Why are authentic workplaces more ethical

How authentically do I lead? Examination of the self-assessment form of the German Inventory of Authentic Leadership


This article from the magazine “Gruppe. Interaction. Organization. (GIO) "deals with the examination and validation of the self-assessment form of the German Inventory of Authentic Leadership.

Authentic leadership has emerged as a central positive leadership concept in recent years. However, so far there is only one instrument for external assessment in the German-speaking countries: The German Inventory of Authentic Leadership (DIAF-F). In this contribution, an instrument for self-assessment of authentic leadership (DIAF-S) with the components self-confidence, transparency in relationships, balanced information processing and internalized moral perspective is presented and tested. For the first validation of the new procedure, data from N = 91 executives and N= 259 employees recorded. Internal consistencies in the satisfactory to good range were found for the overall scale and three of the four components. The DIAF-S showed significant correlations with external organizational criteria (work performance of employees assessed by managers, voice behavior and moral courage). In addition, for a large part of the components, the expected moderate correspondence between self-assessed and external-assessed authentic leadership was shown. Using confirmatory factor analyzes, it was possible to justify the use of the four-factor structure analogous to external assessments. An exploratory factor analysis suggested a three-factor structure of self-assessed authentic leadership. Against this background, alternative models of authentic leadership are discussed. Overall, the DIAF-S is a largely reliable and valid instrument for the self-assessment of authentic leadership and thus an important addition to the DIAF-F.


This paper of the magazine “group. Interaction. Organization. (GIO) ”presents the psychometric properties and validation of the self-assessment form of the German Inventory of Authentic Leadership.

Authentic leadership has become one of the central positive leadership styles in recent years. Up to now, there has only been an other-rating version in German language to assess authentic leadership: The German Inventory of Authentic Leadership (DIAF-F). This article presents and tests the new instrument DIAF-S for self-assessment of authentic leadership with the components self-awareness, relational transparency, balanced processing and an internalized moral perspective. For first validation, the DIAF-S was tested on N = 91 leaders and N = 265 employees. The overall scale and three of four subscales showed satisfying to good internal consistencies. The DIAF-S correlated significantly with criteria of organizational performance (work performance of employees, voice behavior and moral courage rated by leader). In addition, for a majority of the components the expected moderate correlation between self- and other ratings of authentic leadership was evident. The use of the four-factor structure could be verified by confirmatory factor analysis in analogy to the other-rating version of authentic leadership.

An explorative factor analysis revealed rather a three-factor structure of self-rated authentic leadership. In this context, we discuss alternative models of authentic leadership. Overall, the DIAF-S is a widely reliable and valid instrument for self-assessment of authentic leadership and thus an important supplement to the DIAF-F.

The great thing is not to be this or that, but to be yourself. (Søren Kierkegaard)

Kierkegaard already saw the true greatness of a person to reveal their authenticity and the true self. Authenticity means having a deeper understanding of oneself and acting in harmony with one's own thoughts, feelings and needs (Kernis 2003).

Authentic behavior can be understood as an important resource for moral behavior (Hannah et al. 2011). When it comes to moral issues in organizations, executives play a key role, as they influence the behavior of employees through setting an example (Treviño et al. 2014). Correspondingly, authentic leadership behavior by setting an example of trustworthiness and moral values ​​can promote moral behavior on the part of employees (Brown and Mitchell 2010).

Authentic leadership

Luthans and Avolio (2003) characterize an authentic manager as self-confident, transparent, oriented towards morals and ethics and interested in developing employees into managers themselves. Authentic leadership is currently defined as “(...) as a pattern of leader behavior that draws upon and promotes both positive psychological capacities and a positive ethical climate, to foster greater self-awareness, an internalized moral perspective, balanced processing of information, and relational transparency on the part of leaders working with followers, fostering self-development ”(Walumbwa et al. 2008, p. 94). The model of authentic leadership according to Walumbwa et al. (2008) comprises four behavioral components: Self-confidence (SB) is the self-reflection of the manager, the profound knowledge of one's own strengths and weaknesses, the anticipation of one's own impact on others and the gathering of feedback; Transparency in relationships with others (TiB) includes showing the true self to other people without suppressing or disguising thoughts or feelings; Balanced information processing (AIV) records the objective analysis of relevant information in the decision-making process, whereby managers obtain the opinions of employees and provide their own views on the discourse; Internalized moral perspective (VMP) describes how much a manager's decision-making and actions are guided by internalized moral standards and values.

It is assumed that authentic leadership behavior has positive effects on the attitudes and behavior of employees (Walumbwa et al. 2008). Numerous studies have identified positive connections between authentic leadership and behavioral and affective measures such as employee performance, engagement, organizational citizenship behavior and employee job satisfaction (Banks et al. 2016; Hoch et al. 2018).

The German Inventory of Authentic Leadership (DIAF)

The DIAF as a form of external assessment (DIAF-F)

Based on the model of authentic leadership (Walumbwa et al. 2008), the German inventory of authentic leadership in the form of external assessment (DIAF-F, Franke-Bartholdt et al. 2018) was created as a German translation of the Authentic Leadership Inventory (ALI, Neider and Schriesheim 2011) (for the process development of the DIAF-F see Franke-Bartholdt et al. 2018).

The quality and validity of the DIAF-F was checked in several studies (Franke-Bartholdt et al. 2018): When the internal structure was checked by means of a confirmatory factor analysis, a model-compliant four-factor structure resulted. The model was superior to a model with four first-order factors and one second-order factor (total value of authentic leadership) and a model with one factor. The check with an exploratory factor analysis resulted in a two-factor solution in which the items of the components SB, AIV and TiB loaded to factor 1, the items of the component VMP to factor 2. As expected, the DIAF-F correlated highly with facets of ethical leadership (Brown et al. 2005). High, negative correlations between the Abusive Supervision management style (Tepper 2000) and the DIAF-F demonstrated the distinction between the two management forms. With external organizational criteria such as self-assessed performance, work commitment and well-being of the employees, positive, significant associations emerged. Authentic leadership went beyond ethical leadership to clarify variance in the criteria of performance and leader-member exchange.

The DIAF as a form of self-assessment (DIAF-S): goals and hypotheses

In particular when evaluating their own leadership style, managers tend to exaggerate self-assessment (Yammarino and Atwater 1997), so that the presence of a test procedure as a form of self and external assessment provides added value, especially in the area of ​​leadership behavior. Therefore, the present work aims to comprehensively test and validate the DIAF-S as a supplement to the DIAF-F.

The DIAF-S was derived from the DIAF-F (Franke-Bartholdt et al. 2018) by transforming the items into the first-person perspective. The DIAF-S consists of 16 items and forms the four components of authentic leadership Self-confidence (e.g. "I am clearly aware of the effect I have on others.") Transparency in relationships with others (For example, “I openly say how I really feel and do not pretend.“), internalized moral perspective (e.g. "I act in accordance with my beliefs and values.") and balanced information processing (e.g. "I also actively ask about viewpoints that question my beliefs.") with four items each. The items are on a five-point Likert scale (from 1 =I do not agree to 5 =I agree completely) to answer.

In order to obtain information about the suitability of the individual items for the components and the overall scale, the psychometric properties of the DIAF-S were tested. To determine the construct validity, the internal factor structure was determined by means of confirmatory and exploratory factor analysis. For the comparability of the DIAF-S with the DIAF-F, the factorial validity of the DIAF-F was also determined.

To test the criterion validity, relationships with performance and pro-organizational behavior, operationalized as voice (Liang et al. 2012) and moral courage (Hannah et al. 2011), were used. The following hypotheses are to be checked.

Authentic managers promote self-determination and self-regulation of their employees, so that this increases the motivation and satisfaction of the employees, which in turn has positive effects on work performance (Leroy et al. 2015). The positive relationship between authentic leadership and performance has been replicated several times (Hoch et al. 2018; Banks et al. 2016). Various studies showed empirically a small effect (ρ = 0.14) between authentic leadership and the performance of employees. This results in:

Hypothesis 1

There is a low positive correlation between self-assessed authentic leadership and the performance of employees.

Authentic leaders act as role models and reflect integrity, honesty and high moral standards in relationships with employees (Brown and Mitchell 2010). By acting as role models, they promote social learning processes among employees and form the normative context for moral action. Accordingly, authentic managers can influence the moral attitudes and behaviors of their employees, which support the functionality of the organization (Treviño et al. 2014). In this way, authentic managers promote that the employees themselves are more willing to initiate changes and to show what is known as voice behavior (Wong and Cummings 2009), and to adhere to moral standards in the form of moral courage (Hannah et al. 2011). Liang et al. (2012) differentiate between prohibitive voice, actively addressing problems in the work environment and promotive voice, expressing ideas for improving work processes. It has been shown empirically that authentic leadership has an indirect effect, conveyed through trust in the manager, on the voice behavior of employees (Wong and Cummings 2009). Meta-analytically, a relationship of ρ = 0.31 was found between authentic leadership and voice behavior (Banks et al. 2016). A connection between authentic leadership and moral courage could be found in the middle range (r = 0.29) can be found (Hannah et al. 2011). The following hypotheses result:

Hypothesis 2a and 2b

There is a moderate positive correlation between self-assessed authentic leadership and the promotive voice or prohibitive voice behavior of employees.

Hypothesis 2c

There is a moderate positive correlation between self-assessed authentic leadership and the moral courage of employees.

The external assessment of the employees is used as a further criterion. In particular, there is a risk of overestimating oneself when assessing one's own leadership style (Yammarino and Atwater 1997). In a metanalysis by Lee and Carpenter (2018), the correspondence between self-assessment and external assessment was tested for various leadership styles: For relationship-oriented leadership styles, which also include authentic leadership (Lee and Carpenter 2018), a correlation of ρ = 0 was found , 27, which leads to the following assumption:

Hypothesis 3

There is a moderate positive correlation between self-assessment and authentic leadership that is assessed by others.



Participated in the study N = 91 managers (68.5% female) from four organizations in the health sector participated (age: M. = 45,43, SD = 9.61, range 26-61). Of the respondents, 31.1% had been employed in the current field of work for less than five years, 38.9% between 6 and 20 years and 30.0% for more than 20 years. The participants were recruited as part of the DFG project FIDES (Leadership in organizations as a determinant of the silence of employees on morally problematic issues) in which the connection between (un) moral forms of leadership and the behavior of employees in morally relevant situations is examined. A positive ethics vote was given for the project (GZ STR 1005 / 5‑1; WE 1504 / 25-1). The questionnaires were filled out using an online survey or a paper-pencil variant. To check the factorial validity of the DIAF-S, a sample of n = 80 managers (66.7% female, age: M. = 44.86, S.D. = 9.65); for the DIAF-F, the corresponding total sample of employees was used, which entered after the exclusion of employees with missing values n = 620 (76.1% female, age: M. = 42.21, S.D. = 12.32).

A random sample of n = 42 executives (66.7% female, age: M. = 44.00, S.D. = 8.78) and n = 259 employees (76.8% female, age: M. = 40.02, S.D. = 10.98) is used.


Tab. 4 shows the internal consistencies of the processes used. The self-assessment of authentic leadership behavior was determined using DIAF-S using four items per component on a five-point Likert scale (from 1 =I do not agree to 5 =I agree completely) detected.

The DIAF-F was used for the external assessment of authentic leadership (Franke-Bartholdt et al. 2018), which, analogous to the DIAF-S, consists of 16 items (e.g. "My direct superior / my direct superior listens to alternative points of view carefully before he / she draws any conclusions. "). The items were answered on a five-point scale (from 1 =I do not agree to 5 =I agree completely).

Voice was based on the German translation (Rose 2015) of the questionnaire by Liang et al. (2012), in which promotive and prohibitive voices are asked about five items each (promotive voice, e.g. "The employees in my work area actively suggest new projects that are beneficial for the work area"; prohibitive voice, For example, "The employees in my work area dare to point out problems in the work area, even if this could affect the relationship with other colleagues."). Voice behavior was assessed on a five-point scale (from 1 =I do not agree to 5 =I agree completely).

Moral Courage was recorded using six items of the Moral Potency Questionnaire by Hannah and Avolio (2010) in the German translation by Rose (2015). The executives rated their employees with regard to adherence to moral principles on a five-point Likerst scale (from 1 =I do not agree to 5 =I agree completely; z. B. “The employees in my work area take responsibility and intervene if they observe a morally problematic act.”).

The work performance was recorded using two items (Robinson 1996): The manager assessed the performance of their own employees in comparison to other teams and also stated how the employees would presumably rate their own performance in comparison to that of other teams. The assessment was based on a seven-point Likert scale (1 =significantly worse to 7 =significantly better). Since both items were highly correlated with each other (r = 0,58, p <0.001) a total value for work performance was formed.


The data was evaluated using the statistics program IBM SPSS Statistics (Version 25.0., 2017). An outlier analysis did not reveal any multivariate outliers. Missing values ​​occurred occasionally, which were taken into account by excluding them in pairs.

To test the factorial validity of the DIAF-S, various theoretically derived models were tested using confirmatory factor analyzes (CFA). For the comparability of the DIAF ‑ S with the DIAF ‑ F when determining the connection between self and others as assessed authentic leadership, the factorial validity of the DIAF ‑ F was also determined. The calculation was carried out with the program IBM SPSS AMOS (Version 26.0; Arbuckle 2014). To check the global model fit, the χ2-Test used. To assess the overall fit of a model, Hu and Bentler (1999) recommend with a small sample size (N ≤ 250) additionally the combination of different fit indices: Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean Square Residual (SRMR). The cut-off values ​​proposed by Hu and Bentler (1999) were used to interpret the indices: CFI ≥ 0.95, RMSEA <0.06 (at N ≥ 250) or <0.08 (at N ≤ 250) and SRMR <0.09. In addition, the models were selected using χ2-Difference test compared to each other.

In addition, an exploratory factor analysis (EFA) was carried out for a differentiated recording of the internal factor structure of the DIAF-S. Based on the assumption of correlating factors, a principal axis analysis with oblique rotation (Promax, Kappa = 4) was used. The extraction criterion for the number of factors was the parallel analysis according to Horn (1965; implemented with O'Connor 2000).

To test the criterion validity, correlative relationships between DIAF-S and organizational external criteria were considered. Due to moderate correlations between the DIAF-S and the age and length of activity of the managers, connections with external criteria were determined using partial correlations. The α level was corrected for Bonferroni, which led to an α ′ = 0.05 / 9 ≈ 0.006. In order to enable a comparison between DIAF-S and DIAF-F, the four-factor structure of authentic leadership was used to calculate the criterion validity.

To determine the relationship between external and self-assessed authentic leadership, the information provided by the executives was included, for whom a direct assignment to their employees was possible based on the team code. The mean value of all external assessments was calculated for each manager and correlated with the self-assessment. The data sets from n = 42 executives and n = 259 employees available. For each manager there were between 2 and 15 external assessments (Mdn = 6). In order to be able to assess the quality of the interrater agreement, the rFlat shareValue (within-group correlation) calculated. In order to take into account different response tendencies in the assessment by the employees, values ​​were calculated for the assumption of an evenly distributed response pattern (constant equal distribution) and a response tendency that is subject to lenciency bias (behavior tends to be assessed better than it actually is; slightly skewed distribution assumption) (LeBreton and Senter 2008). The α-level was Bonferroni-corrected for the relationships between self-assessment and authentic leadership as assessed by others, which led to an α ′ = 0.05 / 10 ≈ 0.005.


Descriptive statistics of the items, components and overall scale

Tab. 1 shows relevant parameters (descriptive statistics, discriminations) of the 16 items; Tab. 4 the descriptive statistics of the components and the overall scale as well as the correlations between demographic variables and DIAF-S. Based on the observed relationships, age and length of employment were included as control variables in the following calculations of the relationships between the DIAF-S and the criteria.

The item difficulties were in the low range (3.26 ≤M. ≤ 4.53). The difficulty of the components and the overall scale were also low with values ​​between 3.94 (SB) and 4.28 (VMP). All items showed a skew <| 2 | and kurtosis <| 7 | which suggests a normal distribution (West et al. 1995). There was a left-skewed, right-hand distribution for all items.

The selectivity was mostly in the acceptable range: in relation to the overall scale in the range of 0.28 <rit ‑ G <0.62, for the respective components in the range of 0.22 <rit ‑ K <0.67. The low power of discrimination of two items of the component SB with values ​​was noticeable rit ‑ K <0.30 as well as the low selectivity of the items of the component VMP (rit ‑ G <0.40) (Tab. 1).


The internal consistency of the overall DIAF-S scale can be classified as good with a Cronbach's alpha of 0.82. The internal consistencies of the components were in the range of 0.55 (SB) ≤ α ≤ 0.77 (TiB). For the components TiB, VMP and AIV, the internal consistencies can be classified as satisfactory to good, taking into account the small number of items. The SB component was noticeable for its low internal consistency, which is classified as inadequate.

Factorial validity

To check the internal factorial structure of the DIAF ‑ S, four models were tested against each other in CFA: (1) A model with four intercorrelated factors that depict the components of authentic leadership (Walumbwa et al. 2008), (2) a model with four factors first-order and a second-order factor (total value of authentic leadership), (3) a model with two intercorrelated factors according to the findings of the exploratory factor analysis of the DIAF-F (Franke-Bartholdt et al. 2018) and (4) a model with a global factor authentic tour on which all 16 items were loaded. In order to ensure the comparability of the DIAF-S with the DIAF-F, all four models were also tested for the DIAF-F. The results are shown in Tab. 2.

For the DIAF-S, model 1 with four intercorrelated factors (\ (\ overline {r} \) = 0.50) and model 2 with four first-order factors and one second-order factor proved to be the most suitable for the data the self-assessment of authentic leadership. Model 1 and model 2 did not differ significantly from one another, and the various fit indices did not differ in their height either. The fit indices RMSEA and SRMR were below the cutoff values ​​of Hu and Bentler (1999), but the CFI value did not reach the desired cutoff value for either of the two models. Both models explained the data structure significantly better than models 3 and 4. Since more economical models are to be preferred to more complex models (Steiger 1990), the less complex model 1 was focused with four intercorrelated factors. When looking at the modification indices, it became apparent that correlations between the error variances specified between eleven items would lead to a better model fit.

An EFA using parallel analysis revealed three factors for the DIAF ‑ S, which explained 37.96% of the variance and moderately correlated with each other (0.32 <r <0.44). Tab. 3 shows the factor loadings. All four items of the component TiB as well as three items of the component SB loaded on factor 1. The items of the component AIV loaded together with item 1 of the component SB on factor 2. All items of the component VMP had the highest loads on factor 3 and thus showed a simple structure.

For the DIAF-F, model 1 with four intercorrelated factors ( = 0.85) the best model fit. The fit indices RMSEA and SRMR were below, the index CFI above the cut-off values ​​recommended by Hu and Bentler (1999). Model 1 explained the data structure of the DIAF-F significantly better than the three alternative models, whereby model 2 with four first-order factors and one second-order factor and model 3 with two intercorrelated factors also had satisfactory fit indices (Table 2).

Relationship to criteria

The correlations between the external criteria work performance, voice and moral courage and DIAF-S are shown in Table 4.

TiB (r = 0,25, p <0.05), VMP (r = 0,29, p <0.01) and the total value of the DIAF-S (r = 0,30, p <0.01) correlated significantly positively with the performance of the employees as assessed by the manager. For SB (r = 0,11, p = 0.320) and AIV (r = 0,18, p = 0.102) no significant correlations were found.

Promotive and prohibitive voices showed similar connections to authentic leadership: SB had the highest correlation in each case (promotive voice: r = 0,44, p <0.006; prohibitive voice: r = 0,50, p <0.006) with both types of voice. Promotive and prohibitive voice showed with TiB (0.32 ≤r ≤ 0,36, p ≤ 0.006) and AIV (0.28 ≤r = 0,29, p ≤0.01) showed small to moderate correlations. There were moderate correlations for both types of voice (0.44 ≤r ≤ 0,47 p <0.006). For VMP and promotional (r = 0,15, p = 0.172) and prohibitive voice (r = 0,15, p = 0.166) no correlations were found.

In terms of moral courage, there were positive correlations between SB, TIB, AIV and the total value (0.29 ≤r ≤0,36, p ≤ 0.01). No connection could be established between moral courage and VMP (r = 0,09, p = 0.413) (Tab. 4).

Relationships between self-assessment and external assessment

The descriptive statistics as well as intercorrelations of the self and others assessed authentic leadership are shown in Table 5.

For the interrater agreement of the external assessment of the leadership style, values ​​were obtained for various distribution assumptions (steady and slightly crooked) for all components and the overall scale that, according to LeBrenton and Senter (2008), indicate moderate to high agreement within the teams. For the assumption of a constant uniform distribution, the component AIV had the lowest value, the component VMP the highest value in the agreement (0.68 ≤rdue to ≤ 0.81). For the assumption of a slightly skewed distribution, component AIV achieved the lowest value, component VMP the highest (0.52 ≤rdue to ≤ 0,71).

Due to moderate to high correlations between various components and the overall scale of self-assessed authentic leadership with age (0.38 ≤r ≤ 0,55, p <0.05) and with duration of activity (0.44 ≤r ≤ 0,57, p <0.05) partial correlations were then calculated.

Between the external and self-assessment there were (r = 0,46, p <0.005) and for SB (r = 0,35, p <0.05) and TiB (r = 0,62, p <0.005) moderate to high correlations. In AIV, there was no correlation between self and external assessments (r = 0,16, p = 0.325), there was also no correlation for the VMP component (r = 0,09, p = 0,565).


The contribution includes the examination and validation of the first German form of self-assessment of the inventory for recording authentic leadership - DIAF-S.

The quality and the factorial internal structure were tested. To assess the criterion validity, external organizational criteria - performance, voice behavior and moral courage - were used and the connection between self-assessed and others-assessed authentic leadership was considered. The DIAF-S was based on self-assessments by N= 91 executives and external assessments by N = 259 employees checked.

Significant correlations emerged between authentic leadership and age and duration of activity. Older and longer-active managers judged themselves to be more authentic leaders at all levels than younger ones or managers who were less active. One explanation could be that a positive self-image and the perception of one's own leadership behavior are confirmed and strengthened over time through suitable situations.

The item difficulties can be classified as easy. Accordingly, the managers rated themselves high on average for all components, which is also made clear by the comparison with the external assessment form (DIAF-F, Franke-Bartholdt et al. 2018) with medium item difficulties. In a self-assessment procedure such as the DIAF-S, increased mean values ​​are to be expected, since managers in particular can be subject to a positive bias towards their own behavior (Yammarino and Atwater 1997). The items represent desirable behavior so that socially desirable response behavior becomes more likely. The differentiation between people with high and low characteristics is therefore made more difficult. In particular, the VMP component, which records the management's alignment with high moral standards, had the highest mean value and the lowest selectivity. It is all the more important to use the two forms of self-assessment and assessment by others in order to capture both perspectives. The item selectivity for the majority of the items was in the targeted middle range, with two items of the component SB having a selectivity rit <0.30 are to be regarded as critical. Compared to the excellent selectivity of the DIAF-F, however, the selectivity of the DIAF-S is significantly lower, which can be attributed to the homogeneous sample and the associated limitation of variance.

The psychometric properties of the DIAF-S can be rated as satisfactory to good for the most part. The reliabilities of the components TiB, AIV, VMP and the overall scale were in the satisfactory to good range, while the reliability of α = 0.55 for the component SB cannot be regarded as satisfactory. Therefore the items of the component SB should be revised. One approach is provided by the result of the EFA, in which the items of the components SB and TiB were found together with a factor of 1, while item 1 of the component SB together with the items of the component AIV loaded with a factor of 2. An exploratory consideration of the reliability for the factor solution of the EFA resulted in satisfactory internal consistencies: For factor 1 with three items of the component SB and four items of the component TiB a Cronbach's alpha of 0.74, for factor 2 with item 1 of the component SB and four items of the component AIV a Cronbach's alpha of 0.66 and for factor 3 with the items of the component VMP a value of 0.65. By adding item 1 to component AIV, the reliability drops from 0.69 to 0.66, which is why item 1 should be revised.

When examining the internal structure of the DIAF-S and DIAF-F via CFA, analogous to the procedure of Franke-Bartholdt et al. (2018) tested different models: Model 1 with four intercorrelated factors, Model 2 with four factors of the first and one factor of the second order, Model 3 with two intercorrelated factors and Model 4 with one factor.

For the DIAF-S, model 1 with four intercorrelated factors and model 2 with four first-order and one second-order factors achieved the same parameters in the fit indices and did not differ significantly from one another. Both models reproduced the data significantly better than the alternative models, but showed a limited overall fit. For reasons of economy, the less complex model 1 with four intercorrelated factors SB, TiB, AIV and VMP should be used.

For the DIAF-F, the model with four intercorrelated factors showed a satisfactory overall fit and was significantly superior to the three alternative models. Thus, the Franke-Bartholdt et al. (2018) confirmed the four-factor structure for the form of external assessment of authentic leadership.

In summary, a model with the intercorrelated factors SB, TiB, AIV and VMP explained both forms best for assessing authentic leadership compared to the alternative models. The four-factor structure corresponds to that of Walumbwa et al. (2008) postulated, which could be empirically proven several times (Franke-Bartholdt et al. 2018; Neider and Schriesheim 2011). Accordingly, it is possible to compare the self-assessment and the third-party assessment at the component level.

However, no satisfactory overall fit could be achieved for the DIAF-S. The results of the CFA should be due to the small sample size of n = 80 should be interpreted with caution, since estimation problems occur more frequently with small samples (Bühner 2010) and the present number of test subjects is below the recommendations of the required size for the calculation of CFA (Schermelleh-Engel et al. 2003).

The inadequate fit of a model with four intercorrelated factors for the DIAF-S gives rise to further research into the structure of authentic leadership in self-assessment. One approach to this is provided by the result of the EFA, which resulted in a three-factor solution. The items of the components SB and TiB were found on a common factor, only item 1 of the component SB loaded with the items of the component AIV on factor 2. If you look at the item (“I ask for feedback in order to improve my dealings with others . ”) It becomes clear that the content is about gathering information, which suggests a process component that is characteristic of AIV. Due to the low selectivity (rit ‑ K = 0.22) and the high main load on factor 2, the item should be revised or removed linguistically. The factor structure found with exploratory analyzes does not correspond to that of Walumbwa et al. (2008) postulated structure, but can be comprehensibly explained with an alternative model of authentic leadership: In the Theory of Embodied Authentic Leadership (Ladkin and Taylor 2010) the behavioral aspect of Self-esteem, which describes the components SB and TiB. Self-esteem means having a deep awareness of your own emotional life and being able to decide whether emotions are expressed. It is important that managers can also show weaknesses and that employees identify with the manager through their open approach to them. This promotes trust and deeper relationships. Thus, the component SB can be assumed to be the foregoing condition that promotes TiB. The second behavioral component Make references according to Ladkin and Taylor (2010) can be seen as a counterpart to component AIV. It is about the importance of social exchange in order to make decisions for the good of the organization. Opinions may also be expressed here that are not in agreement with those of the manager. As a result, it has to adapt to the contextual requirements and obtain information from employees. Finally, the behavioral component of the Management decisions, that the manager reflects the identity of the group and creates the framework for development processes, which is promoted by shared moral values. This can be assigned to the DIAF ‑ S component of the VMP accordingly. Another three-factor model of authentic leadership is presented by Whitehead (2009). An authentic manager is “(1) (...) self-aware, humble, always seeking improvement, aware of those being led and looks out for the welfare of others; (2) fosters high degrees of trust by building an ethical and moral framework; and (3) is committed to organizational success within the construct of social values ​​”(Whitehead 2009, p. 850). The properties under (1) thus map the components SB and TiB, under (2) aspects of the component VMP are shown and (3) corresponds in a broader sense to the component AIV.

The expected moderate correlations between the external criteria work performance, voice and moral courage could be confirmed for SB, TiB and AIV and the overall value and are to be rated as promising. Similar high correlations were found for the criterion validity as in the DIAF-F. The connections between VMP and promotive and prohibitive voice and moral courage were contrary to hypotheses. Small, insignificant correlations were found here. The executives surveyed assessed themselves as being very homogeneous on all components, but especially on VMP, so that a reduction in variance and thus a reduction in correlation can be expected here. In addition, the items on the internalized moral perspective reflect internal processes and not directly observable behavior. In order for this aspect of authentic leadership to be related to visible, active behavior such as voice and moral courage, the behavior must also be perceptible.

When looking at the correspondence between self-assessment and external assessment, moderate to high values ​​were achieved for the agreement of employees within a team. This means that the formation of the mean can be declared permissible (LeBreton and Senter 2008). It should be noted, however, that the variance in agreement was quite high. The assessment of the leadership style can also vary within a team, as a manager can have different relationship quality due to differences in the leader-member exchange with different employees. This can lead to different patterns of assessment of leadership style (Estel et al. 2019). For the components SB, TiB and the total value, moderate to high correlations were found between self and external assessments. The greatest correlation existed for the component TiB: This depicts the relational aspect between employees and managers and can therefore be assessed equally well by both sides based on the behavior shown. The difficulty of assessing invisible behavior from an internalized moral perspective was also demonstrated by the lack of a connection between external and self-assessment. Metanalytically, ethical leadership that has great overlaps with authentic leadership was found to have the greatest effect size compared to transformational or serving leadership for the difference between self-assessment and external assessment (Lee and Carpenter 2018). It can be assumed that there is a large discrepancy between self and external assessments, especially when it comes to content on moral behavior.

Limitations and recommendations

The testing and validation of the DIAF-S was implemented using a competitive design. This means that no statements can be made about the predictive validity of the DIAF-S. In the future, this could be solved by longitudinal investigations. The executives assess themselves to be very homogeneous with regard to authentic leadership, which may represent the found relationships to a limited extent. In addition, the sample was very homogeneous, as all of the executives surveyed worked in the health sector. Testing the DIAF-S on a larger, heterogeneous sample of executives is therefore recommended, especially with regard to checking the factorial validity. Above all, industries in which moral standards play an important role should be included. In order to be able to make comprehensive statements on the criterion validity, findings on affective criteria of employees (well-being, work commitment), but also on relational criteria (leader-member exchange), as implemented in the DIAF-F, would be informative. Checking the factorial validity using CFA does not currently provide any satisfactory model fits. The three-factor structure found in the EFA shows that the theoretical conception of authentic leadership is not yet complete. A model with three intercorrelated factors should therefore be examined more specifically in further studies. Further research is required in order to sharpen the construct of authentic leadership and to delimit the individual components more precisely or to arrive at a more suitable structure for self-assessment.


With the DIAF-S, in addition to the external assessment form, the DIAF-F, there is now a largely reliable and valid instrument for self-assessment of authentic leadership behavior. The results of this study indicate predominantly good psychometric properties and show promising results with regard to external organizational criteria. The low internal consistency of the SB component must be taken into account, which is why an alternative structure of authentic guidance is recommended when the DIAF-S is used individually. This is based on three factors that can explain the self-assessment form on the basis of theoretical alternative models and whose application resulted in internal consistencies in the satisfactory range (0.65 ≤ α ≤ 0.74).

The four-factor model with the components SB, TiB, AIV and VMP can be used to compare the self-assessment with that of others, and it is therefore recommended that the DIAF-S and the DIAF-F be used jointly. With the help of these two forms of DIAF, statements about the authentic leadership style can be derived from multiple perspectives. This is an immense advantage over looking at management behavior alone from the perspective of oneself or others, as expectations of the management style can be determined together and improvement measures can be initiated if discrepancies arise (Franke-Bartholdt et al. 2018).

The DIAF ‑ S is an important instrument for executive coaching, as the behavioral items can be used to reflect on the understanding of roles and the role model function, especially when it comes to dealing with moral issues. In this area in particular, the findings of this study indicate a large discrepancy between self-perceived behavior and behavior that is perceived by others.

Correspondingly, the comparison between the management behavior perceived by the employees and the self-assessment of the manager can provide an impetus to identify deviations. On the basis of these deviations, managers can learn to deal confidently with possible inadequacies as managers and develop further. When assessing the self-assessment, the context should also be taken into account: If executive coaching is specifically about reflecting on one's own leadership style and developing it further, executives respond more honestly and admit deficiencies more than in other situations (Lee and Carpenter 2018). In the context of determining one's position, for example when personal decisions are made on the basis of leadership style, socially desirable response behavior becomes very likely. For this reason, it is precisely then that external assessments should be obtained in order to reduce this effect.

Furthermore, with the help of the DIAF, the nomological network of authentic leadership can be researched and expanded and knowledge about one of the most important characteristics in the organizational context can be gained: being true to yourself and thus creating a culture of trust that promotes an open discourse.


  1. Arbuckle, J. L. (2014). Amos (version 26.0). Chicago, IL: IBM SPPS. Computer Program

    Google Scholar

  2. Banks, G. C., McCauley, K.D., Gardner, W. L., & Guler, C. E. (2016). A meta-analytic review of authentic and transformational leadership: a test for redundancy. The Leadership Quarterly, 27(4), 634–652.

    Article Google Scholar

  3. Brown, M. E., & Mitchell, M.S. (2010). Ethical and unethical leadership: exploring new avenues for future research. Business Ethics Quarterly, 20, 583–616.

    Article Google Scholar

  4. Brown, M. E., Treviño, L.K., & Harrison, D.A. (2005). Ethical leadership: a social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes, 97, 117–134.