed, certain program contents can suggest specific educational strategies.

As noted earlier, formal definitions of character education sometimes restrict the domain to school settings. This may simply be an artifact of the reference to education, but a more substantive possible explanation for associating character education with schools over out-of-school time programs will be discussed below in the section on identity. However, there are notable instances in which out-of-school programs treat character as a goal for intervention. One example is the positive youth development program implemented by 4-H among others, a program that specifically includes character as one of “5 Cs” targeted for intervention: competence, confidence, connection, character, and caring/compassion (Bowers et al., 2015; see Jones, Dunn, Holt, Sullivan, & Bloom, 2011, for another example).

The focus on school-based programs provides a good basis for comparing definition by prototype with definition based on natural kinds. Traditional definitions that locate character education in schools imply out-of-school programs fall outside the umbrella of character education. Definition by prototype allows for a more flexible perspective. A program that demonstrates many of the other features of a character education program could still be treated as falling within the category even if it takes place outside of school. If two programs are equivalent on all other features, but one is school based while the other is extramural, definition by prototype would suggest that both are likely to be perceived as examples of character education programs, though the former may be perceived as a truer exemplar of the class than the latter.

Something that is to be thought of as a program should demonstrate some level of structure. Many teachers and out-of-school programs incorporate informal or ad hoc elements into their activities that could contribute to the development character, but these are not exemplars of a character education program. A true character education program is built around shared expectations about what is being transmitted or presented to the participants, and how that transmission takes place in that program, though the degree to which those expectations are explicitly spelled out may vary across programs. Ideally, a formal curriculum would be available for review to ensure consistency in implementation, allow for replication, enhance fidelity, and ultimately guide research on program effectiveness.

The term attributes is used here to refer to any characteristic of the program participants that is relatively enduring but also considered malleable. Both qualities are important to identifying attributes that deserve to be considered targets for intervention. Without malleability, intervention would be senseless; without enduring outcomes, the value of intervention would be ephemeral.

Psychological attributes are relatively enduring qualities of the individual that describe the individual’s characteristic approach to interpreting and interacting with the larger environment, particularly other people. At this point in the process of developing the prototype, the reference to positive psychological attributes is broadly conceived, to incorporate any quality of the individual that contributes to social good, or that contributes to personal good and does not typically result in harm to others. That is, positive psychological attributes include any attributes of the indi

vidual that achieve some sort of benefit to the individual and/or community within which that individual operates, without causing purposeful harm to the welfare of that community. The focus on positive psychological attributes represents the most central feature of any program intended to achieve personal growth, though the prototype approach to definition allows for the possibility of developing a personal growth program that does not target positive psychological attributes.

The preceding feature does not distinguish among different types of programs intended to enhance personal growth. A review of such programs suggests an important distinction in the terminology they use to describe the positive psychological attributes that they target. Where many programs are purposely intended to address character strengths or virtues, in other cases the intended targets are referred to or conceptualized as skills. Perhaps the best-known example of the latter is programs that focus on social and emotional learning (Durlak, Weissberg, Dymnicki, Taylor, & Schellinger, 2011).

Conceptual discussions of skills (Fischer, 1980) and character (Peterson & Seligman, 2004) suggest substantial overlap between the two classes of positive psychological attributes. Both can be conceptualized as the products of gradual change in the individual’s characteristic style of responding to environmental events because of person-context interactions, a developmental framework that has more recently come to be called relational developmental systems (Overton, 2015). Both are also conceptualized in terms of evolving cognitive structures that shape behavior. There is no reason a skill-centric program could not influence character and vice versa. In fact, the categories of skills and character are themselves inclusive of elements that are best understood as related based on family resemblance rather than necessity, so the boundary between the two is permeable. For example, both McGrath (2015) and Park, Tsukayama, Goodwin, Patrick, and Duckworth (2016) have recently identified three broad classes of positive psychological attributes, one having to do with interpersonal issues, one having to do with issues of self-control, and the third having to do with intellectual functioning. However, where Park et al. focused on these specifically as academically relevant areas of skill, McGrath referred to them as general virtues, a concept closely connected to that of character. There is also the possibility of hybrid programs, such as the “5 Cs” of positive youth development described above (Bowers et al., 2015), or social-emotional and character development (Elias, Parker, Kash, & Dunkeblau. 2007).

Given the similarities, what distinguishes the two concepts? Based on several sources (review of curricula, literature on virtue theory, and certain sources in character education, e.g., Berkowitz, 2012), it can be hypothesized that character education differs from skills training primarily in terms of the extent to which it focuses on narrative identity. Narrative identity is a term used to refer to an internalized autobiographical perspective that ties together past, present, and future into a cohesive story about the person (McAdams & Manczak, 2015; McAdams & McLean, 2013). The stories we tell ourselves about ourselves are broad in scope: they represent thematically coherent statements about our relationship with our life situation, our values and goals, our understanding of ourselves as individuals, the role of suffering in our lives, our cultural context, and many other pieces of who we are. From this perspective, a central goal of character education is to help us see ourselves as individuals defined in important ways by our positive psychological attributes.

This focus on identity in character education allows it to be compared with skills training in a variety of ways, though in cases where differences exist they differ in terms of degree. Both types of programs are committed to the development of positive behavioral habits. In the case of character education, this association with habit may be traced as far back as Aristotle’s discussion of habit in relation to virtue (Bartlett & Collins, 2007), while the conception of habit in skills-based programs is a more recent product of psychological learning theory. However, where skills training emphasizes the participant achieving behavioral effectiveness, character education is more about the participant developing a greater sense of who he or she is or may become. Where skills training tends to be about current functioning, character education can have more of a focus on a trajectory for future change. Though both are the product of individual-context interplay, skills training focuses more on how to respond in specific situations, character on the stable attributes underlying those responses. Put overly simply, skills training is about “what will I do?” Character education is about “who will I be?”¹

This difference in emphasis offers one hypothesis for why out-of-school time programs might prefer to target skills rather than character. While many out-of-school time programs aspire to contribute to the personal growth of involved youth, the development of character can seem too ambitious a target for programs with limited resources, goals, and youth access.

One of the reasons the distinction between a program that focuses on skills and one that focuses on character is important, even if difficult to make, is that it has some implications for the types of interventions that are likely to be used. Skills training suggests a greater emphasis on behavioral shaping and reinforcement, though gradual change can be an important component in both types of programs. Modeling and storytelling can also be important tools across both, though how they are used may differ. In skills training, modeling and storytelling may be used primarily to demonstrate specific responses to specific environmental events. In character education, the leader may be more important as a global role model for participants (Kristjânsson, 2015), and storytelling as a means of sharing values with participants (Tappan & Brown, 1989).

One technique that particularly highlights the differences between character education and skills training has been called strengths spotting (Linkins et al., 2014). This term refers to various strategies intended to help participants define one or more positive psychological attributes, and identify examples of the expression of that attribute in themselves and/ or others. This strategy makes relatively little sense in the context of skills building, where the assumption is that the participant needs to enhance the skill. Strengths spotting assumes the positive attribute is already present to some degree, and it is through the process of integrating that strength into one’s sense of identity that the attribute is enhanced.

Developmental factors play an important role in how identity is likely be addressed in character education program. There is reason to believe that preadolescent children generally do not demonstrate the cognitive capacity to generate a life story that weaves together personal history, current context, and aspirations for the future (Habermas & Bluck, 2000). Instead, self-statements tend to be associated with specific incidents rather than comprehensive statements about self. For younger children, then, the focus on identity may involve little more than an emphasis on strengths spotting, with the goal of enhancing the child’s labeling of various positive attributes as they are currently constituted.

One further area of overlap between skills-building and character education programs is that they are often implemented for similar purposes, such as, to reduce bullying or enhance resilience. Whether the two types of programs differ in their effectiveness for achieving objective outcomes of this type remains an open question, in part because of the difficulties already discussed distinguishing between the two camps.

In summary, the distinction between skills training and character education is the most subtle discussed so far. At the same time, it may well be the most important for distinguishing between character education and other kinds of personal growth programs.

The concept of character has historically been closely associated with concepts of morality and virtue. Ralph Waldo Emerson (1884, para. 5) referred to character as “moral order seen through the medium of an individual nature.” Baumrind and Thompson (2002, p. 12) captured a common perspective among psychologists about the nature of character when they referred to is as “personality evaluated.” We have already seen that some leaders in the field of character education identify the moral component as central (Berkowitz, 2014; Berkowitz et al., 2012), though programs exist that emphasize the flourishing of individual participants more than the flourishing of the community (e.g., Tough, 2011). The number of essays on character education that have emphasized the moral component of such programs indicates that moral content deserves to be included among the central features of character education. However, it should be noted that many skills-oriented programs include a focus on prosocial skills. It is currently unclear whether character education programs as a set tend to deal with issues of moral development in a different manner than skills-based programs tend to teach prosocial skills (e.g., through a greater emphasis on deontological or universal principles of morality). In the absence of evidence for such a difference, a moral emphasis in itself may not be a good basis for distinguishing between character education and skills-based programs. Research examining how issues of prosociality tend to be addressed in the two types of programs could provide important clarification on this issue.

The final two central features are suggested as prototypical of character education even though they do not seem particularly common foci for existing character education programs. Since prototype definition can be about defining the ideal for a category, they merit inclusion as aspirational goals for character education programs. For this reason, the rationale for their inclusion rests more on philosophical speculations than examples from the character education literature.

Aristotle’s discussion of the nature of virtue has led some commentators to talk about the reciprocity of the virtues, the proposition that a person cannot be considered virtuous unless he or she has achieved proficiency in all the virtues (Irwin, 1988). Reciprocal development among elements of character represents a potential goal for the prototypical character education program, one that fits particularly well with the emphasis on identity and life stories. That is, character education programs ideally should aim to enhance character as a whole, or multiple elements of character. This focus on holistic growth of character is another way in which true character education programs can be differentiated from skills-based programs. Skills tend to be perceived as distinct behavioral tendencies, while the development of character has associated with it a whole-person approach. Programs that focus on specific elements of character to the exclusion of others can be considered less prototypical than a more multidimensional character education program.

Efforts at an integrated character education program can be enhanced by theory about the essential elements of character. It was mentioned previously that McGrath (2015) identified three core elements of character called caring, inquisitiveness, and self-control. Similarly, Berkowitz et al. (in press) suggest four elements for a comprehensive view on character: moral, civic, performance, and intellectual. Programs that highlight only one or even two of these elements may be incapable of ensuring individuals of good general character. The ideal character education program would instead convey to its participants the importance of growth across all the key dimensions of character. Whether there are additional key elements to be targeted by holistic character education programs is a question to be deferred for future research on the essential nature of character.

Another important concept to emerge from Aristotle’s philosophy of virtue is the importance of practical wisdom or phronesis, that is, the deliberate and thoughtful application of elements of character to specific situations (Bartlett & Collins, 2007). According to Aristotle, the truly virtuous person deliberates about which virtues should be applied to what extent based on the situation. The prototypical character education program would not only nurture the essential elements of character, but also the deliberative process of deciding how to act as a person of character in specific contexts (Kristjânsson, 2015). An important element of this practical wisdom according to Aristotle involves avoiding the overuse or underuse of one’s strengths, a topic that has recently received attention in the literature on character (e.g., Freidlin, Littman-Ovadia, & Niemiec, 2017; Peterson, 2006).

The prototype that has been introduced here is not intended as a purely intellectual exercise; the ultimate goal is to provide a basis for evaluating personal growth programs as examples of character education. There are two complementary approaches to making these kinds of judgments. The first is called polythetic classification, the second prototype matching.

Comparing mental disorders with more traditional medical illnesses illustrates one practical problem with using prototypes. In medicine, most diagnoses have an essential biological basis that makes them natural kinds, such as, influenza involves infection by a specific family of virus. Classification based on essential conditions is called monothetic classification.

The absence of a clear dividing line in a prototype for a mental disorder can render diagnosis unreliable. To address this problem, category minima were developed for each of the mental disorders. For example, the minima for a diagnosis of major depression include: (1) The person must demonstrate at least five of the nine central symptoms; (2) At least one of the symptoms must involve depressed mood or loss of interest or pleasure; (3) The person must experience clinically significant distress or impairment resulting from the disorder; and (4) The disorder cannot be attributed to physiological factors or to certain other mental disorders (American Psychiatric Association, 2013). Diagnosis of mental disorders is an example of polythetic classification, classification based on shared features, none of which is essential. In polythetic classification, the category minima for the prototype are recognized as arbitrary, since no natural dividing line exists for the category.

As a starting point for the polythetic classification of programs as character education programs, the following features are suggested as reasonable category minima:

1. The program is structured.

2. The program addresses specific positive psychological attributes.

3. The program addresses identity.

Acceptance of these minima would indicate that a program demonstrating these three conditions can for practical purposes be considered an example of character education, while a personal growth program that does not demonstrate all three would be excluded. The other four central features play no role in deciding whether the program should be considered character education, unless others (researchers, practitioners, or the field as a whole) settle on category minima that include them.

It is inherent to polythetic classification that the minima are heuristic, to some extent arbitrary, and open to reasoned debate. The definition of character education provided earlier from Berkowitz et al. (2012) indicates reasonable objections to the exclusion of an academic setting and a moral focus from the basic requirements, for example. It might even be reasonable for different researchers to set different category minima depending on their goals, though those minima should always be explicitly stated.

Polythetic classification results in a binary decision about a program: do we consider it character education or not? Prototype matching offers a complementary approach based on a review of all seven central features, one that results in a dimensional evaluation of consistency with the prototype.

Two empirical strategies are available that can be considered forms of prototype matching. One provides raters the list of seven central features for the prototype, and asks the raters to evaluate a program’s overall consistency with the prototype on a 1-5 scale (Westen, 2012). The second asks raters to evaluate the program on the presence or absence of each of the seven features. The simplest way to aggregate this information would award one point for each feature that is present. The result would be a 0-7 scale reflecting degree of consistency with the prototype. It is also possible for the features to be differentially weighted. Cooper and Balsis (2009) suggested the use of item response theory for this purpose, which would weigh the less common features more highly; another option would put more weight on the features that comprise the category minima. Whatever weighting system is chosen, and the best weighting system should be based on empirical evidence suggesting a value to one over others, the result would be a dimensional scale where higher scores indicate better match to the prototype.

Polythetic classification and prototype matching serve complementary ends. Poly-thetic classification is more appropriate if the researcher or practitioner wants to decide if a program is or is not an example of character education. This is useful if the goal is to compare character education to other forms of personal growth programing, for example. Prototype matching is more helpful if the goal is to evaluate how closely the program matches the “ideal” of a character education program. This can be particularly useful, for example, when the goal is to identify exemplary character education programs, or to compare across character education programs.

The first step in either polythetic classification or prototype matching requires educating raters about the program. Ideally, the raters would extensively observe the program in action. Less desirable options would include limited observation; interviews with program stakeholders; review of the program curriculum; or a written description of the program. To maximize the reliability and validity of the ratings, multiple raters knowledgeable about character education but external to the program are preferable over single raters, raters naïve to character education, or raters who are associated with the program.

If the goal is polythetic classification, the raters only have to consider those features that comprise the category minima set by the researchers. Prototype matching in contrast requires rating all seven features or, if the raters are just providing a 1-5 rating of prototype match, a description of the features. Rating of the features could take place on a dichotomous (present-absent) or dimensional scale, though the latter would require setting a minimum rating needed to treat the feature as present in the case of polythetic classification. A sample rating form using a dimensional scale may be found in Table 2.

For example, School A has implemented a program with a formal curriculum that focuses on the development of prosocial skills. There is no discussion of how prosocial skills operate in tandem with other skills, or when they should not be used. The program is school-based and structured, targets positive attributes, and is concerned with moral growth. Using the category minima proposed for a character education program, this would not meet criteria for a character education program. There is clear evidence for four of the seven features, so equal weighting of the features would generate a prototype match score of 4.

School B has also implemented a program with a formal curriculum. It focuses on self-determination as a personal attribute, and on the importance of seeing ourselves as self-determined people. Because the program is school-based, structured, focuses on positive psychological attributes, and addresses issues of identity, it meets the proposed category minima for a character education program. However, in terms of prototype matching, it would still receive a score of 4 on the 0-7 scale, because the moral focus has been replaced by a focus on identity.

Whether a score of 4 is a good match to the prototype relative to other character education programs is a question that cannot be answered until a variety of programs have been evaluated to determine what is normative. What we can say is that they are both moderately similar to the ideal for a character education program. It may well be that the School A program will even have lasting effects on character. However, since identity is not a target of the intervention in School A, only the School B program would represent a true example of a character education programs, at least according to the proposed category minima.

There are practical issues inherent to the application of polythetic classification and prototype matching in character education. Reasonable observers can differ in terms of their judgment about whether or not a program is structured, or whether it is truly holistic. As aspirational features, the last two features may be particularly susceptible to disagreement. With concerns about the reliability of measurement in mind, polythetic classification and prototype matching offer an empirical basis for the identification of character education programs. In this way, the proposed prototype provides a starting point for discussing issues in character education with greater clarity than has been possible in the past.

As multiple editions of the diagnostic manual for mental disorders, and literature critical of each of those editions, attest, the development of a technical prototype is an evolving process that is open to debate. The purpose of this article has been to begin that process. Future literature and discussion may well result in substantial changes to the prototype and category minima offered here. There are some conclusions to be drawn from this article that may be of enduring value, though. First, greater clarity in what is meant by character education can only advance the field for both researchers and practitioners. Second, definition by prototype rather than necessary and/or sufficient conditions provides a framework for definition that can accommodate reasonable differences across character education programs and perspectives on the nature of character education. Finally, the general adoption of a prototype for character education would allow for the formal evaluation of programs via polythetic classification or prototype matching. Such a development would have substantial potential for clarifying discussions and research on character education.

I am grateful to Marvin Berkowitz for his comments on earlier versions of this article.