Abstract
We identify a divide between two schools of thought in leadership behavior research. The classical behavioral school (CBS) categorizes leadership behaviors based on their latent functions and meanings—task, relations, change, and moral—with the aim of achieving intersubjective agreement among evaluators. In contrast, the neo-behavioral school (NBS) focuses on the observable forms and manifestations of behavior—choices, embodiment, interactions, textual markers, and temporal patterns—in the pursuit of objective description. In this article, we conduct a metasynthesis to systematically describe, evaluate, and integrate these two schools. We identify their distinct research logics, showing that CBS categories enable cross-contextual generalization but conflate behavioral description with evaluative judgment, thereby risking causal indeterminacy and limiting actionability. Conversely, NBS approaches enable causally well-determined analyses of what leaders concretely do but risk reductionism and limited generalizability of meaning when studied in isolation. Building on this analysis, we identify school-specific best practices and distinguish limitations that can be mitigated within each school from those that require cross-school integration. We then develop an integrative framework that assigns distinct causal roles to concrete behavioral manifestations and to evaluations of their function and meaning, explicating how leadership behaviors acquire meaning across contexts and how CBS evaluations emerge as downstream consequences of NBS-specified behavioral aspects and various contextual factors. Finally, we propose a generative research agenda that advances leadership behavior research across schools of thought. The framework also has practical implications, enabling behaviorally grounded leadership development through actionable feedback, as well as more equitable and evidence-based leadership selection anchored in observable behavior.
Studying the effects of leadership behaviors on various outcomes of interest has a long history and dates back to the pioneering study of autocratic, democratic, and laissez-faire leadership by Lewin, Lippitt, and White (1939). Since then, scholars have conducted thousands of studies on the effects of leadership behaviors (Yukl, 2012). Scholars typically represent concrete behaviors in general constructs (Yukl & Gardner, 2019). Examples are leadership styles such as authentic, ethical, and servant leadership (Lemoine, Hartnell, & Leroy, 2019), destructive forms of leadership (Mackey, Frieder, Brees, & Martinko, 2017), transactional-transformational leadership (Avolio, Bass, & Jung, 1999), initiating structure and consideration behaviors (Judge, Piccolo, & Ilies, 2004), or single workplace behaviors (e.g., feedback seeking; Ashford, De Stobbeleir, & Nujella, 2016). We refer to this type of research as part of the classical behavioral school of thought (CBS). In this school of thought, behaviors are defined as categorizations of leadership behaviors based on their latent and valenced function and meaning. For example, a behavior may be defined as whether a leader serves others or behaves unethically.
There is also a nascent alternative that we call the neo-behavioral school of thought (NBS). Research in this school studies leadership behavior too, but it directly represents concrete and separate aspects of leadership behavior by focusing on their form and manifestation. Examples are specific words used in speeches (Antonakis, d’Adda, Weber, & Zehnder, 2022), open questions asked in interactions (Lehmann-Willenbrock, 2024), making strategic organization-level decisions (Samimi, Cortes, Anderson, & Herrmann, 2022), patterns of eye-gazing (Maran, Furtner, Liegl, Kraus, & Sachse, 2019), or the temporal patterning of behaviors (Gerpott, Lehmann-Willenbrock, Voelpel, & Van Vugt, 2019). In this school, behaviors can be defined as the form and manifestation of concrete and separate aspects of leadership behaviors.
In this article, we conduct a meta-synthesis to describe and evaluate the research logics of the two schools of thought and integrate them into a unified framework. This framework specifies the causes and consequences of leadership behaviors and clarifies the causal relationships linking CBS and NBS classifications, from which we derive theoretical pathways and empirical opportunities for future research.
Our meta-synthesis makes three contributions. First, we describe the research logic of the two schools in terms of their conceptualization, measurement, and causal explanation. We provide an in-depth analysis of existing research that goes beyond broad overview articles (e.g., Avolio, Walumbwa, & Weber, 2009; Liden, Wang, & Wang, 2025; Lord, Day, Zaccaro, Avolio, & Eagly, 2017) and textbook chapters (e.g., Antonakis & Day, 2017; Yukl & Gardner, 2019) that map the field without uncovering its underlying research logic. We characterize the classical behavioral school as relying on a general function- and meaning-based classification logic. In line with previous research, we identify four sublogics of classifying behaviors in the CBS, judging the extent to which they serve task-, relations-, change-oriented, or moral goals and purposes (e.g., Lemoine et al., 2019; Stogdill, 1963; Yukl, 2012). Different from previous research, however, we argue that these categories do not describe different behaviors but represent different classification logics, such that the same behavior can be evaluated according to multiple sublogics at the same time. For example, when a manager meets with an employee and suggests an alternative work procedure, their behavior can be simultaneously classified as problem-solving (task-oriented), supporting followers (relations-oriented), and encouraging innovation (change-oriented), which refer to different leadership behaviors in Yukl’s (2012) taxonomy.
We characterize the neo-behavioral school as relying on a form- and manifestation-based classification logic, and we identify five sublogics of classifying behaviors in the NBS—namely choice, embodiment, interaction, textual, and temporal behavioral aspects. Like in the CBS, the same instance of leadership behavior can be classified according to each sublogic at the same time. In this school, validity presupposes high accuracy in capturing behavioral manifestations. Returning to the example of a manager who suggests an alternative work procedure, the episode can be decomposed into multiple neo-behavioral aspects, such as the words used to articulate the suggestion (textual), the timing of the suggestion (temporal), eye gaze or voice pitch (embodiment), patterns of turn-taking or interruption (interaction), and the decision to raise the issue in a one-on-one meeting (choice). Our description of the two research logics goes beyond previous reviews within the CBS (Anderson & Sun, 2017; Eva, Robin, Sendjaya, van Dierendonck, & Liden, 2019; Lemoine et al., 2019) and the neo-behavioral school (Aguinis & Bakker, 2021; Cheng et al., 2022; Eichenseer, 2023; Lehmann-Willenbrock, 2024; Liu, Chambers, & Moore, 2023; Reh, Van Quaquebeke, & Giessner, 2017; Samimi et al., 2022) that take stock of and critically examine research within but not across schools of thought.
Second, we evaluate the research logic of the two schools of thought and find that they have complementary strengths and weaknesses. Categories in the CBS have a cross-contextually generalizable meaning, but they are limited by a tendency to conflate descriptions of what leaders actually do with evaluations of the intentions underlying those behaviors, the quality with which they are enacted, and their presumed effects. As a result, CBS constructs often do not precisely capture the concrete behaviors leaders display. Instead, multiple distinct elements are bundled into a single construct, making it unclear which component is responsible for observed relationships with outcomes. Consequently, two leaders may receive identical CBS scores despite engaging in substantively different behaviors. This lack of behavioral specificity leads to causal indeterminacy, because the same empirical association may reflect differences in behavior or differences in interpretation (Fischer & Sitkin, 2023). Conversely, the neo-behavioral school of thought specifies leadership behavior at the level of concrete form and manifestation and therefore avoids conflating behavioral description with evaluations of intention, performance quality, or effects. By isolating what leaders actually do, NBS constructs capture behavioral variation directly rather than indirectly through evaluative abstractions. As a result, relationships between neo-behavioral constructs and other variables are causally determined, because it is clearer which specific behavioral elements are responsible for observed effects. However, this gain in behavioral specificity comes with important trade-offs. Because NBS constructs deliberately abstract away from evaluations of function and meaning, they may lack cross-contextually generalizable significance. The same behavioral manifestation can acquire different meanings across situations, relationships, and cultural settings. Moreover, by focusing on isolated behavioral aspects, NBS research risks reductionism, as individual elements are removed from the broader situational and relational settings in which leadership behavior unfolds.
These complementary strengths and limitations imply distinct strategies for advancing research within each school of thought. In the CBS, counterfactual categories that separate behavioral description from evaluation can reduce causal indeterminacy, and behavioral classification schemes with functionally independent categories can help address construct redundancy. In the NBS, studying multiple behavioral aspects simultaneously can mitigate concerns of reductionism, and examining contextual moderators can clarify when and how specific behavioral manifestations acquire stable meanings across contexts. Thus, we offer the first general critical appraisal of neo-behavioral leadership research, and we go beyond review-based critiques of the classical school (e.g., Alvesson & Einola, 2019; Banks, Woznyj, & Mansfield, 2023; Fischer & Sitkin, 2023) by examining both the strengths and limitations of the research logic to offer a more balanced assessment. We also build on and extend critiques of construct redundancy to show how and to what extent this limitation can be addressed (Anderson & Sun, 2017; Banks, Gooty, Ross, Williams, & Harrington, 2018).
Third, we offer an integrative framework that specifies how concrete behavioral aspects take on specific meanings and functions across contexts, moderated by the situation, interaction history, and follower characteristics. Thus, the framework accounts for the different positions in the nomological network of the descriptive neo-behavioral aspects versus evaluative classical school-categories. Specifically, neo-behavioral constructs capture what leaders concretely do and thus function as behaviorally specified explanatory variables, whereas classical behavioral constructs capture how such behavior is evaluated in terms of its function and meaning and therefore occupy a downstream position in the causal chain. Furthermore, the framework enables causally determinate and cross-contextually generalizable analysis by explicitly linking de-conflated neo-behavioral constructs to downstream evaluative classical categories. Moreover, meaning-based classifications of CBS constructs allow for cross-contextual generalizability. Thus, the integrative framework shifts the focus from diagnosing problems in leadership research to providing concrete solutions for cumulative theory development.
In the remainder of the article, we first outline the procedure of the meta-synthesis. Then, we describe the research logic of the two schools of thought. Subsequently, we evaluate their inherent strengths and weaknesses and identify school-specific best practice standards. Afterwards, we integrate the two schools and their behavioral categories into a unified framework that leverages the complementary strengths. Finally, we outline a generative agenda for future research to guide theory advancement and empirical testing.
The Procedure of the Meta-Synthesis
Scope of the Meta-Synthesis
Our meta-synthesis takes stock of research that seeks to explain the relationship between leadership behaviors and other variables, particularly their antecedents and consequences. Accordingly, we delimit the scope of the meta-synthesis to explanatory research that seeks “to conceptualize and order topics, and use specific forms of reasoning (such as propositional reasoning) to progressively zoom in on the underlying causal forces or mechanisms that explain the manifestation, dynamics, and outcomes of the topic” (Cornelissen, Höllerer, & Seidl, 2021: 6). Consistent with this focus on causal explanation, interpretivist approaches that prioritize meaning construction over causal inference fall outside the scope of our review. This explanatory focus situates our review within the functionalist research paradigm, which assumes that “society has a concrete, real existence, and a systemic character oriented to produce an ordered and regulated state of affairs” (Morgan, 1980: 608). We examine schools of thought nested in this paradigm as “different ways of approaching and studying a shared reality or world view” (Morgan, 1980: 607). 1 By clarifying the causal relationships linking leadership behaviors, antecedents, and outcomes, our meta-synthesis provides a foundation for practically useful theory and evidence-based management, which depend on credible causal explanations that are also actionable (Dietz, 2026; Rousseau, 2020).
Our meta-synthesis takes stock of and critically examines review articles on leadership behavior research published since the year 2000 in a broad range of management and applied psychology journals. We chose the year 2000 because behavioral leadership research broadened its scope after Yukl’s (1999) critique of charismatic and transactional-transformational leadership. Moreover, Yukl et al.’s (2002) hierarchical taxonomy of leadership behaviors followed shortly thereafter and was an influential synthesis of the classical school of thought. In addition, neo-behavioral leadership research also gained momentum only during this millennium.
Article Identification and Retention Procedure
The article identification and screening procedure unfolded in three steps. First, we identified potentially relevant articles for an initial screening by searching in Web of Science for research that has been published between the years 2000 and 2023 in journals ranked as three-star or better within a broad range of disciplines in the 2023 Academic Journal Guide ranking (known as the Association of Business Schools’ ABS list). We searched for articles that were either a review or a meta-analysis of leadership research, using “(review OR meta) AND leader*” as a keyword combination in the category “topic.” We identified 84 journals and 1,211 articles for initial screening. Second, we screened which of the articles studied leadership behaviors, excluded those that examined unrelated topics, and retrieved 280 articles. Of these, we excluded four articles that were not review articles, 32 articles that, upon closer inspection, were found not to be examining leadership behaviors (e.g., narcissism), and 22 articles that were not explanatory (i.e., research focusing on observer interpretations, but on neither causes nor consequences of leadership behaviors). Third, we systematically coded the remaining 222 articles. We provide a summary of the search and screening process in a PRISMA diagram (see Online Appendix A).
The Coding Procedure
The coded data and Appendix A are available on the Open Science Framework project page (https://osf.io/4bgxt/?view_only=9f43c038f96a463ca67dab69983a2bd4). The coding procedure for review articles had three steps. First, we distinguished whether the articles fit into the classical behavioral or neo-behavioral school of thought by assessing whether the focus was on evaluating the function and meaning of behaviors or describing their form and manifestation. Several review articles include elements associated with both the CBS and the NBS. For example, Fosse, Skogstad, Einarsen, and Martinussen (2019) reviewed the literature on destructive leadership in a military context. Although we classify this review as part of the CBS, it contains isolated references to neo-behavioral elements, such as physical embodiment and behavioral choice, which are mentioned but not examined systematically. Consistent with our coding approach, we classified such articles based on their primary analytical focus.
In addition, and going beyond the systematic coding procedure, we identified qualitatively the logic of how leadership behaviors are conceptualized, measured, and used for causal explanation, and we examined the features and limitations of each school of thought. We did so by revisiting seminal reviews of research in the classical school (e.g., Fleishman, Mumford, Zaccaro, Levin, Korotkin, & Hein, 1991; Lord, 1977; Yukl, 2012; Yukl & Gardner, 2019), comparing their conceptual and explanatory logic with the CBS reviews in our sample and contrasting this logic with the reviews in the NBS. In the description of the research logics of the two schools, we offer additional explanations on how these two schools differ and how this insight is grounded in existing work.
Second, we set out to code articles in the CBS according to the taxonomy of Yukl et al. (2002), which distinguishes task-, relation-, and change-oriented behaviors. In the process of coding articles, however, we gained two insights that led us to revise the taxonomy: (i) We added the category of moral behaviors to reflect the recent rise of moral approaches to studying leadership (Lemoine et al., 2019). (ii) Unlike in previous research, we find that the categories are independent sublogics of classifying behaviors rather than mutually exclusive categories for describing behaviors (Lemoine et al., 2019; Stogdill, 1962; Yukl, 2012). In line with our reasoning that any single behavior can serve multiple functions and have multiple meanings, we allowed review articles and leadership behavior constructs to belong to multiple categories. For example, we coded transformational leadership as relations-, change-, and morally oriented; ethical leadership as relations- and morally oriented; servant leadership as relations- and morally oriented, contingent reward behaviors as task- and relations-oriented; and destructive leadership as (negatively) relations- and (im-)morally oriented. Using a 10-review article subsample (total k = 206) in the CBS category, coding percent agreement was 90% and Cohen’s kappa = 0.80 across two coders, which is a conservative estimate that accounts for the risk of agreement by chance. Fleiss, Levin, and Paik (1981) suggested that a kappa < 0.40 is poor, 0.40 to 0.75 is acceptable, and 0.75 or higher is excellent.
Third, in the neo-behavioral school of thought, we identified a five-fold scheme that covers choice, embodiment, interaction, textual, and temporal aspects of leadership behaviors. Coding agreement was 83% and Cohen’s kappa = 0.66 across a 20% subset of review articles in the NBS (acceptable by Fleiss et al.’s, 1981, recommendations). All review articles in the NBS category were double-coded (k = 16), and any disagreements were resolved through discussion. We also conducted open coding of the antecedents and consequences of leadership behaviors across the two schools of thought, and additional open coding for primary NBS articles (e.g., disciplinary origins). As in the CBS, some articles in the NBS were classified into multiple categories (e.g., focusing on the embodiment and temporal aspects of behavior).
Describing the Two Distinct Schools of Thought
We now discuss how the CBS and NBS differ in the research logics that underlie their conceptualization and measurement of behaviors and causal explanations. Figure 1 provides an overview.
A Descriptive Overview of the Research Logic of the Two Schools of Thought
Describing the Research Logic of the Classical Behavioral School of Thought
General Research Logic of the CBS
Exemplary CBS constructs are leadership styles such as ethical or transformational leadership and single behaviors such as giving or seeking feedback or influencing others. In conceptual terms, behaviors in the CBS are classified based on their latent functions and meanings. Although the behavior itself is observable, it is necessary to judge whether it meets the specified function and meaning. For example, a leader may suggest an alternative work procedure, which is an observable act. Yet, whether these behaviors should be classified as problem-solving, supporting followers, or rather interfering with and micro-managing the employee requires an evaluative judgment. This judgment typically considers the intended or actual effect of the behavior on the employee, and these effects are typically valenced, relating to either a positive or a negative function and meaning. This depiction is in line with Tepper’s (2000) conceptualization of abusive supervision as a subjective evaluation because “[t]he same individual could view a supervisor’s behavior as abusive in one context and as nonabusive in another context, and two subordinates could differ in their evaluations of the same supervisor’s behavior” (p. 178). This logic implies that CBS constructs primarily capture how leadership behavior is evaluated rather than specifying which concrete behaviors leaders should enact in a given situation. As a result, CBS categories are well-suited for explaining why leadership behaviors have certain effects but are less informative for deriving prescriptions at the level of specific behavioral acts.
The logic of functional classification in research becomes apparent when revisiting seminal reviews of the CBS. For example, Lord (1977) classified leadership behaviors into two overarching functions (task-related and socioemotionally related), as well as into more specific functional categories that are nested within them, such as “[f]acilitating information exchange” and “[r]educing or preventing conflict” (p. 117). Concrete behaviors that foster these functions are then called functional behaviors. Similarly, Fleishman et al. (1991) offered a taxonomy that is explicitly functional too, defining leadership behavior as “a form of organizationally-based problem solving, implemented in a social context, where an attempt is made to bring about goal attainment by influencing the actions of other subsystems” (pp. 258–259; italics in original). Their leadership behavior categories, such as “Communicating Information” or “Motivating Personnel Resources” (pp. 260–261), reflect this functional logic. Likewise, Yukl et al. (2002) classified behaviors based on the primary purpose they serve, yet go beyond Lord (1977) and Fleishman et al. (1991) by adding a change-oriented category that captures the “way leaders initiate and implement change in organizations” (p. 16).
The functional logic also characterizes CBS research on moral leadership behaviors. For example, ethical leadership includes “the promotion of such [i.e., normatively appropriate] conduct to followers” (Brown, Treviño, & Harrison, 2005: 120) and servant leadership entails “[the leader’s] moral responsibility not only to the success of the organization, but also to his or her subordinates, the organization’s customers, and other organizational stakeholders” (quoted in Lemoine et al., 2019: 151). The logic is slightly different for authentic leadership. Exemplary behavior subcategories such as “internalized moral perspective” or “self-awareness” (Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008) reflect the meaning of the behavior to the observer, but not necessarily its function. Across these streams, however, the overarching logic remains the same: raters infer latent functions or meanings from observed behavior.
Yukl and Gardner (2019) emphasized that behavior categories “are derived from observed behavior [. . .] but they do not exist in any objective sense” (pp. 42–43). This does not imply that categorizing leadership behaviors would be arbitrary. Instead, CBS research typically treats intersubjective agreement as a validation ideal, such that it is desirable for observers to converge in their evaluation of a behavior’s function or meaning (e.g., whether a leader’s suggestion is judged helpful or undesirable interference).
In terms of measurement, psychometric procedures reflect the focus on intersubjectivity. CBS constructs are validated through interrater agreement and through discriminant and criterion-related validity, which assess whether agreed-upon evaluations are empirically distinct and related to outcomes such as performance or satisfaction. These procedures illustrate that CBS research prioritizes intersubjective agreement over strict objectivity. At the same time, CBS research is not primarily concerned with explaining systematic variation in individual interpretations of behavior, which is the focus of interpretivist leadership research (Fischer & Alvesson, 2025; Smircich & Morgan, 1982).
Causal explanation in the CBS is therefore grounded in evaluations of a behavior’s function and meaning. For example, what matters is whether a leader’s behavior is evaluated as supportive—and thus high on relations orientation—rather than whether this evaluation arises from suggesting alternative work procedures or from expressing trust in the employee’s existing approach. Accordingly, CBS research neither attributes causal status to specific behavioral manifestations (as in an objectivist, neo-behavioral logic) nor to idiosyncratic, observer-specific interpretations (as in a subjectivist or interpretivist logic). Rather, causality in the CBS operates through evaluations of behavior, such that leadership effects arise from how behaviors are classified in terms of their function and meaning.
Sublogics and behavioral categories of the CBS
Although all CBS constructs follow the same general logic of classifying behaviors according to evaluations of their function and meaning, the specific functions and meanings emphasized vary. Building on previous research, we classify leadership behavior along four dimensions—namely tasks, relations, change, and moral issues. According to Yukl (2012), “the primary purpose of task-oriented behaviors is to ensure that people, equipment, and other resources are used in an efficient way to accomplish the mission of a group or organization” (p. 69); relations-oriented behaviors are displayed “to enhance member skills, the leader-member relationship, identification with the work unit or organization, and commitment to the mission” (p. 71); and change-oriented behaviors have the goal “to increase innovation, collective learning, and adaptation to external change” (p. 72). The fourth dimension, morality, captures the ethical considerations underlying leadership behavior, including consequentialist, deontological, and virtue-based evaluations (Lemoine et al., 2019).
The usefulness of this four-fold typology lies not in its claim to represent leadership behavior exhaustively, but in its conceptual value as an evaluative abstraction that helps make sense of leadership phenomena; accordingly, no single classification scheme can be considered “true” independent of its purpose (Yukl & Gardner, 2019). Historically, the identification of task- and relations-oriented leadership behaviors took place in the post-war era that was characterized by hierarchical organizational forms and relatively stable institutional environments. Over the last decades, however, change orientation has gained prominence as organizations flattened and competitive pressures increased (Zhang, 2023). Similarly, the weakening of institutionalized normative rules has heightened the relevance of moral leadership (Brown et al., 2005). Together, these developments underscore why the change and moral dimensions are essential for understanding leadership in contemporary contexts. Importantly, these four dimensions are not orthogonal. A single leadership behavior can simultaneously serve multiple functions and meanings, a point to which we return later.
In total, we identified 206 (92.79%) articles that fall into the classical school of thought. Of these, 135 (65.53%) fall into the task category, 200 (97.09%) into the relations category, 107 (51.94%) into the change category, and 164 (79.61%) into the moral one. Thus, most articles bridge several categories. A reason for this is that research in the CBS is dominated by broad constructs of leadership styles. For example, transactional and transformational leadership are intended to capture the full range of leadership (Antonakis, Avolio, & Sivasubramaniam, 2003) and thus cut across dimensions such as relations, change, and moral orientation.
Describing the Research Logic of the Neo-Behavioral School of Thought
General research logic of the NBS
In conceptual terms, the NBS represents concrete and separate aspects of behaviors directly based on their observed form and manifestation. An NBS construct focuses on a single behavioral aspect (e.g., choices or decisions, facial expressions, or speaking time) and, unless studied multimodally, abstracts from other aspects of the behavior. As a result, NBS research adopts a deliberately reductionist focus, which helps to avoid conflating the description of behaviors with the evaluation of whether the behavior was intended to or successful in pursuing a certain function.
In terms of measurement, NBS research pursues objectivity as a validation ideal by aiming to capture behavioral aspects as accurately as possible, for example, through unambiguous choices (e.g., degree of pro-sociality in economic games; executive orders regarding mergers and acquisitions) or technological assistance (e.g., eye-trackers). Accordingly, each behavioral aspect is treated as having an observer-independent value, such as the number of times a “we” or “I” pronoun has been used (Fladerer et al., 2021), the speaking time a person had (MacLaren et al., 2020), or the type of facial expressions this person displayed (e.g., a raised or lowered eyebrow line; Reh et al., 2017). In this logic, intersubjective agreement is not the primary validation criterion, and a single accurate measurement is, in principle, sufficient. At the same time, objectivity remains an aspiration rather than a guarantee. For example, interactional aspects are often captured through behavioral coding, which can reintroduce subjectivity when abstract coding schemes require judgment calls (Güntner, Meinecke, & Lüders, 2023). 2
In terms of causal explanation, the focus is on the form and manifestation of the behavior and not which function it serves or the meaning it has. Although NBS research remains rooted in a functionalist paradigm insofar as it examines the effects of behavioral aspects on outcomes of interest (cf. Morgan, 1980), its functionalism is deliberately circumscribed. Specifically, functionalist reasoning is applied to the causal link between observed behaviors and outcomes, but not to the conceptual classification of behaviors themselves. For example, NBS research examines the causal effects of raising one’s voice as a specific behavioral manifestation, rather than of “abusive supervision” as a higher-order construct defined by inferred intent or function.
As a clarification, we would like to highlight that the neo-behavioral school is unrelated to the behaviorist approach in the tradition of classical or operant conditioning. Behaviorism in psychology is a theory of learning and explanation that emphasizes environmental contingencies and reinforcement histories while downplaying the role of mental states and meaning (Locke & Latham, 2002). By contrast, the distinction we draw between the neo-behavioral and classical behavioral schools does not concern how behaviors are learned or what causes them. Rather, it concerns different logics of behavior classification. Both schools of thought are compatible with cognitively rich, intention-based theories of leadership; they differ not in their psychological assumptions but in how behavior is conceptualized and incorporated into causal explanation.
Sublogics and behavioral aspects of the NBS
We distinguish five aspects in the neo-behavioral school of thought—namely, the choice, embodiment, interaction, textual, and temporal ones that we depict in Table 1. We identified these behavioral aspects in an abductive process, using the classification schemes of Banks et al. (2023) and Fischer (2023) as a starting point. During the coding of articles, we settled on the five categories mentioned. In total, we identified 16 articles (7.21%) that fall into the neo-behavioral school of thought. Of these, 6 articles (37.50%) fall into the choice category, 13 articles (81.25%) into the embodiment category, 8 articles (50.00%) into the interaction category, 11 articles (68.75%) into the textual category, and 7 articles (43.75%) into the temporal category.
Describing the Behavioral Categories in the Neo-Behavioral Schools of Thought
Note: The different dimensions of this taxonomy co-occur and are not mutually exclusive. Multimodal research that cuts across these categories would be particularly valuable (see also Lehmann-Willenbrock & Hung, 2023).
First, the choice category focuses on leadership decisions. At the micro level, research examines consequential choices in economics games, such as resource allocation in public goods games (Eichenseer, 2023). At the macro level, scholars study strategic decisions made by executives, including foreign direct investment choices (Herrmann & Datta, 2006; Samimi et al., 2022). This stream draws primarily on behavioral economics (Smith, 1982) and strategic management (Hambrick & Mason, 1984), and typically relies on experimental or archival data.
Second, the embodiment category captures nonverbal aspects of leadership, such as gestures, eye gaze, and vocal features. For example, research has examined how entrepreneurs use hand gestures during pitch presentations to influence audiences (Clarke, Llewellyn, Cornelissen, & Viney, 2021) or how leaders’ eye gaze relates to perceptions of charisma (Maran et al., 2019). This work is rooted mainly in psychology and physiology, with early foundations in research on facial expressions and emotion (Ekman et al., 1969).
Third, interaction aspects concern behavioral patterns that emerge in social exchange, such as turn-taking, interruptions, and speaking networks. These behaviors are often analyzed using social network approaches (Carter et al., 2015). Foundational work includes early interaction coding schemes for small groups (Bales, 1950), alongside later developments in social network analysis (Borgatti et al., 2009).
Fourth, the textual category focuses on linguistic features of leadership communication, including pronoun use and sentiment (Liu et al., 2023). While rooted in communication studies, contemporary work draws heavily on computational linguistics and data science. For instance, scholars used text from CEO letters or annual reports to assess entrepreneurial orientation or social performance (Short, McKenny, & Reid, 2018; Smulowitz, Pfarrer, Cossin, & Lu, 2025).
Fifth, temporal aspects examine the timing, duration, and sequencing of leadership behavior. Examples include speaking time during meetings as a predictor of leader emergence (MacLaren et al., 2020) or the speed of information sharing as a cue for ethical leadership (Van de Calseyde, Evans, & Demerouti, 2021). This stream integrates perspectives from organizational psychology and data science, with recent reviews synthesizing how duration, frequency, timing, and behavioral sequences shape leadership processes (Aguinis & Bakker, 2021).
Overall Mapping of the Field of Research
In our meta-synthesis, 206 (92.79%) of the review articles are part of the CBS, while 16 (7.21%) are part of the NBS. The relative dominance of the classical behavioral school of thought reflects the NBS’s nascent status compared to the CBS’s established status. Nevertheless, the neo-behavioral review articles in our sample are based on hundreds of primary studies in the natural (e.g., evolutionary biology) and social (e.g., psychology, anthropology) sciences, indicating the growing importance of this school of thought in academia. Overall, our coding shows that all reviews fall into one of the two schools of thought.
Evaluating the Features and Limitations of the Two Schools of Thought
In this section, we assess the distinctive features and potential risks associated with the two schools of thought (see Table 2 for an overview). We also identify best practices specific to each school that help mitigate these risks. Finally, we outline the limitations inherent to each school that can be only overcome through cross-school integration.
Evaluating the Research Logic within the Two Schools of Thought
Evaluating the Classical School of Thought
Strength of the CBS
The strength of the function- and meaning-based classification in the classical school of thought is the generalizability of behavior categories across contexts. Hannah, Sumanth, Lester, and Cavarretta (2014) exemplified this point when arguing that the meaning of more general specifications remains stable across contexts for the ethical leadership item “Conducts his/her personal life in an ethical manner” (Brown et al., 2005: 125) but not for more specifically defined behaviors, such as charitable or self-sacrificing acts that would fall into the NBS. According to Hannah et al. (2014), more specific acts are highly prescriptive and may be ethical in one context but not another (e.g., judged as tax evasion or virtue signaling instead). Similarly, Rook, Leroy, Zhu, and Anisman-Razin (2025) emphasized that there are many ways to act authentically, yet that there is a shared deeper meaning of authenticity, even though different concrete behaviors are required to achieve it. In other words, CBS categories are cross-contextual, while the specific behaviors that match these categories are context-contingent.
Risks of the CBS
First, there is the risk of conflation and causal indeterminacy, which has been outlined by Fischer and Sitkin (2023) in their general review of leadership style research. They argued that general item formulations lead to conflating descriptions of behaviors with positive or negative evaluations of intentions, performance (i.e., quality of execution), and effects. As a result, such constructs nominally refer to behavior but largely capture evaluative judgments. For instance, leading one’s life ethically, which is an ethical leadership item (Brown et al., 2005), describes not only what a leader is doing but also that the behavior is driven by ethical intentions, is performed in an ethically appropriate manner, or has ethical consequences. In this sense, CBS constructs are not sufficiently behaviorally specified explanatory variables but also evaluative outcomes. Fischer and Sitkin (2023) concluded that this conflation leads to causal indeterminacy because it is unclear whether the relationships between leadership style constructs and outcomes of interest are driven by leader behaviors or by followers’ subjective evaluations. Review-based critiques of single leadership styles come to similar conclusions (e.g., Banks, Fischer, Gooty, & Stock, 2021; Fischer, Tian, Lee, & Hughes, 2021; Van Knippenberg & Sitkin, 2013). This insight aligns with measurement-oriented work, which emphasizes the fact that the function or meaning of a behavior cannot be directly observed but only judged because it taps mostly into the rather evaluative semantic memory and less into the rather descriptive episodic memory (Hansbrough, Lord, Schyns, Foti, Liden, & Acton, 2021). Accordingly, the strengths and risks of the CBS are closely intertwined: the same evaluative abstraction that enables cross-contextual generalization also blurs the underlying behavioral acts required for a causally determinate explanation.
Second, there are also several calls for research in the CBS to limit construct overlap (Anderson & Sun, 2017; Banks et al., 2018). Commonly, overlapping or highly similar item formulations are considered the source of construct redundancy. However, our analysis of the CBS research logic suggests an additional explanation: the same concrete behavior can fall into multiple dimensions of leadership behaviors, even if these dimensions are conceptually distinct. This interpretation departs from much CBS research, including Stogdill (1962), Yukl (2012), and Lemoine et al. (2019), which tends to treat these dimensions as describing different types of leadership behaviors. Consider, for example, the example introduced earlier of a manager who suggests an alternative work procedure to an employee. We have seen that this behavior can be classified both as high on task, relations, and change orientation. This multiplicity does not imply conceptual overlap among the dimensions. Rather, the same behavioral episode simultaneously serves multiple functions. Thus, leadership behavior can be evaluated along each dimension independently. Consequently, correlations among CBS dimensions do not necessarily indicate poorly differentiated constructs. Instead, they may reflect the multifunctional nature of leadership behavior itself. Accordingly, the well-documented empirical overlap across classical leadership behavior constructs (Banks, McCauley, Gardner, & Guler, 2016; Hoch, Bommer, Dulebohn, & Wu, 2018) can stem from two different sources: first, function- and meaning-based overlaps in item formulations, and second, leadership behaviors or styles serving multiple functions or meanings. In the latter case, empirical overlap is theoretically meaningful in terms of the CBS rather than an artifact of functionally indiscriminate categorization.
Best practices and inherent limitations in the CBS
Our evaluation of the research logic of the classical behavioral school of thought (CBS) allows us to distinguish between limitations that are structural to function- and meaning-based classification and those that can be mitigated through school-specific best practice standards. First, developing counterfactual constructs and items that either describe or evaluate—but do not mix description with evaluation—can reduce the risk of conflation and causal indeterminacy. As Fischer (2023) outlined, many leadership behavior constructs in the CBS lack counterfactuals because they rely on double-barreled, interpretive, or nonbehavioral specifications, thereby inviting evaluations already at the level of construct definition. However, even with counterfactual items, an important limitation remains. Although counterfactual specifications can prevent conceptual conflation, they cannot fully prevent empirical conflation at the level of rater cognition. For example, an item such as “Schedules the work to be done” (Neubert, Kacmar, Carlson, Chonko, & Roberts, 2008: 1233) is counterfactual and descriptive, yet raters may still assign higher scores to successful than unsuccessful managers. This pattern is consistent with evidence that rater judgments draw on evaluative semantic memory in addition to episodic recall (Gioia & Sims, 1985). Raters may not only use their episodic memory that refers to behavior descriptions but also their semantic memory that draws on behavior evaluations (Hansbrough et al., 2021). Thus, counterfactual constructs can reduce, but not eliminate, causal indeterminacy.
Second, not all instances of construct overlap in the CBS reflect limitations that can be fully remedied through improved construct specification. While some function- and meaning-based overlap can be reduced through more functionally distinctive classification schemes, other overlap is inherent to leadership behavior itself. Anderson and Sun (2024) outlined how leadership style research can draw on insights from personality research—most notably the development of the Big Five—to clarify functional distinctions among constructs. In the moral dimension, for example, Lemoine et al. (2019) refined the conceptualization of authentic, ethical, and servant leadership by distinguishing virtue-based, deontological, and consequentialist orientations, thereby reducing conceptual overlap. More broadly, the four-dimensional taxonomy builds on previous research (e.g., Lemoine et al., 2019; Yukl et al., 2002) and represents a further step toward conceptual clarity through functional differentiation. However, even when functional dimensions are conceptually distinct, they may overlap empirically because the same behavior can serve multiple functions simultaneously. Such overlap is neither an artifact of poorly specified categories nor a consequence of response bias but may reflect the multifunctional nature of leadership behavior itself. Accordingly, functional distinctiveness is best understood as a best practice for reducing artifactual overlap, rather than as a means of eliminating empirically meaningful overlap altogether.
Finally, the multifunctional nature of leadership behaviors poses challenges for psychometric procedures such as confirmatory factor analysis (CFA) when applied to function-based behavioral classifications. CFA assumes that observable indicators load on a single latent variable (Furr, 2021), an assumption that is difficult to reconcile with behaviors that legitimately serve multiple functions. One possible mitigation strategy is to ask raters to classify behaviors based on their primary function only, such that each behavior maps onto a single dominant category. Conceptually, this approach restores unidimensionality. Practically, however, it is often unclear which function is primary, and raters may disagree. In the case of a manager who publicly challenges an employee’s idea, it is not obvious whether the act primarily clarifies expectations, develops skills, or encourages innovation. Asking for a primary function thus constitutes only a partial remedy. More generally, these challenges parallel those faced by personality researchers in modeling the Big Five, where even functionally distinct dimensions exhibit empirical overlap (Marsh et al., 2010). These considerations suggest that some limitations of CBS measurement reflect inherent tensions between evaluative classification and the multifunctional nature of leadership behavior, rather than remediable flaws in research design.
Evaluating the Neo-Behavioral School of Thought
Strength of the NBS
The form- and manifestation-based approach avoids conflation and the related risk of causal indeterminacy (Fischer, 2023). By directly representing observable behavioral acts—such as specific words used, timing of interventions, gaze patterns, or concrete decisions—the NBS captures what leaders actually do, rather than how those actions are subsequently evaluated. This strength is of high importance to address growing concerns that leadership research does not study concrete behaviors enough (Banks et al., 2023) and that the conclusions drawn from existing research evidence may be limited by severe flaws (Fischer & Sitkin, 2023; Van Knippenberg & Sitkin, 2013).
Importantly, NBS research specifies the behavioral building blocks through which function- and meaning-based leadership categories are enacted. In doing so, neo-behavioral constructs provide the kind of causally specified inputs required for actionable theories, because they identify the concrete actions and behavioral manifestations through which leadership functions are realized (Dietz, 2026). Rather than classifying behavior as task-oriented or ethical, NBS research pinpoints the specific communicative, interactional, or decision-related behaviors that give rise to such classifications. As a result, neo-behavioral constructs enable more causally determinate tests of leadership effects and offer a direct bridge from abstract leadership categories to actionable guidance at the level of observable conduct.
Risks of the NBS
This approach comes with important risks, too. First, capturing behaviors counterfactually implies a lack of cross-contextually generalizable meaning because a specific behavior can have a different meaning across situations (see, e.g., Hannah et al., 2014; Tepper, 2000). For example, Bastardoz, Jacquart, and Antonakis (2024) demonstrated that the same behaviors are evaluated differently in crisis versus stable situations. Fischer and Alvesson (2025) took an interpretivist and socially constructivist stance to identify and theorize inherent interpretive ambiguity in concrete NBS-type leadership behaviors. Similarly, a constant eye-gaze can signal either attentiveness or intrusive social behavior (Cheng et al., 2022). The underlying reason is the reductionism inherent in research aimed at objectivity (Burrell & Morgan, 2019 [1979]) and the a-contextual nature of conceptualizing and measuring behaviors in the NBS.
Second, the objectivist ideal and reductionist approach of the NBS may miss the contextual complexity of leadership. For instance, NBS scholars might seek to study an arm gesture, yet when considering the case of Elon Musk’s gesture that was interpreted by many as a Nazi salute (Guardian, 2025), merely capturing the movement is unlikely to shed light on the social meaning. NBS scholars who only measure the gesture, without taking into consideration the political context, would miss the rich complexity around what the behavior may mean. Similarly, not all NBS-type conceptualizations of leader behavior achieve the same level of objectivity. For example, Ekman categorized 18 different types of smiles, from a forced smile to a spontaneous smile to a genuine smile. The determination of the names of these different smiles has a subjective element. The muscle movements can be objectively measured. In the case of a “genuine smile,” Ekman measured crinkles in the orbicularis oculi muscle, also referred to as the “crow’s feet,” around one’s eyes (Gabriel, 2025). As NBS research advances, it is important to remember that the goal is to improve the objective measurement of these specific muscles, and that this is distinct from what it means to observers (i.e., the smile is genuine). Commercial tools, such as those by Microsoft and Amazon, do not perform perfectly when evaluated for recall (“of all the smile images shown, how many were correctly classified?”) and precision (“of all the predicted smiles, how many were correct?”). Thus, there is an important warning that, for example, interaction coding, which is commonly considered objective, invites considerable degrees of subjectivity, especially if the coding categories are abstract (Güntner et al., 2023).
Best practices and inherent limitations in the NBS
Our evaluation of the research logic of the NBS enables us to differentiate between limitations that are inherent and those that can be mitigated by school-specific best-practice standards. First, studying contextual moderators increases our knowledge of the cross-contextual generalizability of NBS research. For example, examining the conditions under which an extended eye gaze is associated with positive or negative follower reactions reveals how the effects of a given behavioral manifestation vary across situations, thereby clarifying the scope conditions under which generalizations are warranted. Still, there is a potentially unbounded set of relevant contextual moderators, and no research design can exhaustively account for all of them; accordingly, limited cross-contextual generalizability remains an inherent constraint of behaviorally specific constructs.
Second, multimodal research that studies the interplay of multiple rather than isolated aspects can mitigate the risk of reductionism. Such research is highly feasible given recent advances in the study of behaviors as multimodal signals (Lehmann-Willenbrock & Hung, 2023). Returning to the example of a manager who suggests an alternative work procedure, followers do not attend to a single behavioral cue in isolation. Instead, their inferences are based on triangulating multiple behavioral signals. In this episode, the leader uses particular words (textual), delivers them with a specific tone of voice and facial expression (embodiment), speaks up at a particular moment and for a certain duration (temporal), and chooses to address the issue privately rather than publicly (choice). These signals occur in response to the employee’s behavior and jointly constitute a behavioral episode (interaction). Even when studied multimodally, however, such analyses remain confined to describing behavioral form and manifestation. This preserves the core strength of the NBS—nonconflated and causally determinate description—while also retaining an inherent limitation: the absence of function- and meaning-based abstraction required for cross-contextual generalization.
An Integrative Framework of Leadership Behaviors
This section introduces an integrative framework that combines the behavioral specificity of the neo-behavioral school with the meaning-based abstraction of the classical behavioral school. By assigning distinct causal roles to concrete behaviors and to evaluations of their function and meaning, the framework explains how observable actions become meaningful leadership behaviors across contexts. We first outline the framework’s logic of causal explanation, then discuss how it captures the dynamic, episodic nature of leadership, and finally show how integrating the two schools overcomes limitations that persist when either is studied in isolation—laying the foundation for the generative research agenda presented in the next section.
Causal Explanations in the Integrative Framework
The integrative framework explains five causal links (see Figure 2): (1) how neo-behavioral aspects from the NBS influence evaluations of functions and meanings of behaviors in the CBS; (2) how this link is moderated by contextual variables, namely situational, follower, and relational characteristics; (3) how people adapt their display of behavioral aspects from the NBS based on the intended meaning from the CBS that is also context-contingent; (4) how behaviors—both in terms of their function and meaning as well as form and manifestation—have distinct antecedents and consequences due to their different position in the nomological network; and (5) how outcomes of leadership, such as follower performance or satisfaction, can influence evaluations of the function and meaning of behaviors.
An Integrative Framework of Leadership Behaviors
First, neo-behavioral forms and manifestations of behaviors systematically influence evaluations of the function and meaning of behaviors in the CBS. Drawing upon the meta-synthesis, the NBS is composed of behaviors that represent choices, embodiment, interaction, textual, and temporal aspects of leader behavior. Followers often experience all of these behaviors simultaneously and over the course of time (Hemshorn de Sanchez, Gerpott, & Lehmann-Willenbrock, 2022). Accordingly, concrete behaviors are ultimately judged regarding what they mean in terms of the CBS logic. That is, drawing upon the CBS reviews, we know that observers judge how well leaders behave as well as which effects these behaviors have on various outcomes. In this sense, CBS constructs attach causal status to evaluations of behavior, whereas NBS constructs attach causal status to the behaviors themselves.
Second, the link between concrete behaviors and follower evaluations of those behaviors may be context-dependent rather than deterministic. According to the review by Johns (2006), context refers to the “situational opportunities and constraints that affect the occurrence and meaning of organizational behavior as well as functional relationships between variables” (Johns, 2006: 386). That is, context alters the meaning of behaviors, such that the same neo-behavioral aspect could be evaluated differently across contexts. The leadership context may comprise the work event (Morgeson, Mitchell, & Liu, 2015), characteristics of the follower (Wang, Van Iddekinge, Zhang, & Bishoff, 2019), and the interaction history between the leader and follower (Gottfredson, Wright, & Heaphy, 2020). More generally, subjective evaluations of the meaning and function of concrete neo-behavioral aspects are influenced jointly by behavioral manifestations and the contexts in which they are displayed.
Third, context influences people’s display of neo-behavioral aspects because they choose their behaviors based on the anticipated meaning and function. This implies that across contexts, different neo-behavioral aspects need to be displayed to express the same function and meaning. The path-goal theory of leadership makes this argument when outlining that leadership behaviors are chosen to support follower performance, which calls for different leadership behaviors depending on the context (House, 1996). The logic is that when there are already supportive structures in place, there is less need for displaying interpersonal support—a point made by the closely related substitutes of leadership theory (Kerr & Jermier, 1978). For instance, theorizing on destructive leadership suggests that leaders assess follower susceptibility and situational constraints before adapting their behavioral signals accordingly (Padilla, Hogan, & Kaiser, 2007).
Fourth, CBS research suggests that evaluations of leadership behaviors can have a distinct impact on leadership outcomes, for example, by influencing leader and follower behavior, which subsequently affects more distal outcomes (Güntner, Klonek, Lehmann-Willenbrock, & Kauffeld, 2020; Wang et al., 2019). Although such sequences are theoretically compelling, empirically testing the full causal chain remains challenging because it requires observing both behavioral manifestations from the NBS and evaluative processes from the CBS.
Fifth, there is a recursive causal link such that outcomes of leader behaviors influence evaluations of the function and meaning of behaviors. This logic parallels the halo effect (Nisbett & Wilson, 1977), whereby positive impressions in one domain influence judgments in another. Leadership outcomes—ranging from performance to ethical achievements—can thus retroactively shape how later behaviors are interpreted.
Dynamic Nature of the Integrative Framework
There is a dynamic element to all causal steps of the integrative framework. First, every interaction between a leader and a follower alters the interaction history. In turn, changes in interaction history influence how neo-behavioral aspects are interpreted by followers (Step 2) and, via leaders’ anticipation of follower interpretations, influence which neo-behavioral aspects they display in the first place (Step 3). Second, at any moment, events in the environment (Morgeson et al., 2015) can change the evaluation of leaders’ behaviors. What may be considered appropriate at one moment may later be reinterpreted as inappropriate given new information or a changing context (Step 2), and leaders may adjust their behavior accordingly (Step 3).
This within-person dynamic represents the episodic nature of leader behaviors (Kelemen, Matthews, & Breevaart, 2020), and our integrative framework helps shed light on how and why leaders adapt their behaviors. Such a dynamic view avoids missing within-person variation and is particularly relevant when an average or stable behavioral style may not manifest (Alvesson, 2019). For instance, evaluations of ethical leadership may reflect a pattern of episodes rather than a stable style, and discrete behaviors may be evaluated at different times as ethical or as abusive.
How the Integrative Framework Overcomes Limitations Within Schools of Thought
We outline how the integrative framework leverages the complementary strengths of the two schools to address limitations that persist when either is studied in isolation. First, distinguishing between NBS aspects and CBS categories constitutes de-conflation by separating behavioral description from evaluative classification. Importantly, this distinction clarifies that CBS constructs are not fully actionable independent variables (Dietz, 2026) but point evaluative mechanisms that profit from behavioral specification through NBS constructs. Such de-conflation enables causally determinate analysis of leadership effects (Fischer & Sitkin, 2023). For example, Jensen et al. (2023) showed that value-laden rhetorical tactics increase evaluations of leader charisma, which in turn predict follower compliance.
Second, the integrative framework advances theory by specifying how behavioral aspects, evaluations, and context jointly shape leadership processes (Steps 1–3 in Figure 2). CBS and NBS constructs occupy different positions in the nomological network, such that neo-behavioral aspects function as action-level causes, whereas classical school constructs refer to evaluative mechanisms related to intentions, execution, and effects. For example, a leader who intends to act charismatically (CBS, intention level) may deploy specific rhetorical tactics, such as the use of metaphors or value-laden language (NBS, textual; Antonakis et al., 2022), as well as distinctive nonverbal cues, such as eye gaze or facial expressions (NBS, embodiment; Maran et al., 2019). These behavioral manifestations, in turn, shape followers’ evaluations of charisma (CBS, effect level). Contextual factors—such as culture, situational demands, or value congruence—may moderate which behavioral aspects are displayed and how they are evaluated (Ernst et al., 2022).
Third, integrating the NBS and CBS acknowledges the socially embedded nature of meaning formation. This integration mitigates reductionism and a-contextuality in NBS research and addresses the context deficit in CBS research. As Johns (2023) argued, evaluative confounding in leadership research hampers the detection of contextual moderators. Integrative studies—such as Wilms, Bastardoz, El Dahan, and Jacquart (2024), who examined value congruence in charismatic rhetoric—illustrate how combining behavioral description and evaluative classification enables more context-sensitive leadership research.
A Generative Agenda Toward an Integrative Study of Leadership Behaviors
Our school-specific best practices and the integrative framework suggest eight specific theoretical pathways for future research that we bundle in four broader sets and summarize in Table 3. Next, we first outline how to conduct research across the two schools of thought and along the integrative framework, then how to advance research within the two schools of thought, and ultimately how to draw on interpretivist scholarship to examine in more detail how the same leadership behaviors can be evaluated differently.
A Generative Future Research Agenda Toward an Integrative Study of Leadership Behaviors
Across-School Research Along the Integrative Framework
Investigating the relationship between the two schools of thought
As a first set of theoretical pathways, we propose that leadership scholars investigate relationships across the two schools by theorizing and testing the core causal links specified in the integrative framework (Steps 1–3 in Figure 2). Step 1 focuses on meaning formation—that is, how neo-behavioral aspects give rise to evaluations of leadership function and meaning in the CBS. For example, Antonakis, Fenley, and Liechti (2011) showed that a set of predominantly verbal rhetorical tactics leads to systematically higher evaluations of leadership charisma. This research logic can be extended to other neo-behavioral aspects (choice, embodiment, interaction, temporal) to examine how concrete behavioral manifestations translate into meaning for observers. Such behaviors may be examined as single episodes or as recurring patterns that shape leader reputations over time (Hemshorn de Sanchez et al., 2022).
Step 2 addresses contextual contingency. The same behavior can be evaluated differently depending on situational features, follower characteristics, or interaction history (Johns, 2023). For example, rhetorical tactics often work better in crises or when people share values (Wilms et al., 2024). Signaling theory provides a useful lens for theorizing such contingencies by studying behaviors as signals whose interpretation depends on visibility, fit, timing, and receiver attention (Connelly, Certo, Reutzel, DesJardine, & Zhou, 2025). Complementarily, social information processing theory emphasizes how contextual cues and norms shape how behaviors are construed and evaluated (Kelemen, Matthews, Whitney, & Matthews, 2025; Salancik & Pfeffer, 1978).
Step 3 centers on behavioral selection—that is, which neo-behavioral aspects leaders display when they intend to enact particular leadership functions or meanings, and how this selection varies across contexts. For example, Bastardoz et al. (2024) showed that leaders use more value-laden language in crises, raising questions about the intentions underlying such behavioral adaptations. Integrating intention attribution into this step remains a critical but underdeveloped avenue for future research because perceived leader intention shapes follower reactions (for reviews, see Bastardoz & Van Vugt, 2019; Güntner et al., 2020). Taken together, this line of theorizing highlights the need for empirical designs that explicitly connect leaders’ intended meanings, their concrete behavioral displays, and followers’ subsequent evaluations.
Such integrative designs already exist in a small but informative subset of studies. Specifically, we conducted a supplemental review of primary studies in the neo-behavioral school, which revealed that although most studies operate exclusively within the NBS, some explicitly bridge NBS and CBS elements. These studies exemplify the type of cross-school research called for here and are summarized in Online Appendix B. For example, Antonakis et al. (2011) linked charismatic leadership tactics (NBS: textual, embodiment) to higher evaluations of charisma (CBS: relations, moral). MacLaren et al. (2020) showed that longer speaking duration (NBS: temporal) predicted leader emergence ratings (CBS: task, relations). Van de Calseyde et al. (2021) demonstrated that faster inclusion of others in consequential decisions (NBS: temporal, choice) was associated with higher evaluations of honesty (CBS: moral, relations).
Extending the specificity of the nomological network
As a second set of theoretical pathways, we propose extending the specificity of the nomological network of leadership behaviors in line with the integrative framework (Steps 4 and 5 in Figure 2). Step 4 focuses on specifying the causes and consequences of leadership behaviors across both behavioral manifestations and evaluative classifications. Scholars can examine how neo-behavioral aspects give rise to evaluations of function and meaning, and how both behaviors and their evaluations can exert distinct causal impacts on leadership outcomes—thereby developing a more complete understanding of the causal network of leader behavior. This approach opens new opportunities for theory building in both the CBS and NBS by explicitly modeling behaviors as action-level antecedents with direct effects, and evaluations as downstream mechanisms that may additionally shape outcomes in their own right. For example, theory can specify how evaluations of concrete behaviors depend on configurations of multiple behaviors and the context in which they unfold (for a review, see Ernst et al., 2022). Because evaluations of function and meaning are likely shaped by repeated behavioral episodes over time, this pathway also invites more dynamic theorizing about how leadership effects accumulate.
Step 5 addresses recursive effects, namely how outcomes of leadership behaviors feed back into subsequent evaluations of function and meaning. This is an important distinction that captures a rationalization of behavior. A CEO or coach who enjoys tremendous success, for example, is likely to have a halo effect that frames the evaluation of any subsequent behavior, and vice versa if the performance of their unit is poor (Nisbett & Wilson, 1977). Hence, leader behavior and the evaluations of such behavior may be influenced by a number of factors other than the behavioral manifestation itself (for a review, see Güntner et al., 2020).
Advancing Research Within the Two Schools of Thought
As a third set of theoretical pathways, we propose ways to advance research within each of the schools of thought, based on our best-practice standards. There are a number of ways to accomplish this within the CBS. Leadership scholars should ensure counterfactuals in item formulations to reduce concerns of causal indeterminacy (Fischer, 2023). Scholars should also assess the extent to which dimensions in existing taxonomies are functionally independent to reduce construct redundancy.
In addition, within the NBS, it is critical that scholars conduct multimodal research to reduce concerns of reductionism (for a review, see Lehmann-Willenbrock, 2024). Such designs combine multiple behavioral aspects (e.g., embodiment, textual, temporal) rather than examining isolated signals. In addition, scholars should examine contextual moderators of neo-behavioral effects to assess the limits of cross-contextual generalizability. Pentland (2012) illustrate the value of this approach. The researchers equipped teams with electronic badges that captured tone of voice, body language, interaction partners, and speaking time. The evidence showed that patterns of communication—integrating multiple neo-behavioral aspects—were the strongest predictors of team success, exceeding the predictive power of cognitive ability, personality, skills, or task content. Since cognitive ability is typically regarded as one of the strongest predictors of performance (Schmidt & Hunter, 1998), this finding underscores the explanatory power of multimodal behavioral patterns. Against conventional assumptions about efficiency, organizational leaders reorganized coffee breaks to increase informal interaction opportunities (Pentland, 2012). Overall, this example highlights how multimodal NBS research can generate theoretically novel and practically actionable insights.
We now illustrate how within-school improvements in the CBS and NBS can be used in a complementary manner. Consider the idea that emotions are contagious and that leaders can influence followers through their emotional displays, regardless of whether these displays fully correspond to felt emotions. Historically, CBS scholars examining emotional contagion have relied on diary or survey designs, which have been criticized for causal indeterminacy and lack of behavioral counterfactuals (Shao, 2024). Conversely, NBS scholars might use artificial intelligence (AI) to capture facial expressions—such as a leader’s smile (Padmaja, Mishra, Mishra, Tembra, Paramasivan, & Rajest, 2024)—but such approaches risk reductionism if they ignore verbal, vocal, or interactional cues, as well as lack of cross-contextually generalizable meaning if interaction history is omitted.
Drawing on the integrative opportunities outlined previously, NBS scholars can adopt a multimodal approach to measuring the signaling of joy. AI can be used to capture the degree of smiling via lip curvature and eyebrow movement (embodiment), large language models can assess expressed joy in speech (textual), and audio analysis can capture vocal tone (embodiment). These signals can be tracked across the duration of an interaction, such as a 30-minute Zoom meeting (temporal), allowing researchers to examine both overall levels and within-episode variation. In addition, applying the same measures to followers enables analysis of covariation between leader and follower affective signals (interaction), thereby situating leader behavior within a broader behavioral episode.
However, even such richly specified multimodal approaches remain limited to behavioral form and manifestation. On their own, they can identify what leaders do and how they do it, but not what these behaviors come to mean to followers or which leadership functions they fulfill. To capture the function and meaning of these behaviors, CBS methodologies remain essential. Conversely, CBS approaches alone cannot specify which concrete actions give rise to particular evaluations and therefore remain weakly actionable and causally indeterminate at the behavioral level. Questionnaire-based measures using counterfactual items and functionally independent leadership categories can assess how followers evaluate a leader’s emotional signaling, for example, as supportive, patronizing, or manipulative. In this way, CBS evaluations provide the interpretive layer that assigns meaning and function to neo-behavioral signals, while NBS measures specify the behavioral building blocks through which these meanings are enacted. Integrating NBS and CBS approaches enables researchers to examine not only how joy is signaled, but also when and why such signaling is evaluated positively or negatively—a question that neither school can adequately address in isolation.
Connecting Functionalist with Interpretivist Leadership Research
In the integrative framework, Step 2 suggests a final theoretical pathway: examining more deeply how the same leadership behaviors can be evaluated differently across contexts. This causal step is compatible with insights from interpretivist leadership research (Alvesson, 2019) and reflects a core premise of interpretivist organizational analysis—namely, that meaning is socially constructed and context-dependent (Morgan, 1980). At the same time, our framework remains grounded in an explanatory and functionalist orientation: contextualized meaning variation is treated as a causal mechanism to be theorized and tested across cases, rather than as the primary object of interpretive inquiry aimed at reconstructing situated meaning. What this implies in practice is that evaluative variability is not treated as noise or irreducible subjectivity, but as a systematic outcome of contextual conditions.
For example, a leader who scores high on relationship-oriented behaviors may be judged as highly supportive by some followers and as patronizing by others. Conversely, a leader who displays few relationship-oriented behaviors may be judged as highly trusting by some followers, while others may judge them as disengaged. Similarly, a follower may assess a behavior as trusting in one context, but the same behavior may be evaluated as disengaged on another occasion (c.f. Johns, 2006, 2023). Importantly, such divergent evaluations are not arbitrary; they may be traceable to differences in situational demands, relational histories, cultural norms, or follower expectations. In this way, evaluative variation becomes an explanandum within a functionalist framework, rather than a reason to abandon causal explanation.
The same logic extends beyond relationship-oriented behaviors to the other dimensions of the CBS. First, a high task orientation may be judged as either guiding or micromanaging, while a low task orientation may be judged as either granting autonomy or laissez-faire. Second, a high change orientation may indicate adaptation or instability, while a low change orientation may indicate stability or inertia. Fourth, a high moral orientation may be judged as moral or moralizing, while a low moral orientation may be judged as value-neutral or technocratic (i.e., leads with evidence-based, ideologically free decisions). Across these dimensions, evaluations vary systematically as a function of contextual cues and interpretive frames rather than randomly across observers. Future research can theorize and empirically test the contextual conditions under which particular evaluations are more likely to emerge, thereby extending the nomological network while remaining firmly within an explanatory logic.
Importantly, although interpretivist approaches were beyond the scope of this meta-synthesis, elements of interpretivist scholarship are highly informative for functionalist research. In particular, interpretivist work sensitizes scholars to the multiplicity of meanings that can be attached to the same behavior and to the social processes through which such meanings arise. Our integrative framework incorporates these insights by explicitly modeling context-dependent evaluation processes (Step 2), while retaining a commitment to causal explanation. At the same time, future research could move beyond this instrumental use of interpretivist insights and examine leadership behaviors from a genuinely interpretivist paradigm, in which meaning construction itself constitutes the primary object of inquiry. Such work could further explore whether interpretivist leadership research comprises distinct schools of thought—ranging from relativist approaches that document divergent interpretations to critical perspectives that interrogate the social, political, and motivational origins of dominant leadership meanings. In this way, functionalist and interpretivist approaches can be understood as complementary but paradigmatically distinct avenues for advancing leadership theory.
Discussion
The current meta-synthesis describes, evaluates, and integrates the research logic underlying two schools of thought of leadership behavior research. Specifically, we distinguish between the classical behavioral school (CBS), which classifies leadership behaviors based on evaluations of their latent and valenced function and meaning, and the neo-behavioral school (NBS), which focuses on the direct observation of behavioral form and manifestation. We show that each school rests on a distinct logic of causal explanation and faces characteristic limitations: CBS research offers cross-contextual generalizability but is vulnerable to conflation, causal indeterminacy, and construct redundancy, whereas NBS research provides behavioral specificity and causal determinacy but risks reductionism and limited generalizability. We further identify school-specific best practice standards that can mitigate—though not fully resolve—these limitations. In the CBS, counterfactual constructs and items can reduce conflation and indeterminacy, while functionally distinct behavior classification schemes can reduce construct redundancy. In the NBS, multimodal research can reduce reductionism, and studying cross-contextual moderators can help assess generalizability. We argue that fully addressing these limitations requires cross-school integration. To this end, we develop an integrative framework that assigns distinct causal roles to concrete behaviors and to evaluations of their function and meaning, thereby enabling causally determinate, theoretically grounded, and context-sensitive leadership research. Building on this framework and the identified best practices, we propose a generative research agenda that outlines multiple pathways for advancing leadership behavior research across schools of thought.
Leadership Inaction as an Additional Research Avenue
This behavior-centric focus across the CBS and NBS leaves underexamined a theoretically important boundary case: leadership inaction. A partial exception is laissez-faire leadership (Judge & Piccolo, 2004), which is typically treated as a CBS construct. Importantly, inaction itself is subject to evaluation and interpretation: followers may construe a leader’s failure to decide as fearful avoidance or, alternatively, as politically savvy restraint. Despite its evaluative relevance, inaction remains underexplored as a distinct behavioral phenomenon in NBS research (Levitis, Lidicker Jr, & Freund, 2009). This gap is increasingly consequential in contemporary contexts—such as executive responses to climate change—where both speaking up and remaining silent send powerful signals. Integrating CBS evaluations with NBS-based analyses of when and how leaders refrain from concrete action would enable more precise theorizing about the causes and consequences of leadership inaction.
Implications for Leadership Development
The integrative framework has direct implications for leadership development because it specifies how people can learn to enact particular leadership functions through concrete behaviors. Neo-behavioral approaches make leadership behaviors observable and trainable: leaders can practice specific actions—such as turn-taking, word choice, gaze, or timing—while these behaviors are recorded and analyzed in real time. Importantly, this allows learners to receive behavioral feedback first, separated from judgment. Feedback research shows that behavior-focused information supports learning more effectively than evaluative feedback, which often triggers defensiveness and impairs self-regulation (Kluger & DeNisi, 1996).
Building on this logic, leaders can then receive feedforward-oriented guidance that links observed behaviors to future improvement rather than past performance (Kluger & Nir, 2010). With recent technological advances, this process can be continuous: leaders practice in meetings or simulations, receive real-time feedback on observable behaviors (e.g., interruptions, speaking duration, linguistic patterns), and subsequently learn how these behaviors are evaluated by others using CBS-based assessments. This two-step process—first observing what leaders do, then examining how these actions are experienced—enables leaders to iteratively adjust their behavior to enact desired leadership styles (e.g., empowering, ethical) while avoiding conflation between observation and evaluation. In this way, leadership development becomes a structured learning cycle grounded in behavioral practice, contextualized evaluation, and actionable improvement.
Implications for Leadership Selection
The integrative framework also has important implications for leadership selection, particularly with respect to reducing bias and improving predictive validity. Selection decisions that rely predominantly on CBS-type evaluations—such as authentic or servant leadership—are vulnerable to halo effects and evaluative conflation (Fischer, Dietz, & Antonakis, 2024). By contrast, incorporating NBS-informed, behaviorally specified criteria anchors selection decisions in observable actions rather than inferred traits or intentions. Recent experimental evidence published in Science supports this logic. For example, Arslan, Chang, Chilazi, Bohnet, and Hauser (2025) demonstrated that behaviorally designed training interventions—focused on specifying and standardizing how evaluators assess candidate behaviors—led to more diverse hiring outcomes. From the perspective of the integrative framework, such interventions work because they constrain subjective evaluation by foregrounding concrete behaviors while still allowing evaluative judgments to play a role. In this way, combining neo-behavioral specification with classical evaluation enables selection systems that are both more equitable and more closely aligned with the behaviors that leaders are expected to enact once hired.
Conclusion
The current meta-synthesis describes, evaluates, and integrates two schools of thought in leadership behavior research with distinct research logics. In doing so, it offers a meta-theoretical reflection that yields an integrative framework capable of overcoming school-specific limitations. This framework also opens up several theoretically grounded pathways for future research. Overall, the review integrates and redirects leadership behavior research toward a more comprehensive, causally rigorous, and actionable understanding of leadership—shifting the field’s focus from diagnosing problems to developing solutions.
