Empirical Followership Research Since the Publication of the Formal Theory of Followership by Uhl-Bien et al. (2014) — A Systematic Review

In their seminal work, Uhl-Bien et al. (2014) postulated that followership studies should privilege the role of the follower in the leadership process so that the study of followership aims to better understand the nature and impact of followers and following in the leadership process. In this way, they clearly distinguished followership approaches from prior approaches (i.e., leader-centric, follower-centric, and relational approaches, see Uhl-Bien et al., 2014), which privileged the role and contribution of the leader instead. That is, those prior approaches focused, for instance, on the nature and impact of leaders and leading in the leadership process (i.e., leader-centric approaches). Moreover, follower-centric approaches drew attention to the followers’ perspective, but still focused on leader and leadership constructions (such as implicit leadership theories or the romance of leadership) instead of follower or followership constructions (such as implicit followership theories or follower role orientations, see Uhl-Bien et al., 2014). Rather than studying the nature of followership or the followers’ contribution to leadership success, those follower-centric approaches explored, for instance, how followers attribute charisma to a leader (e.g., Bligh et al., 2004) or how followers rate their leaders in accordance with their cognitive schema of leader behaviors (e.g., Eden & Leviatan, 1975). Other studies considered leadership as a social exchange process, but positioned the leader as the driving force for the exchange and the relationship building (i.e., traditional relational approaches; see Uhl-Bien et al., 2014). Hence, prior approaches regularly discussed followers, but not necessarily followership in accordance with the FTF (Uhl-Bien et al., 2014). To provide a clear theoretical framework that can enhance and truly advance genuine followership research, Uhl-Bien et al. (2014) proposed the FTF. This integral theoretical framework contains (i) definitions of followership characteristics and behaviors (see above), (ii) two fundamental approaches to followership (i.e., co-production and co-construction), and (iii) a set of variables that were considered pertinent for the study of followership.

According to Uhl-Bien et al (2014), followership necessarily occurs in the context of hierarchical relationships with leaders and is clearly associated with a follower role or with the act of following. Hence, followership characteristics or behaviors clearly differ from general employee characteristics (e.g., demographics; see Oc et al., 2023) or general behaviors at work that do not occur in relation to leaders (such as general proactivity or work engagement; see Uhl-Bien et al., 2014).

The first of the two proposed approaches to followership (i.e., the “reversing the lens” or co-production approach) has been described as follows:

Thus, people in the role of followers are considered to be co-producers of leadership and its outcomes along with their leaders (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). Uhl-Bien and Carsten (2018) later specified that this approach might also involve an informal role perspective, in which leader and follower role-switching can occur in social relations. However, within the role-based “reversing the lens” approach, followership is enacted from a predefined formal or informal follower rank or position (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018).

In contrast, the second approach to followership has been defined as a constructionist approach (see also DeRue & Ashford, 2010) that goes beyond fixed hierarchical role assignments for leaders and followers and explores leadership and followership as a social construction process. This means that the co-construction or leadership process approach to followership investigates how individuals mutually interact and engage in social and relational contexts to construct (or not construct) leadership and followership, while these relational interactions do not necessarily align with formal hierarchical roles (i.e., superiors might not lead and subordinates might not follow; see Uhl-Bien et al., 2014). It provides a framework, in which followership studies explore the dynamic interactions between leading and following patterns, take a close look on what characterizes social constructions of following behaviors and identities, or analyze how and why non-following occurs (see Uhl-Bien et al., 2014). Hence, within the co-construction approach, followership does not mean showing behaviors that are carried out from a follower (i.e., subordinate) role, rank or position, but emerges from combined acts of leading and following, leader and follower identity claims or grants, and from the meaning-making processes of different social actors (see Uhl-Bien & Carsten, 2018).

In addition to these two approaches to followership, Uhl-Bien et al. (2014) provided a set of potential constructs and variables that they considered pertinent for the study of followership. According to the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018), variables in followership research must reflect the unique context and research questions associated with followership. That is, followership occurs in the context of hierarchical relationships with leaders or refers to the act of following within a social interaction process. Furthermore, Uhl-Bien et al. (2014) echoed Shamir’s (2007) call for considering leaders and followers as co-producers of leadership, followership, and its outcomes rather than only “reversing the lens” by studying just the same variables that have been used in leader-centric research. This means that followership research should not just mirror or replicate leadership research from the followers’ perspective, but it should provide research questions and models that allow for a better understanding of the specific nature and impact of followers and following (see Uhl-Bien et al., 2014).

Among the proposed followership characteristics in the FTF are, for instance, followers’ implicit followership theories, follower role orientations, follower identity, or the followers’ political skill (see Uhl-Bien et al., 2014). Proposed followership behaviors are, for instance, obedience, resistance, upward influence, initiative-taking, or voice. The authors also suggest leader characteristics (i.e., for instance, satisfaction with followers or leader identity) and leadership behaviors (i.e., for instance, feedback seeking or follower development) as pertinent variables for followership studies. Furthermore, pertinent outcomes are supposed to be genuine individual follower outcomes (e.g., organizational advancement), individual leader outcomes (e.g., motivation), relationship outcomes (e.g., trust), and leadership process outcomes (e.g., unethical conduct; see Uhl-Bien et al., 2014).

To systematically review empirical followership research since 2014, we selected empirical studies based on the following procedure. We searched for empirical followership studies within the fields of psychology, social science, and business research that were published as peer-reviewed journal articles between 2014 and 2022 in English language, and that referred to work and organizational contexts. In the following, we will first describe our search strategy in more detail. We will then elaborate on the criteria for study inclusion and outline the review procedure.²

For our systematic literature review, we developed a search string and applied it to six databases (i.e., Scopus and APA PsychInfo, APA PsychArticles, SocIndex, Business Source Primier, and Econlit via EBSCO) on February 10th, 2022 (all details of the search string can be found in Supplementary Material 1). The search string contained four basic elements.

With the first element of the string, we intended to find every article that had the words “follower” or “followership” in the title. With the second element of the string, we intended to additionally perform a more nuanced search to find studies that investigate the followership constructs Uhl-Bien et al. (2014) had suggested. Therefore, in addition to the title search for “follower” and “followership”, we also searched for “follower”, “member”, or “employee” in combination with followership characteristics or behaviors such as “resistance”, “dissent”, or “influence tactics” (see Uhl-Bien et al., 2014) in titles, abstracts and keywords.³ With the third element of the string, we intended to additionally find followership studies that might use new or additional followership constructs, which were not mentioned by Uhl-Bien et al. (2014). Hence, we added a few more general terms to the title, abstract, and keyword search, for instance, “characteristic”, “skill”, or “identity” in combination with “follower”. In the fourth element of the string, we added some limitation criteria (i.e., for instance, time range [2014-2022], scientific disciplines, and English as the publication language). With the “AND NOT” operator, we excluded hits outside our region of interest (i.e., for instance, “animal”, “child”, or “social media”). Furthermore, given our strict focus on empirical followership studies, we excluded notes, editorials, reviews, conference papers, and books. With a separate search, however, we tried to ensure that there was no other systematic review that conflicts with our research interest.

Our search resulted in 6,330 hits total (see Figure 1). In addition, we conducted a citation forward search of Uhl-Bien et al. (2014), which resulted in additional 356 hits. After removing duplicates, we obtained 4,716 articles for abstract screening. The screening of the abstracts resulted in 182 articles that we selected for full-text screening, because either the study was considered eligible for our review or we could not yet make a decision based on the abstract. The first author performed the full literature search and study selection. Following common practice (e.g., Fischer et al., 2021; Forner et al., 2023; Gullifor et al., 2023), another author additionally performed one half of the full-text screening for inter-rater reliability with substantial agreement (84%; k = .67, p < .001; Cohen, 1960; Landis & Koch, 1977). All discrepancies between the two authors were discussed and resolved, which finally led to 100 percent agreement. A total of 89 studies were included in our review.

To identify followership studies in our systematic literature search, we referred to the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). Hence, we included followership research that either (i) “reversed the lens” (see Shamir, 2007; Uhl-Bien et al., 2014) or (ii) studied followership from a leadership process/co-construction perspective (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). In contrast, we excluded leader-centric, follower-centric, or relational approaches that neglected the role and contribution of the followers in leadership. Based on the theoretical principles and definitions of the FTF (see above; Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018), we specified clear inclusion and exclusion criteria for the study selection of our systematic review (see Table 1).

To answer our research questions, we analyzed the 89 included studies and extracted and synthesized the following information: (i) descriptive information (i.e., for instance, types of samples and sample size, or countries where the research was conducted), (ii) the research question of the study, (iii) the followership constructs investigated, (iv) the applied measures to operationalize the followership constructs, (v) the investigated non-followership variables, (vi) the applied methods, and (vii) the main findings. We also categorized each study according to the following approaches: (a) “reversing the lens”/co-production” approach to followership (see Shamir, 2007; Uhl-Bien et al., 2014), (b) “leadership process/co-construction” approach to followership (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018), and (c) scale development/validation study.

Furthermore, we assessed the quality of each study to review the state of followership research and which evidence the selected studies were able to provide. To do so, we largely followed the Study Design and Implementation Assessment Device (Valentine & Cooper, 2008) to assess the quality of the included studies. Specifically, we evaluated the (i) study design, (ii) operationalization and measurement, and (iii) statistical approaches. We categorized each study as “adequate”, “fair”, or “questionable” in these three domains. For (i) study design, we assessed whether the research design was appropriate to address the aims of the research. Specifically, we determined if sample sizes were reported, justified, and whether the sample sizes were adequate to provide sufficiently precise estimates of effect sizes (see Valentine & Cooper, 2008; for details, see also Supplementary Material 2). In addition, we identified whether the outcome was measured at an appropriate time for capturing the proposed effect and to what extent directions of effects could be identified for important measured outcomes (see Valentine & Cooper, 2008). For (ii) operationalization and measurement, we evaluated to what extent variables were assessed in a way that is consistent with the definitions of the study and its proposed effects. That is, measures should represent the content of interest (i.e., they should at least have face validity), and measures should be sufficiently reliable to allow adequately precise estimates of the effect sizes (see Valentine & Cooper, 2008). For (iii) statistical approaches, we determined whether the statistical tests were adequately reported and whether effect sizes and their standard errors were accurately reported (see Valentine & Cooper, 2008). Similarly, the qualitative data analysis in qualitative studies should also be reported precisely and plausibly. Further details concerning the quality assessment procedure can be found in Supplementary Material 2.

Although followership research is still only a small part of leadership research, Figure 2 shows that it can indeed be described as an emerging field. The majority of studies were conducted in North America or Europe (n = 52), followed by Asia (n = 37). Sample sizes ranged from 56 to 212,223 participants for quantitative studies. The median was 297. If we treat the extraordinarily large sample of 212,223 participants as an outlier, the other studies realized an average sample size of 342. For qualitative studies, sample sizes ranged from 4 to 92 participants with an average sample size of 39. The median was 29. The majority of studies were single-study reports (n = 65), while 24 studies were multi-study reports. The 89 studies comprised 128 different samples. A comprehensive overview of all included studies that contains information on the chosen followership approach, the unit of analysis, the methodological approach, the sample, and a brief summary of results can be found in Supplementary Material 3.

Table 2 shows the distribution of the studies with regard to their followership approaches and their methodological approaches. In a first step, we sought to answer our first two research questions: Which approaches to followership have been applied in empirical studies since 2014? And is there possibly a lack of balance between the two fundamental approaches? To do so, two authors independently assigned the selected studies to one of the followership approaches as part of the coding process to obtain information on inter-rater reliability (see the search strategy as described above). With only one discrepancy, which was later resolved by discussion, substantial agreement was achieved (97%; k = .65, p < .001; Cohen, 1960; Landis & Koch, 1977).

The vast majority of studies (n = 78) that we identified for this review used a “reversing the lens”/co-production approach to followership (Shamir, 2007; Uhl-Bien et al., 2014). Those studies were role-based, because followership referred to a static formal follower role or position (see Uhl-Bien & Carsten, 2018). One study (Falls & Allen, 2020), however, investigated the dynamics and switching between leader and follower roles within individuals and, hence, explored followership from an informal role perspective (see Uhl-Bien & Carsten, 2018). We could identify only three studies that used a leadership process/co-construction approach to followership (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018), which indicates an asymmetry between the two fundamental approaches within the FTF framework. Furthermore, 14 out of 89 studies were scale development or psychometric validation studies.

In a second step, we sought to answer the following research questions: Which methods have been applied to investigate followership in empirical studies since 2014? And were different methods used for the “reversing the lens”/co-production approach than for the leadership process/co-construction approach?

Table 2 shows that most of the reviewed studies used a quantitative methodological approach. Within the “reversing the lens”/co-production followership approach (Shamir, 2007; Uhl-Bien et al., 2014), most studies used a cross-sectional quantitative approach (n = 43), 27 studies used a time-lagged or longitudinal approach, and 12 studies used an experimental design. Qualitative methods were applied in nine role-based followership studies. All three leadership process/co-construction (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018) studies used exclusively qualitative methods. For scale development and validation, a cross-sectional quantitative approach was primarily used (n = 14), while four studies also applied qualitative methods for scale development.

The assessment of study quality revealed that the study designs of 30 out of 89 studies were questionable (see Table 3). One of the main reasons for this assessment was that those studies used a cross-sectional quantitative design to test cause-effect relationships (e.g., consequences of followership behaviors). Another frequent problem of certain study designs was a sample that lacked sufficient statistical power. Furthermore, for 12 studies, the operationalization or measurement of the main variables were questionable. For instance, some studies (e.g., Essa & Alattari, 2019; Ivanoska et al., 2019) used Kelley’s (1992) questionnaire in its original form to assess his two proposed dimensions of followership behavior (i.e., active engagement and critical thinking toward the leader). However, different validation studies (see, for instance, Blanchard et al., 2009; Ribbat et al., 2021) showed that the underlying factor structure of this questionnaire differs from what Kelley (1992) had expected. Moreover, several studies reported either questionable reliability of certain scales or no reliability information at all. Furthermore, the statistical approaches were questionable for six studies due to incomplete reporting of the procedure or the results (i.e., for instance, no degrees of freedom were reported for structural equation modeling).⁴

We assessed 40 out of 89 studies to be at least “fair” in all categories (i.e., study design, operationalization and measurement, and statistical approaches), which suggests that their findings should be rather robust (see Valentine & Cooper, 2008). Those studies were mainly designed as qualitative, quantitative longitudinal/time-lagged, or experimental studies. Furthermore, some cross-sectional study designs were considered fair as they were validation studies or were only interested in descriptive information, such as how some followership styles were distributed in certain samples. In addition to those 40 studies, we found another 12 multi-study reports to be partially fair designed. That is, one or more studies of the multi-study report were adequately or fairly designed, while other studies were not. Similarly, we found one study (Kim & Schachter, 2015) to be partially fair operationalized.

With regard to investigated followership constructs (i.e., followership characteristics and behaviors), we sought to answer the following research questions: Which followership constructs have been investigated in empirical studies since 2014? Which followership constructs have been neglected? And are there additional followership constructs that Uhl-Bien et al. (2014) did not suggest in their review? Table 4 shows the followership constructs that were investigated in the reviewed studies. We found several followership characteristics and behaviors in our systematic review that Uhl-Bien et al. (2014) had suggested in their theoretical framework, variables that brought new perspectives on those proposed constructs, and constructs that went beyond Uhl-Bien’s et al. (2014) suggestions. Following our quality assessment as described above (see also Table 3), in this section, we only refer to those studies with at least fair operationalizations and measurements of the followership constructs.

First, several studies investigated followership characteristics and behaviors as conceived by Uhl-Bien et al. (2014; see Table 4). (i) Implicit followership theories (e.g., Junker et al., 2016), (ii) Kelley’s (1992) followership behaviors and styles (e.g., Gatti et al., 2014), (iii) upward influence tactics and impression management (e.g., De Clercq et al., 2021), (iv) voice behaviors (e.g., Carsten et al., 2018), and (v) feedback seeking (e.g., Gong et al., 2020) were the most studied constructs and, thus, can be highlighted as the main focus of empirical followership research since 2014. In contrast, advising was the only construct that was conceptualized as followership behavior by Uhl-Bien et al. (2014) and not investigated in the reviewed studies at all. Furthermore, political skill, Machiavellianism, motivation to lead, and romance of leadership were considered potential followership characteristics by Uhl-Bien et al. (2014), but were not under investigation in the reviewed studies.

Second, some studies brought up new perspectives to those research lines that were suggested by Uhl-Bien et al. (2014), that is, while the constructs were part of Uhl-Bien et al.’s (2014) model, they were investigated in new ways. For instance, some authors introduced group-level variables to the framework of followership constructs. Leung and Sy (2018), for instance, studied group-level implicit followership theories that represent shared conceptions of followers at the group level. Babalola et al. (2021) and Sessions et al. (2020) focused on group-level voice behaviors, which represent combined contributions or shared suggestions and concerns that are put forward to the leader in order to challenge the status quo.

In addition to those constructs that were suggested by Uhl-Bien et al. (2014), we identified 23 more followership constructs (i.e., six followership characteristics and 17 followership behaviors) that were investigated in the reviewed studies (see Table 4). With regard to followership characteristics, Ellis et al. (2021), for instance, studied an equivalent to implicit followership theories with implicit voice theories (i.e., “socially acquired beliefs, or implicit theories, about what makes voice risky in social hierarchies”; Detert & Edmondson, 2011, p. 462). Metwally et al. (2018) investigated the followers’ power profiles as the sum of power sources that let the leader comply to the followers’ influence attempts. Granger et al. (2020) introduced leader-related political knowledge as a followership characteristic, which refers to the follower’s perceived understanding of the relationships, demands, resources, and preferences of the leader as the target of influence. A few studies (Wen et al., 2021; Zheng et al., 2019; Zhong et al., 2021) went in a similar direction by investigating subordinate moqi as a followership characteristic that is considered to be more specific to Asian culture. It refers to the followers’ proactive understandings about leaders’ unspoken requirements, expectations, intentions, and desires based on non-verbal cues (such as body gestures, facial expressions, and voice tone; Zheng et al., 2019).

Subordinate-supervisor guanxi is another concept from the Asian background, which is related to leader-member-exchange. It involves close personal links that emerge from informal connections and relationship building outside the workplace (H. Ren & Chen, 2018). Ren and Chen (2018) studied the individual perception of group-level guanxi practices along with subordinate’s guanxi-building behaviors. The individual perception of group-level guanxi refers to the perception of whether guanxi typically influences their supervisor’s decisions (H. Ren & Chen, 2018). Hence, it can be considered a followership characteristic. Furthermore, Vriend et al. (2020) investigated follower motivation in the form of reciprocity motives, defined as the intention to reciprocate former experiences with the leader either in favor of the leader or in favor of oneself.

In addition to those followership characteristics, we could also detect several forms of followership behavior that extend the list of potential followership constructs and variables as conceived by Uhl-Bien et al. (2014; see Table 4). That is, for instance, some followership behaviors that we identified aim to mimic prominent leadership behaviors, such as abusive followership (Camps et al., 2020). Furthermore, we found several forms of behaviors that provide a more nuanced understanding of proactive followership, such as followers’ helping behaviors related to their leaders (Hoption, 2016), courageous communication toward the leader (Baker et al., 2016), and relationship- or guanxi-building behaviors that were carried out by the followers (Baker et al., 2016; Lapalme et al., 2017; H. Ren & Chen, 2018). Three studies (Baker et al., 2016; Q. Huang & Zhang, 2021; Liu & Dong, 2020) investigated the follower’s taking of the leader’s perspective as followership behavior. Tessema and Folovorito (2021) distinguished routine and strategic followership behaviors. Routine behaviors referred to regular, repetitive actions without taking into account benefits to one’s future career or position, whereas strategic behaviors focused on the follower’s action grounded in one’s future career or position aspiration.

Apart from followers’ facilitative proactivity, some studies focused on avoidant behaviors. That is, some studies applied previously suggested followership behaviors but inspected feedback avoidance (Arain et al., 2020) instead of feedback seeking, or interaction avoidance (Q. Huang & Zhang, 2021) instead of proactive behavior. With reference to emotional labor research in the context of service work, Xu et al. (2014) and Yang et al. (2021) adapted the concepts of emotional masking, surface and deep acting to the followership context. This means that followership behaviors were investigated in form of followers’ attempts to disguise their true inner feelings and the modification of their displayed affect in front of their leaders (J. Xu et al., 2014; J. Yang et al., 2021).

De Jong et al. (2021) brought up another perspective on followership behaviors as they explored the consequences of subordinates’ psychological contract breach (i.e., the failure to meet expectations about obligations and benefits within the employment relationship). In contrast to previous leader-centric research, de Jong et al. (2021) focused on the consequences of the subordinates’ psychological contract breach vis-à-vis the leader and, thus, presented another construct within the framework of followership behavior. Similarly, Gloor (2021) investigated social norm violations of followers and studied the followers’ sarcastic interactions with their leader as a form of followership behavior. Finally, the three leadership process/co-construction studies (Blom & Alvesson, 2014; Larsson & Nielsen, 2021; Van De Mieroop, 2020) explored individuals’ claiming and granting of the follower role or identity in different social interactions.

In addition to the identification of followership characteristics and behaviors that were investigated in the reviewed studies, we also sought to answer the following research questions: Which measures were used for the different followership constructs? And which new ways to measure followership have been developed, applied, or adapted in empirical studies since 2014? We provide an overview of such measures in Table 5. Following our quality assessment of the reviewed studies (see also the previous chapter and Table 3), we additionally present information on whether the specific measure refers to a previously established measure, whether a validation of a measure was provided within the reviewed study, whether the instrument displays at least face validity, or whether its validity is questionable according to our assessment (see Table 5).

In sum, 52 different measures were used in the reviewed studies to investigate various forms of followership characteristics and behaviors. Among these 52 measures, previously established instruments (n = 31) were applied to measure followership characteristics and behaviors.⁵ In addition, 13 new measures were developed since 2014. Four of these 13 newly developed measures were also applied in other studies that we reviewed (i.e., beyond the study introducing them). Furthermore, nine newly developed measures were validated within the followership studies that we reviewed. The other four measures displayed at least face validity, even if they were not validated within the reviewed studies.

Twelve existing measures were adapted to comply with the followership context (i.e., with the followership role) and, hence, could be considered followership characteristics or behaviors. That is, for instance, Yang et al. (2021) used the instrument to measure emotional labor by Grandey (2003) and modified the reference person of the items by replacing “customers” with “supervisor” to measure the extent to which subordinates engaged in deep and surface acting directed at their leader. Similarly, existing measures were adapted for the follower’s power profile (Metwally et al., 2018), supervisor-subordinate guanxi building behaviors (H. Ren & Chen, 2018), perspective taking vis-à-vis the leader (Liu & Dong, 2020), subordinate psychological contract breach (de Jong et al., 2021), emotional masking (J. Xu et al., 2014), and abusive followership (Camps et al., 2020). All 12 adapted measures display face validity according to our assessment (see Table 5).

Apart from identifying followership constructs (i.e., followership characteristics and behaviors) and its measures, we also sought to answer the following research question: Which variables other than followership characteristics and behaviors have been investigated in followership studies since 2014? By analyzing these variables and the central findings of the studied research models, we also address the following research question: How has the field of empirical followership research developed since 2014? Table 6 shows the non-followership specific variables that were explored within the reviewed studies with their function in the different followership models (i.e., as another independent variable, moderator variable, mediator variable, or dependent variable). Furthermore, Table 6 indicates, which of these variables were investigated in studies that were assessed to have at least fair quality.

In what follows, we proceed largely in correspondence with the scheme of theoretical constructs and variables for the study of followership as suggested by Uhl-Bien et al. (2014). However, we present the actually studied variables since 2014 and extend the scheme by using a somewhat more differentiated approach: In our presentation, we clearly distinguish followership characteristics and behaviors from other variables in the authors’ nomological network. While followership characteristics and behaviors (which were analyzed in the previous chapter), by definition, must refer to the followership role, the variables presented in the current part of the review represent all other, non-followership specific variables that were part of the studied models (i.e., as another independent variable, moderator variable, mediator variable, or dependent variable).

The first column of Table 6 clusters those studied variables into six different categories. These categories are (i) follower characteristics, (ii) follower behaviors, (iii) leader characteristics, (iv) leader behaviors, (v) followership outcomes, and (vi) contextual/situational variables. Follower characteristics and behaviors (i.e., follower-related variables) here represent all characteristics and behaviors of followers that do not refer to the follower role (i.e., for instance, general characteristics of the followers such as age or self-esteem, or general behaviors of the followers such as work effort or help-seeking behavior that was not directed at the leader) and, hence, could not be considered followership characteristics or behaviors. Leader characteristics and behaviors are all leader-related variables that were studied in the reviewed studies. Followership outcomes refer to individual follower outcomes, individual leader outcomes, relationship and leadership process outcomes of followership characteristics or behaviors as conceived by Uhl-Bien et al. (2014). All other variables were categorized as contextual/situational variables. The second column of Table 6 clusters the studied variables by their function in the followership models (i.e., the variables were either used as an additional [independent] variable, as a mediator variable, as a moderator variable, or as a dependent variable of followership characteristics and/or behaviors). Finally, the third column of Table 6 provides the different variables along with the corresponding references.

Several variables that Uhl-Bien et al. (2014) suggested for the study of followership, however, were not considered in the reviewed studies. These variables were (i) the leaders’ satisfaction with followers, (ii) democratic or autocratic decision making, (iii) feedback seeking, and (iv) consultation with followers. Moreover, several leadership process outcomes that Uhl-Bien et al. (2014) proposed were not directly investigated in the reviewed studies (i.e., (i) goal accomplishment, (ii) mission fulfillment, and (iii) advancing change/maintaining the status quo.

To provide an overview of the evidence that the reviewed followership studies were able to provide, we summarize their central findings in this section. We only refer, however, to those studies that had at least a fair quality according to our assessment (see Table 3). As described above, we found several studies with questionable designs, operationalizations, or reporting of statistical procedures or results. Hence, those studies could not ensure sufficient robustness of their findings (see Valentine & Cooper, 2008) and are thus left out of our in-depth overview.

First, a wide variety of studies provided evidence for the presence of different followership characteristics, styles, and behaviors. That is, for instance, the findings of numerous studies (e.g., Braun et al., 2017; Coyle & Foti, 2022; Junker et al., 2016; Knoll et al., 2017; Veestraeten et al., 2021; A. Yang et al., 2022; Y. Yang et al., 2020) support the assumption that followers have distinct cognitive schemas about ideal or prototypical and counter-ideal or anti-prototypical followers (i.e., different implicit followership theories), about different follower identities (e.g., Peters & Haslam, 2018; Schneider et al., 2014), and different follower role orientations (e.g., Carsten et al., 2018, 2021). Numerous studies support Kelley’s (1992) assumption that active engagement and independent, critical thinking toward the leader are two distinctive followership behaviors (e.g., Bell, 2020; Gatti et al., 2014; Ghislieri et al., 2015; Ribbat et al., 2021). Furthermore, several qualitative studies uncovered distinct active and passive followership behaviors (such as contributing to the overall vision or simply executing commands; e.g., Benson et al., 2016; Gesang & Süß, 2021; Gordon et al., 2015; Kim & Schachter, 2015), or routine and strategic followership behaviors (Tessema & Florovito, 2021). St-Hilaire et al. (2019) inductively developed a taxonomy of followership behaviors that promote the leader’s health. Almeida et al. (2021) inductively identified different types of followers that were confronted with a destructive leader (i.e., resisters, obedient, and mixed-behavior followers). Moreover, the findings of Falls and Allen (2020) suggest that managers need to be flexible to switch between leader and follower roles in order to be effective.

Second, proactive and constructive followership identities and behaviors were largely found to be positively related to what are usually considered desirable individual, interpersonal, or organizational variables, and mainly negatively related to what are usually considered undesired variables. For instance, three studies found that implicit theories, role orientations, or identities that associate the ideal follower with proactivity or a co-production belief were positively related to effort, performance, voice, or positive leader emotions, whereas associating a follower with passivity and/or disobedience rather had the opposite effects (Carsten et al., 2018, 2021; Schneider et al., 2014). Similarly, Kelley’s (1992) active and independent followership behaviors were positively associated with organizational citizenship behaviors (Gatti et al., 2014; Ribbat et al., 2021) and the follower’s emergence as a leader (Jiang et al., 2021). Furthermore, they were negatively associated with the followers’ disengagement (Gatti et al., 2014). In addition, the findings of Gesang and Süß (2021) and Benson et al. (2016) also suggest that proactive rather than passive followership behaviors were preferred by the leaders.

However, inconsistent results were found for the two dimensions of Kelley’s (1992) followership behavior (i.e., active engagement and independent, critical thinking toward the leader) with regard to job-related attitudes (Gatti et al., 2014, 2017; Ribbat et al., 2021), emotional exhaustion (Gatti et al., 2014; Ribbat et al., 2021), and leader-member exchange (LMX; Gatti et al., 2014; Ribbat et al., 2021). That is, for instance, active engagement was found to be positively related to job satisfaction (Ribbat et al., 2021), while critical thinking was found not to be related to job satisfaction at all (e.g., Gatti et al., 2014). Furthermore, according to the findings of Veerstraeten et al. (2021), followers might reduce their engagement at work, when they have the general belief that followers should be hardworking and productive, but feel that their leader does not convey high expectations. In addition, the findings of Knoll et al. (2017) suggest that implicit followership theories could either increase (i.e., the schema of being a “good citizen”) or decrease (i.e., the schema of “insubordination”) the followers’ tendencies to contribute to unethical leadership. These findings also point to the specific situational context as an important boundary condition, under which implicit followership theories unfold their effects (Knoll et al., 2017).

Furthermore, the findings of eight studies suggest that voice behavior directed at the leader can have a positive impact for both the individual and the organization. For instance, voice was found to be related to greater follower well-being (Yousaf et al., 2019), leader receptivity and recognition (Howell et al., 2015; Z. Zhang et al., 2020), better performance evaluation (Howell et al., 2015), follower endorsement by the leader (R. Ren et al., 2022), perceived follower support, leader motivation, contribution to goal attainment (Carsten et al., 2018), LMX (A. J. Xu et al., 2021), as well as lower abusive supervision (Babalola et al., 2021) and lower emotional exhaustion for the leader (Sessions et al., 2020). In addition, the findings of three studies suggest that the follower’s feedback-seeking from the leader represents a form of proactive, constructive followership behavior, since it was found to be positively associated with the follower’s career adaptability (Gong et al., 2020) and subsequent perceptions of moqi with the leader (Zheng et al., 2019). Furthermore, Mao (2022) found that directly asking the leader for feedback was negatively related to abusive supervision. However, trying to get feedback by simply watching the leader’s reactions was found to be positively related to abusive supervision (Mao, 2022).

Additionally, two studies found upward influence tactics and impression management to be predictors for personal success or favorable leader reactions, such as social influence over others (De Clercq et al., 2021), or positive idea assessments by the leader (Lu et al., 2019). However, the findings by Klotz et al. (2018) suggest that the use of impression management can have negative implications for the followers themselves (i.e., in form of resource depletion) and it might lead to harmful behavior from an organizational perspective. Additionally, follower moqi was found to be a helpful resource for desired followership outcomes. The findings of three studies (Wen et al., 2021; Zheng et al., 2019; Zhong et al., 2021) suggest that follower moqi can be useful for followers to be effective within the leadership process and the organization (i.e., for instance, with regard to social influence exerted over the leader or getting rewards from the leader; see Zheng et al., 2019).

Third, inconsistent findings were reported for masking, opposing, and destructive followership behaviors. That is, the findings of Yang et al. (2021) suggest that emotional masking might have a negative impact on LMX quality, while followers’ attempts to actually change their underlying affective experience (i.e., deep acting) might rather benefit LMX quality (J. Yang et al., 2021). With regard to opposing followership behaviors, Garner (2016) reports a general openness from the interviewed leaders to constructive dissent in his qualitative study. Resisting behaviors, however, might appear either constructive (Gloor, 2021) or destructive (Güntner et al., 2021) for the leader and the organization. Two studies linked destructive followership behaviors with negative consequences for the leader, such as emotional exhaustion (de Jong et al., 2021), perceived interpersonal injustice by the leader, and with abusive supervision (Camps et al., 2020). The findings of Gloor (2021), however, suggest that a social norm violation in form of sarcastic interaction with the leader can reduce the leader’s self-interested behavior (i.e., leader self-overpay). Inconsistent results were also found for the effects of reciprocity motives of followers on the intention for unethical conduct. For instance, Vriend et al. (2020) found that the relationship between a positive reciprocity motive and intended pro-self unethical behavior was negative in one of their studies and positive in their other study.

Finally, the studies that applied the constructionist approach to followership provided insights into follower identity claims and grants within different social interactions. Blom and Alvesson (2014) explored who influenced, inhibited, and initiated managerial leadership among engineers in two organizations. Results indicate that subordinates rather than their manager defined the leadership situation, although subordinates temporarily and partially accepted a followership identity (Blom & Alvesson, 2014). Moreover, Larsson and Nielsen (2021) examined how people collaboratively construct identities in eight organizations. Their conversation analysis of different team and department meetings revealed that leader and follower roles remained abstract in workplace interactions and that participants rather focused on negotiated, task-oriented, and practical identities (such as expert or non-expert identities, see Larsson & Nielsen, 2021). Furthermore, they worked out risks and challenges of claiming a follower identity (i.e., for instance, the challenge to identify a leader identity at play and creating an appropriate follower identity, see Larsson & Nielsen, 2021). Additionally, Van De Mieroop (2020) also analyzed the construction of leader and follower identities by using a discourse-analytical approach. She worked out how participants of various meetings of healthcare workers either actively co-constructed the superior’s leader identity or projected a leader identity upon the superior by actively enacting their identities as followers (Van De Mieroop, 2020).

Our systematic review of empirical followership research since 2014 revealed that an increasing number of studies conceptualizes followers as relevant contributors, co-producers, or co-constructors of leadership and its outcomes. While not all studies that we included in our review referred to the followership framework explicitly, a wide variety of constructs were investigated in line with the FTF by Uhl-Bien et al. (2014). Based on the inclusion and exclusion criteria that we developed in line with the FTF, we could examine the different followership approaches, the studied followership characteristics and behaviors, several measures for their assessment, its nomological network, and the different methods used. By analyzing the reviewed studies, however, we also uncovered some shortcomings, research gaps, and promising avenues for future research that we discuss in the following. Thereby, we refer to our following two research questions: Which new impulses for followership research arise from the empirical studies since 2014? And what has been neglected so far?

Based on our key findings and current shortcomings, which we have highlighted in this section, we have summarized potential areas for future research in Table 7. We also identified possible research topics and open research questions for each of these areas.

In our analysis of empirical followership research since Uhl-Bien et al. (2014) presented their FTF, we found that most followership studies “reversed the lens” (Shamir, 2007) or studied the followers’ leadership co-production. Only three out of 89 studies applied a leadership process/co-construction approach. These findings correspond to the concerns that Uhl-Bien et al. (2014) themselves expressed when they introduced their theory: They assumed that leadership researchers would prefer the role-based approach over the constructionist approach, as it appears to be easier to study (Uhl-Bien et al., 2014).

Neglecting the leadership process/co-construction approach is a relevant shortcoming of followership research, since the constructionist perspective is an integral part of the FTF (see Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). The constructionist approach allows for a deeper understanding of the interplay of leading, following, and non-following and, thereby, of the co-construction of leadership and followership as a dynamic social interaction process between social actors regardless of their formal position (i.e., supervisor or subordinate). Our analysis, however, shows that empirical followership research—apart from a few exceptions—has missed the opportunity to move forward to a new and better understanding of the social processes that co-construct leadership and followership, as most studies stuck to the predetermined labels of leaders and followers.

Moreover, several studies from the context of leadership identity construction processes could not be included in our review in the first place because they were leader-centric (e.g., Ali et al., 2020; Cook et al., 2021). For instance, an increasing number of studies focus on leader emergence or shared leadership as consequences of leadership identity construction (see, for instance, Wu et al., 2020; Zhu et al., 2018), without asking, however, why, when, or how people claim or grant a follower identity (i.e., the construction of followership; see also Uhl-Bien & Carsten, 2018). Thus, while we found a trend toward more pertinent research activity, the application of the two fundamental approaches to followership is unbalanced. Neglecting the constructionist approach to followership reveals a gap that should be filled in future studies (see Table 7).

Our systematic review reveals a large number of studied constructs and variables that fit into the integrative framework of the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). However, our findings also reveal that empirical followership research still lacks depth in several ways. Firstly, a few constructs were studied (much) more often than others, which indicates an imbalance of constructs under investigation. In fact, most of the followership constructs that we identified in our review appeared in only one or two studies, thus preventing cumulative knowledge (see Table 4). It is not surprising, however, that Kelley’s (1992) followership behaviors or Sy’s (2010) implicit followership theories were among the most studied variables, as they belong to those prominent (and early) theoretical approaches that are a core part of the followership concept. Still, we would have expected more progress for empirical tests of further prominent theoretical approaches that were specifically developed to conceptualize followership. For instance, Chaleff’s (1995) courageous followership or the partnership model by Hurwitz and Hurwitz (2015) were not investigated at all in the reviewed studies and, hence, are still lacking empirical exploration.

Secondly, even the most studied followership constructs (i.e., Kelley’s [1992] followership behaviors and styles, implicit followership theories, upward influence tactics and impression management, voice behaviors and feedback seeking) were investigated in a rather fundamental way: This research included validation studies of new or existing measures (e.g., Gatti et al., 2014; Petruş, 2018), related studies often focused on a limited number of variables, and some studies were even limited to the description of (preferred) followership styles (e.g., Essa & Alattari, 2019; Ivanoska et al., 2019). While some research lines do have a longer tradition that exceeds the time period that we reviewed (for an overview of upward influence tactics research, for instance, see Lee et al., 2017), we, thus, observe that most followership research lines have just begun empirical exploration.

Third, we found that several variables that Uhl-Bien et al. (2014) suggested for the study of followership were not investigated in the reviewed studies at all (e.g., the followers advising or various followership outcomes such as goal accomplishment and mission fulfillment). Hence, there are several opportunities for future research that our analysis could identify.

Thus, going forward, one promising avenue would be to explore the variety of constructs presented in our review in more depth. It might be particularly promising to further test the various prominent theoretical approaches that are a core part of the followership concept (i.e., for instance, Carsten et al., 2010; Chaleff, 1995; Hurwitz & Hurwitz, 2015; Kelley, 1992; Sy, 2010) to make progress both theoretically and empirically (Table 7).

Additionally, it is important to provide comparative tests of different followership approaches and concepts. This can help to establish theoretical parsimony by avoiding construct redundancy. Hence, future studies may focus on comparing different followership approaches in terms of their utility and incremental validity to predict organizational outcomes (cf. Hoch et al., 2018; Montano et al., 2023; Tonidandel & LeBreton, 2011). According to our analysis of followership studies between 2014 and 2022, proactive and constructive followership characteristics and behaviors seem to be positively related to what are usually considered desirable individual, interpersonal, or organizational outcomes, and mainly negatively related to what are usually considered undesired outcomes. Future research could compare which of these followership concepts is more effective and for instance results in better performance outcomes or greater satisfaction. Such comparative research can also be meaningful to further establish and distinguish followership from leadership.

In addition to followership characteristics and behaviors, we also analyzed the variables that extend or contribute to the nomological network of followership constructs. Since Uhl-Bien et al. (2014) echoed Shamir’s (2007) call for considering leaders and followers as co-producers of leadership and its outcomes rather than “reversing the lens” by studying just the same variables that have been used in leader-centric research, we asked as one of our research questions: In how far did the studied variables truly reflect the unique context and research questions associated with followership? Our findings show that there is in fact a large overlap between the variables that have been investigated in the context of followership characteristics and behaviors and variables that have been traditionally studied in leader-centric research (e.g., wellbeing, organizational commitment, performance, LMX). However, we found several leader-related variables (such as perceived follower support) and followership outcome variables (such as LMX) that correspond to the proposed scheme by Uhl-Bien et al. (2014). In addition, several followership outcomes, which were not proposed by Uhl-Bien et al. (2014), can still help followers to learn about how to be effective in creating a desired work environment and, thus, should also be considered suitable for the study of followership. For instance, follower well-being is an important goal for successful leadership in organizations, not least since it is also associated with better performance (see, for instance, Inceoglu et al., 2018; Montano et al., 2023). Furthermore, our findings show that several studies included leader characteristics and behaviors as additional independent variables (e.g., the leader’s implicit followership theories), as mediator variables (e.g., perceived follower support), or as moderator variables (e.g., leader’s emotion control) in their followership models. Hence, those followership studies explored the contribution and impact of followers (i.e., they were followership studies by definition), while considering both followers and leaders as co-producers of leadership (i.e., these studies answered the call by Uhl-Bien et al. [2014] and Shamir [2007]).

Some variables that we found in the reviewed studies, however, do not appear unique to the context of followership and, therefore, rather mirror traditional leadership research questions from a different perspective. For instance, general behaviors of followers that were not enacted from the standpoint of a follower role were also studied as dependent variables in the reviewed models (such as customer orientations, employee proactivity, or work engagement). Furthermore, the attempts to explain leadership styles (e.g., transformational leadership behavior) from a followership perspective are indeed at risk of just mirroring leader-centric research from the follower’s perspective. However, it is a plausible assumption that leaders alter their behavior in the wake of their experiences with certain followers (see, for instance, Güntner et al., 2021; L. Li et al., 2020).

Hence, apart from a few exceptions, a clear exclusion of variables, which were also studied in traditional leadership research, based on conceptual or theoretical grounds is, in fact, hardly possible. That is, leadership and followership are closely related and, therefore, a certain overlap of the variables of interest is natural to some extent. In addition, we found the derivation of theoretical hypotheses in the reviewed studies largely to be plausible. Since the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018) provides a rather integrative framework for the study of the nature and impact of followers and following in leadership, followership studies usually have to rely on additional theoretical foundations in order to derive well-grounded and plausible hypotheses. Still, if future studies set out to advance knowledge on followership, they should carefully focus on outcome variables that relate closely to the follower role or the act of following. For instance, future research could explicitly explore the satisfaction with followership behavior as a performance measure that fits the followership framework instead of assessing employee performance in general (as many studies that we reviewed did, see Table 6).

We found strengths and weaknesses in existing research in our analysis of the methodological approaches, which can guide future research. First of all, many methods were applied in the reviewed studies, which is, of course, a strength (see Table 2). Hence, those studies echoed the call by Uhl-Bien et al. (2014) for methodological approaches that can result in both deductively and inductively developed models. Most studies, however, used a quantitative approach rather than a qualitative approach and survey measures were primarily used to inspect the followership variables. Therefore, future research might consider using different methodological approaches (and especially mixed-method approaches), which also corresponds to the argument that a range of paradigmatic perspectives is needed for a true scholarly advance (see Uhl-Bien et al., 2014; Uhl-Bien & Ospina, 2012).

In addition, it is noteworthy that the reviewed studies varied widely with regard to their methodological conceptions and complexity. Moreover, we found quality concerns in a large number of the reviewed studies. A key problem was that most of the quantitative studies were cross-sectional. This is a relevant shortcoming of current followership research, because research questions almost always refer to the temporal link between followership variables and related outcomes, which cannot be properly determined when both are measured at the same time (see, for instance, Mitchell & James, 2001). Even if we found a clear trend toward more followership research activity in our systematic review, taking these methodological issues into account, extant empirical evidence is still limited. The problem of relying on cross-sectional data, however, is not unique to followership research and it is also prevalent in leadership and other organizational studies (see, for instance, Ployhart & Vandenberg, 2010). Future research, therefore, should conduct more longitudinal studies and use more experimental designs or even mixed-method approaches. These methodological approaches would help to gain a better understanding of the cause-and-effect relationships in leadership and followership, which is essential to test the appropriateness of both a “reversing the lens” and a co-construction approach. Regarding the co-construction approach, quantitative studies can also supplement existing qualitative studies. Extensive longitudinal studies (e.g., diary studies; see Gabriel et al., 2019) can be used to examine the reciprocity or the mutual interplay of claiming and granting of leadership and followership. Furthermore, behavioral observation designs offer the opportunity to investigate leader-follower interactions with high temporal resolution (Klonek et al., 2019).

In our review, we could identify a large number of measures that were applied in the reviewed studies. While only a few instruments were developed to measure followership explicitly (e.g., Kelley’s [1992] followership behaviors and styles, implicit followership theories, or follower role orientation), our findings provide a wide variety of measures that fit in the integral followership framework of the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018; see Table 5). Moreover, our analysis revealed options to adapt measures from other contexts to comply with the followership role (e.g., Metwally et al., 2018; J. Yang et al., 2021), which opens up new ways to assess various kinds of characteristics and behaviors that followers might display to contribute to or withdraw from the leadership process. However, many studies that we reviewed relied on face validity. Some studies even displayed questionable operationalizations and measurements. Therefore, more research is needed to demonstrate the required validity and reliability of the numerous followership measures (e.g., for abusive followership or emotional masking, see Table 5).

Another cause for concern is that only six out of the 26 included studies that were published after 2020 considered potential impacts of the COVID-19 pandemic. This means that most studies neglect this important topic, although the pandemic challenged both organizations and their members to adapt to a drastically changed environment of an unexpected external crisis. Organizations and their members were forced to react constantly to changing conditions, lock downs, and varying legal requirements that led to internal adjustments of work organization and forms of collaboration–at least temporarily. Carsten et al. (2021), for instance, pointed out that the COVID-19 pandemic created a context of high physical separation and varying interaction frequency between leaders and followers as increased numbers of employees began to work remotely (see also Hickman & Robison, 2020). Their study on follower adjustments to distal leadership during COVID-19 suggests that the link between interaction frequency with the leader and the followers’ level of engagement depended on the followers’ role orientation (Carsten et al., 2021). Hence, more studies would have been and still are desirable that study potential impacts of the pandemic on followership experiences, behaviors, and outcomes.

Moreover, future research could further explore the role of followership in contexts of global crises. Nohria (2020), for instance, suggested that in the complex and uncertain environment of a sustained, evolving crisis, the most robust organizations will not simply rely on centralized leadership or specialized risk management teams, but on the networks and members’ adaption abilities within the organizations. Andersson (2018) argued that the cooperative relationships between leaders and followers facilitate the mobilization of resources, especially in times of crisis. He concluded that developed followership is an important social resource for organizational resilience. In consideration of future challenges that could emerge from for instance a pandemic, technological change, disruptive innovation, or climate change, a better understanding of beneficial and destructive followership has become even more relevant (see also Ribbat et al., 2023). For instance, critical followership may become an essential skill to question AI-based decisions that are at risk to be incomprehensible, unfair, or biased due to scarce or false data (see, for instance, Guan et al., 2022; Tambe et al., 2019). Hence, future research should find out how followership has to and will change in relation to AI-based management. Furthermore, the digital transformation of work and organizations raise important questions for change management: Can followership theory help to explain why some leaders fail to communicate or effectively implement their vision for the organization’s digital future? Or how can followers be the driving force for change when leaders resist to adapt to technological change?

Our review contributes to both the further identification and conceptual clarification of followership constructs. We tested whether the theoretical principles of the FTF could be applied to identify the proposed followership approaches and variables within published empirical studies. Additionally, we tested whether these principles were sufficient to clearly delineate between followership variables and studies and non-followership variables and studies. In this process, we identified the need for an important clarification to be able to decide what can and what cannot be classified as a followership study in the strict sense. In this way, we clearly differentiate from the analysis of broadly construed follower-related predictors in leadership by Oc et al. (2023). That is, from an analytical standpoint, we have to distinguish between ‘true’ followership constructs (i.e., followership characteristics and behaviors) on the one hand and follower-related variables that refer to general characteristics or behaviors at work on the other when evaluating or constructing potential followership research. It is important to make such a distinction, because a preferably concrete determination of the followership domain is essential for FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018) to be and further become a valuable theoretical framework. Hence, we were able to advance insights on followership from existing reviews (e.g., Oc et al., 2023) with more theoretical clarity and parsimony.

Reflecting a lack of theoretical clarity, we found some misconceptions in the operationalization of followership characteristics and behaviors during the study selection for our systematic review. Specifically, we had to exclude a few studies from our systematic review because of the misconception of such variables, even if those studies claimed to be followership studies. For instance, several studies (e.g., Ahmad et al., 2021; A. J. Xu et al., 2019) “reversed the lens” by using proactive employee behaviors (such as organizational citizenship behaviors) to predict leadership outcomes (such as LMX). These behaviors were, however, not specifically related to a leader (and, thus, also not to a follower role). Therefore, those studies were not followership studies in accordance with the definitions by the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018). These examples show that even if studies refer to the framework of the FTF, they do not automatically correspond to the definitions of the FTF. Followership research aims to better understand the role and contribution of the followers and following in the leadership process (see Uhl-Bien et al., 2014). Therefore, general characteristics (such as personality, skill or well-being) or behaviors (such as engagement or proactivity) of employees are not followership constructs per se, even if those employees are technically subordinate to a leader.

Our systematic review, however, reveals several ‘true’ followership characteristics and behaviors that were explored within the reviewed studies and we discovered followership constructs that go beyond Uhl-Bien’s et al. (2014) suggestions. In sum, we could identify 23 followership characteristics or behaviors that extend the followership framework (see Table 4). Some authors, for instance, introduced group-level variables to the framework of followership constructs (e.g., group level implicit followership theories). This raises awareness of the need for multi-level perspectives in followership research: Followers of the same leader (i.e., within the same team) might share certain similarities. With the help of the related primary studies, we thus contribute to a more nuanced understanding of followership and add to the framework of the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018).

Our systematic literature review shows that more attention should be paid to the contributions of followers to leadership and organizational success, since an increasing number of studies conceptualizes and demonstrates followers as relevant co-producers or co-constructors of leadership and its outcomes. Various findings associate proactive and constructive followership with desirable outcomes, while destructive and passive followership were rather associated with negative implications. This highlights potential benefits of followership trainings (see, for instance, calls of Bufalino, 2018, or Hoption, 2014) and integrated development programs for leaders and followers. Reflecting on findings about followership and abusive supervision or leader well-being/exhaustion, trainings, in which followers and their leaders can jointly learn how to engage in constructive and mutually supportive behaviors, may be promising.

Furthermore, our review outlines which existing followership measures were applied in the reviewed studies, which measures were valid and reliable, and which were not. Thereby, we provide an overview of practical tools not only for researchers, but also for organizations that might want to assess followership behaviors and/or characteristics to develop followership competencies. Given the various endeavors of organizations to select and develop leaders, organizations should generally be aware of the important role of followers and following in the leadership process. This is especially important, as modern organizations have an increasing focus on participation and empowerment (e.g., Maynard et al., 2012; Parker et al., 2019), placing followers in a more influential role.

Some limitations of this review have to be discussed. We only included published peer-reviewed articles in our review. While this is a common strategy for systematic reviews (e.g., Boon et al., 2019; H. Chen et al., 2022), which also helps to avoid double inclusion of studies (e.g., from dissertations), we might have missed relevant published work (e.g., book chapters) or unpublished work (due to publication bias of non-significant findings; see, for instance, Siddaway et al., 2019). Furthermore, we only included studies written in English, which involves the risk of an ethnocentric bias (see Fischer et al., 2021; Steel et al., 2021). A considerable number of studies that we included in our review, however, came from non-English speaking countries (e.g., China or other Asian countries). Additionally, we excluded studies that were published before 2014. Hence, we might have missed relevant research, which was published before 2014, that would meet our theoretical inclusion criteria for the co-production or co-construction approach. However, with our research questions, we intended to review the empirical followership research since the publication of Uhl-Bien’s et al. (2014) seminal work.

Finally, we could not present the findings of each study in detail. Since we discussed how the FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018) is a rather broad framework that comprises a large number of different followership constructs, we could only provide a concise overview of general findings. However, with our research questions, we intended to identify and analyze the studied variables, the applied followership approaches, the methodological approaches and the main and preferably cumulative results rather than the specific results of the individual studies. Furthermore, several studies had strong limitations due to their cross-sectional nature or small samples (see also our quality assessment). Hence, their findings have to be interpreted with caution.

Our systematic review of research in line with FTF (Uhl-Bien et al., 2014; Uhl-Bien & Carsten, 2018) revealed that FTF provides a valuable theoretical framework to integrate a wide variety of research that contributes to a better understanding of the role of followers and following in leadership. However, we argued that it is critical to delineate between ‘true’ followership constructs and constructs that refer to followers’ general characteristics or behaviors at work. That is, future followership studies should consider the definitions by Uhl-Bien et al. (2014) to actually advance followership research and to avoid simply reproducing familiar leadership research from another perspective. With our review, we could provide both a further identification and conceptual clarification of followership constructs.

Our systematic review revealed that followership research is indeed an emerging field. However, most empirical studies are still rather fundamental. Methodological and conceptual issues also currently limit empirical evidence. We found that the study of the two fundamental followership approaches within the FTF was unbalanced, which was also the case for the study of various followership constructs. Hence, even if empirical followership research has developed quantitatively, it still lacks depth in several ways, thus preventing cumulative knowledge. In sum, our analysis of empirical followership research since Uhl-Bien’s et al. (2014) seminal work offers various opportunities for future studies to advance the current knowledge about the role of followers and their followership, both theoretically and methodologically.

Contributed to conception and design: MR, JH

Contributed to acquisition of data: MR

Contributed to analysis and interpretation of data: MR, KNK

Drafted and/or revised the article: MR, KNK, JH

Approved the submitted version for publication: MR, KNK, JH

We have no known conflict of interest to disclose.

We pre-registered our research questions and procedures at the Open Science Framework: https://osf.io/h4k8a/?view_only=5efb33a185b44b8fa7788760755babdf. The search string for our literature research as well as a comprehensive overview of all included studies that contains information on the chosen followership approach, the unit of analysis, the methodological approach, the sample, and a brief summary of results can be found in the supplementary material.