Good intentions gone awry: investigating a strategically oriented MLD program

Abstract

Purpose

Complex organizations increasingly rely on middle managers as strategic linking-pins between the top and bottom levels of the organization. Using social identity theory and commitment theory as the foundation, this study evaluates a management and leadership development program (MLDP) intended to engage middle managers as strategy creators and implementers. We also evaluate the cascading effects of leadership development by assessing changes in subordinates' identification with the leader, and commitment to the work unit and organization.

Design/methodology/approach

Using a sample of 107 manager participants and 913 of their subordinates, this study measures differences in both manager and subordinate identification and commitment prior to and after the completion of a 6 months strategically oriented MLDP.

Findings

Despite the organizations' best intentions, manager identification with and commitment to the organization decreased after completion of the MLDP. Similarly, subordinates identification with the leader and commitment to the organization also decreased at Time 2.

Research limitations/implications

The results paint a complex picture of the nuances of social identification as an outcome of MLDPs, and problematize the notion of cascading effects on subordinates within the organization. Researchers are encouraged to further examine organizational attitudes and perceptions as outcomes of MLDPs.

Practical implications

Suggestions are offered regarding how practitioners can manage strategically oriented MLDPs in order to avoid identity confusion and promote strategic action.

Originality/value

Strategically oriented MLDPs are increasingly popular in organizations. This study is one of the first to evaluate the theoretical mechanisms through which these programs may affect managers and problematize these effects for complex organizations.

Citation

Larsson, M., Carsten, M. and Knudsen, M. (2020), "Good intentions gone awry: investigating a strategically oriented MLD program", Journal of Management Development, Vol. 39 No. 3, pp. 334-354. https://doi.org/10.1108/JMD-12-2018-0373

Good intentions gone awry: investigating a strategically oriented MLD program

In today's dynamic and complex organizations, middle managers are increasingly asked to assume more strategic responsibility and align themselves with the strategic imperative of the organization (Floyd and Wooldridge, 1994; Rouleau, 2005; Rouleau and Balogun, 2011). Yet, in complex organizations, this task becomes increasingly complicated due to the variations in organizational subsystems, including different structural entities (i.e. divisions, units and locations), as well as different professions with their own unique subcultures (Bloor and Dawson, 1994). Thus, how middle managers enact their role as linking-pins between top management and lower level employees is critical for strategic alignment and implementation (Herzig and Jimmieson, 2006; Rouleau, 2005; Rouleau and Balogun, 2011) because they are in the best position to connect strategic and operational subsystems (Herzig and Jimmieson, 2006; Rouleau, 2005; Smith and Woodworth, 2012).

In many organizations, MLD programs are utilized to strengthen middle managers' capacity to support and implement strategic initiatives (Bonn, 2005; Day and Dragoni, 2015). Yet, the current state of the MLD literature does not include much theoretical or empirical evidence that strategically oriented MLD programs have the desired organizational effect (Clarke, 2012; McGurk, 2010). On the contrary, both HR managers and executives have lamented the lack of connection between MLDPs and strategic or organizational level outcomes (Ardichvili et al., 2016; Bolden, 2016; Fernandez-Aráoz et al., 2017; Gurdjian et al., 2014). The MLD literature has repeatedly demonstrated positive effects on an individual participant level, including gains in personal skill acquisition such as transformational leadership behaviors (Chaimongkonrojna and Steane, 2015; Dvir et al., 2002), self-efficacy (Holmberg et al., 2016; Packard and Jones, 2015), political skill and similar psychological variables (see Collins and Holton, 2004; Day and Dragoni, 2015 for a review). The prevalent assumption in the MLD literature is that increases in individual capacities easily translate to organizational level outcomes. To date, however, this assumption has been largely unsubstantiated. Moreover, the theoretical understanding of the link between positive individual outcomes and positive organizational or strategic outcomes is neither clearly understood nor theorized (Bolden, 2016; Clarke, 2012).

In this paper, we posit that organizational identification and organizational commitment might be important intervening mechanisms that link MLDPs with strategic outcomes. Strategically oriented MLDPs deliver messaging intended to redirect managers' role definition and priorities toward strategic success. As managers refine their role definition, they may increase their identification with the organization (Peters et al., 2013) and begin to see their own success as intertwined with the organization's success (Haslam, 2004; Hogg et al., 2017). Thus, if strategically oriented MLDPs promote greater organizational identification, managers might begin to devote their energy toward meeting strategic goals. Conversely, if the MLDP fails to promote identification and commitment, and managers do not engage in strategic action, it may explain why MLDPs have been criticized for not producing the desired organization level effects (Bolden, 2016; Ardichivili et al., 2016; Gurdjian et al., 2014).

In addition to examining the direct effects of MLD on managers' identification and commitment, we follow the previous literature by exploring the indirect effect of MLDP on subordinates (Barling et al., 1996; Dvir et al., 2002). Specifically, manager's skill acquisition during MLDPs has been shown to have cascading effects, spreading to subordinates (Avolio et al., 2009). Yet, it is unclear whether these cascading effects occur beyond individual skill acquisition, or whether a strategically oriented MLDP may indirectly influence subordinates' organizational identification and commitment. If cascading effects follow previous research, we would expect subordinates identification and commitment to increase when they see their managers become more identified and committed to the organization (Barling et al., 1996; Dvir et al., 2002). However, the complex nature of the organization and the multitude of potential identity targets (Ramarajan, 2014), make it possible that cascading effects work in the opposite direction (Peters et al., 2013); decreasing identification and negatively impacting the manager's ability to lead. Given that cascading effects of identification and commitment have not been studied, we examine whether subordinates are impacted by their managers' engagement with a strategically oriented MLDP.

Therefore, the purpose of this study is to investigate the effects of a strategically oriented MLD program on managers' and subordinates' organizational identification and commitment in the face of organizational complexity (see Figure 1). We examine these mechanisms in a complex health administration organization with multiple subsystems, occupations and subcultures, and investigate whether changes in manager identification and commitment have a “cascading” or “diffused” effect on subordinates. Given that organizational identification and commitment have not been studied as outcomes of MLDPs, we present theoretical justification for why strategically oriented MLDPs may impact role definitions, increasing organizational identification and commitment, and thus leading to more strategically oriented thinking and action. However, our findings reveal that although participants were satisfied with the program overall, organizational identification and commitment actually decreased. Our study reveals that these mechanisms may be more complex than originally thought, and that attempts to modify role definition, and identification with the organization may have instead activated identification with other targets (i.e. the profession or manager cohort). Our findings reveal a story of mystery, suggesting that strategically oriented MLDPs may affect many different identification and commitment targets in complex ways.

Theoretical background and hypothesis development

The early literature in social psychology differentiated between one's individual identity (i.e. defining oneself according to personal attributes) and one's social identity (i.e. defining oneself by group membership; Tajfel and Turner, 1979). Applied to organizations, social identification with the organization occurs when one perceives him or herself as intertwined with the fate of the organization and experiences the organization's successes and failures as his or her own successes and failures. Tajfel and Turner (1979) contend that individuals can hold multiple social identities at any given time. For example, an employee may socially identify with their sub-unit, organization, profession or all of the above (Cain et al., 2019; Ramarajan, 2014; Mitchell and Boyle, 2015). Each of these social identities holds different representations for the employee, and identity with any one of these groups may be stronger than others (Hogg and Terry, 2000; Meyer et al., 2006).

Research suggests that employees typically have a stronger identity with their immediate (proximal) work group as opposed to the more distant (distal) group or organization (Becker et al., 1996; Riketta and Van Dick, 2005). This social identification with the work group helps employees distinguish between in-group from out-group members, where in-group members are favored for their unique characteristics and competencies (Haslam, 2004). Research also suggests strong organizational identification may unite individuals across professional divides; however, promoting organizational identification is only successful when it seeks to compliment, rather than replace, professional identities (Cain et al., 2019; Dovidio et al., 2007).

MLDPs that focus on strategic alignment aim to shift a manager's role definition, thus making the organization the salient identification target (Peters et al., 2013). It does this by emphasizing the part of the managers' role that supports strategic initiatives, and realigning a manager's role definition to emphasize organizational and strategic outcomes (McDonald et al., 2018; Peters et al., 2013). Although not intended to replace work-unit identification, strategically oriented MLDPs seek to position middle-managers as strategic linking-pins to promote strategic alignment throughout the organization. If curriculum and messaging about what managers should do to enhance strategic success becomes internalized (Peters et al., 2013) the manager's organizational identification should be activated and strengthened. On the other hand, previous research notes the risks associated with making one identification target more salient than others (Cain et al., 2019; Peters et al., 2013). Specifically, identification may decrease if the strategically oriented MLDP produces conflicts between organization identification and managers' existing identification with their work unit, profession or position cohort. However, given the aim of the program was precisely to enable strategic action, we hypothesize that organizational identification should increase.

H1.

Managers' organizational identification will increase after completing the MLD program.

A similar but distinct concept is organizational commitment: an affective feeling that binds an individual to a collective, group or course of action (Meyer et al., 2006). Whereas identification refers to feelings of oneness with a social entity, commitment relates to one's drive and behavioral intentions to achieve relevant goals on behalf of a target that is psychologically distinct from the self. Although the concepts are related, empirical research has found that they are quite distinct in terms of the way they develop, their volatility, and their behavioral implications (Meyer et al., 2006; Van Knippenberg and Sleebos, 2006). For example, organizational commitment may exist in the absence of identification (Meyer et al., 2006), and has been found to account for unique variance in both job satisfaction and turnover intentions (Van Knippenberg and Sleebos, 2006).

Research suggests that HR practices such as training and development may also create changes in managers' organizational commitment (Klein et al., 2012). For example, a strategically oriented MLD program that increases one's alignment with strategic objectives may also increase the manager's behavioral intentions and drive to advance strategic goals.

Managers who accept and internalize program messaging regarding the need to redefine their role as strategic linking-pins may become more committed to organizational goals. Indeed, research suggests that increasing role clarity and goal salience has a positive influence on commitment (Klein et al., 2012). Conversely, if managers fail to internalize their new role definition, or perceive that strategic action is not possible, organizational commitment may not increase. However, given that the program was designed to emphasize the managers' role in pursuing strategic goals (Klein et al., 2012), we expect that that these managers will become more committed to the organization after completing the program.

H2.

Managers' organizational commitment will increase after completing the MLD program.

In addition to examining the effects of MLD on participating managers, we are also interested in whether the MLD has a cascading effect on the manager's subordinates. A number of researchers have suggested that leadership development has a diffused or cascading effect whereby the leader who improves upon her skills or competencies affects those below her, who in turn affect others in the organization (Avolio et al., 2009; Bass et al., 1987; Berson and Avolio, 2004). While this effect has been examined for programs focusing on individual skill acquisition, there are several reasons to believe that it may be different for a strategically oriented MLDP delivered in a complex organizational environment.

MLDP messaging about strategic orientation, priorities and goals may not diffuse the same way as individual skills or competencies. In complex organizations, there are often deep divides between the strategic and operational levels of the organization that may serve as barriers to cascading messages about strategic orientation (March and Simon, 1958; Mintzberg, 1983). Whereas the strategic (top) levels of management are focused on far-reaching, often financial outcomes, the operational levels are more dedicated to customer care and specific service delivery. As a result, subordinates at the operational levels may see themselves as a distinct and unique subsystem or subculture that, at best, is loosely coupled with the strategic level (Orton and Weick, 1990).

The divide between upper management and lower level employees has important implications for social identification with the organization. Sub-units of subordinates who have a strong social identification with the organization are expected to feel united with other sub-units under a common vision (Cain et al., 2019). However, in complex organizations, subordinates may not demonstrate high levels of identification with the organization, thus magnifying divisions and differences between hierarchical levels (Dovidio et al., 2007). As such, subordinates may begin to see the strategic level of the organization as the out-group, thus emphasizing the differences, rather than similarities, between the two groups (Haslam, 2004). This is particularly problematic when the subordinates' manager, who they once may have identified as an insider and “one of them” (Ellemers et al., 2004; Haslam, 2004), begins to espouse the values and strategic imperative of the organization. Thus, despite the organization's intention to increase identification and commitment throughout hierarchical levels, there are theoretical reasons for expecting different effects. Subordinates who see their leader as more aligned with organizational strategies and objectives may instead perceive that the leader no longer represents them and instead represents the “out-group” (Ellemers et al., 2004). Thus, Branscombe and Ellemers (1998) posit that employees who highly identify with their work unit may perceive leader disloyalty as a threat and begin to dis-identify with the leader.

Subordinates who dis-identify with the leader may also experience decreased identification and commitment to the organization (Ashforth et al., 2008; Cain et al., 2019; Klein et al., 2012). Specifically, subordinates rely on their leader to support and represent their work unit and operational goals. If subordinates perceive a decline in leader support, they may begin to feel even more disconnected from the organization as they continue to focus on operational excellence without perceived leader support. If they feel as though their ability to pursue work unit goals has been challenged by the leader and the organization (i.e. out-group, Ellemers et al., 2004), the salience of organization goals may be weakened as their focus on operational goals becomes more central (Klein et al., 2012). Thus, although in direct contrast to the organization's intentions, we find theoretical reasons to hypothesize:

H3a.

Subordinates' identification with the leader will decrease after completion of the MLD program.

H3b.

Subordinates identification with the organization will decrease after completion of the MLD program.

H3c.

Subordinates organizational commitment will decrease after completion of the MLD program.

Dis-identification with the organization and the leader may impact how subordinates view their work unit and how much they identify with that social group. Specifically, we posit that the leader aligning with the organization will be perceived as a source of uncertainty by the group, or a shift in the context that threatens the group's stability (Branscombe and Ellemers, 1998). This sort of instability has been shown to make associations within the group even stronger as group members turn inward for support and constancy. In essence, the threat creates an in-group bias effect (Mullen et al., 1992), where similar others are elevated in status and out-group members degraded more harshly (Wann and Grieve, 2005)

H4.

Subordinates' identification with the work unit will increase after completion of the MLD program.

Finally, in addition to anticipating significant differences from time 1 to time 2, we further hypothesize that changes in managers' identification and commitment will predict changes in subordinates' identification and commitment. Specifically, as managers increase their identification with and commitment to the organization, we anticipate that subordinate identification with the leader and organization, and commitment to the organization will decrease, and identification with the work unit will increase.

H5.

Manager changes in organizational identification and commitment will predict subordinate changes in (a) leader identification, (b) organizational commitment, (c) organizational identification, and (d) work unit identification after completion of the MLD program.

Method

Program and participants

Middle managers employed in a regional health administration in Denmark completed a six-month MLDP. Developed in collaboration with an association of private consulting firms and the regional healthcare administration in Denmark, the MLDP comprises 14.5 days divided into six modules spread out over the course of about six months.

According to program documentation, the purpose of the leadership development program was to help middle-managers better connect to the goals and strategy of the organization. In order to meet this objective, managers were presented with content on the steering instruments of the organization, as well as content to improve their inter- and intra-personal skills and communication. Throughout the program, managers were presented with messages regarding the importance of connecting with and realizing strategic efforts of the organization.

Prior to collecting survey data to test our stated hypotheses, we conducted a short pilot study to ensure that managers were aware of the program objectives regarding strategic orientation. Two of the authors collected qualitative, open-ended interview data with approximately 22 participants enrolled in the first wave of the program. Participants were asked “What was your understanding of the purpose of the program?” and “How did the content of the program resonate with your current work role?”, and responses were recorded and transcribed. Interview data were then coded for responses that reflected participants' understanding and/or acceptance of the program message regarding strategic orientation. Selected quotes from program participants are included below.

Quantitative survey data were collected across three waves of participants enrolled in the MLDP. In each wave of the program, all managers enrolled were asked to participate in the study (N = 183).The time 1 survey was administered during the first week of the MLDP and included scales measuring the manager's organizational identification and commitment. The time 2 survey was administered two months after the conclusion of the program and collected data on the same identification and commitment measures administered in time 1.

Participating managers were asked to identify a team of subordinates who reported directly to them and to provide names and email addresses. All identified subordinates (N = 2,342) were sent an introductory email explaining the study and asking for voluntary participation. The subordinate time 1 survey was administered one month after the program began and measured subordinates identification with the organization, the leader, and the work group, as well as commitment to both the work group and the organization. The time 2 survey for subordinates was administered at the completion of the program and collected the same identification and commitment measures administered in time 1.

Demographic data on both managers and subordinates can be found in Table I. The manager time 1 survey yielded 150 complete responses, however, this number dropped in time 2 to 110. After matching time 1 and time 2 data, our final sample for the study included 107 managers (58.4% response rate). The subordinate time 1 survey yielded 1,303 responses, however, only 935 subordinates responded at time 2. Again, after matching the time 1 and time 2 surveys, the final subordinate sample included 913 complete responses (38.9% response rate).

Several precautions were taken both before and after data collection to prevent common method bias in our results. First, we followed the suggestions by Podsakoff et al. (2003) and designed our survey such that items on the same scales were separated and counterbalanced. In addition, we used Harman's one-factor test (Chang et al., 2010) to assess whether a single common factor (associated with common method variance) accounted for a large proportion of variance in our data. Finally, although we did not have data from individuals who did not complete the surveys, we did test for significant differences between those who only completed survey 1 and participants who completed both surveys on variables such as age, gender, organizational tenure, organizational and work unit identification and commitment. No significant results were found in either the single-factor tests or the tests between participant groups. Thus, we were confident that common method bias was not a contributing factor to the variance in our data.

Measures

All scales included in the surveys utilized a five-point response scale ranging from 1 (strongly disagree) to 5 (strongly agree). Manager surveys included measures of organizational identity and organizational commitment. Subordinate surveys included identical measures of organizational identification and commitment, as well as work unit identity and leader identification.

Organizational identification was measured with Kark et al.'s (2003) six-item social identification scale (manager time 1 α = 0.76; manager time 2 α = 0.80). A sample item from this scale is “When someone criticizes this organization, it feels like a personal insult.” Work unit identification was also measured with the same six-item scale with a sight modification of each item to target identification with the “work unit” instead of the “organization” (subordinate time 1 α = 0.80; subordinate time 2 α = 0.85).

Subordinate identification with the leader was also measured with Kark et al. 's (2003) ten-item scale (time 1 α = 0.92; time 2 α = 0.93). This scale uses similar items with a modification that replaces the word “organization” with “leader”. A sample item from this scale is “The values of my leader are similar to my values.”

Organizational commitment was measured with Schoemmel et al. (2015) four-item multi-target commitment scale (manager time 1 α = 0.77; manager time 2 α = 0.80; subordinate time 1 α = 0.79; subordinate time 2 α = 0.80). A sample item from this scale is “I really care about this organization.” Work unit commitment was also measured with the same four-item scale with a sight modification of each item to target commitment to the “work unit” instead of the “organization” (manager time 1 α = 0.77; manager time 2 α = 0.81; subordinate time 1 α = 0.84; subordinate time 2 α = 0.87).

Analytic strategy

Prior to running our analyses, we assessed both reliability and agreement within subordinate work units to ensure that variables could be aggregated to the group level. We assessed the internal consistency of subordinate responses with intra-class correlations (ICCs) by selecting a one-way random effects model. Results demonstrate that time one responses demonstrated adequate levels of consistency for leader identification (ICC1 = 0.22; ICC2 = 0.74), work unit identification (ICC1 = 0.13; ICC2 = 0.51), and organizational commitment (ICC1 = 0.12; ICC2 = 0.55). Similarly, we obtained time 2 values for leader identification (ICC1 = 0.17; ICC2 = 0.64), work unit identification (ICC1 = 0.08; ICC2 = 0.30), and organizational commitment (ICC1 = 0.05; ICC2 = 0.17). The ICC values for subordinate organizational identification, however, were not acceptable (Time 1 ICC1 = 0.02; ICC2 = 0.09); (Time 2 ICC1 = 0.02; ICC2 = 0.07). Given that several of our ICC values (i.e. organizational commitment) fell outside of normal range as suggested by by Bliese (2000) and LeBreton and Senter (2008), we also assessed within-team agreement for the three variables with acceptable ICCs by computing the r_wg(j) index (James et al., 1993). The average r_wg(j) values were all above the advised cut-off of 0.70 used to indicate adequate levels of agreement. Specifically, time 1 values of leader identification, work unit identification, organizational commitment were 0.80, 0.84 and 0.81 respectively. Results from the time 2 data were equally promising, demonstrating r_wg(j) values of 0.77, 0.79 and 0.81, respectively. Thus, a decision was made to aggregate all subordinate rated variables to the group level with the exception of organizational identification.

Given that our data included managers and subordinates from multiple departments, locations and positions within the organization, we ran analyses to ensure that our data were not impacted by these contextual influences. We found no significant differences among the different hospital locations or training cohorts. However, we did find significant differences between work units on the variable of aggregated subordinate organizational commitment. As a result, we decided to include work unit as a control variable in our regression analyses.

We tested our hypotheses following a two-stage analysis strategy. In stage one, we conducted repeated measures t-tests on the two leader variables as well as the three aggregated subordinate variables to assess changes in scores from time 1 to time 2. In stage two, we conduced hierarchical linear regression to assess whether changes in the leader's identification and commitment predicted changes in aggregated subordinate identification and commitment. In step one of the regression analysis, we entered a number of control variables including the subordinate's work unit, hospital location and tenure with the supervisor. In step one, we also entered the time 1 value of the aggregated subordinate variable to control for baseline values and allow us to explain only the variance associated with change in the dependent variable. In step two, we entered the difference scores for managers' organizational identification and organizational commitment.

Results

Pilot study

The purpose of the pilot study was to understand whether managers enrolled in the program were (1) understanding and (2) buying into the organization's messaging about the need for better strategic orientation and integration. With regard to the first objective, responses revealed that managers indeed understood that at least part of the training was to help them become more integrated in strategy creation and strategic decision-making. For example, one respondent said:

In relation to organizational strategy, it is a very smart way to get us all on the same track, so that overall, [we] more or less walk in the same direction.

Another respondent echoed this by stating:

In the beginning, we started with this organizational strategy, and talked about the meaning of the work.

With regard to the second objective, responses were mixed in the extent to which the managers accepted and integrated the messaging about becoming better strategic actors. For example, one respondent put it this way:

With the strategic leadership and leading upwards, that…is a part that often gets forgotten because there is not time or space for it in everyday work…so I think it was very good to have it in the program and to give it more focus and be more conscious about it.

Similarly, another respondent mentioned the challenges associated with thinking and acting on a more strategic level:

Strategic management and managing upwards, I know that, but… That is the part which I think tends to fade out in the everyday work which becomes much more about fire extinguishing and operational tasks.

Taken together, the pilot interviews suggest that participants did understand the MLD messaging regarding strategic orientation and alignment, yet they commented on the challenges associated with having the time and resources to think and act strategically in the face of more immediate operational demands.

Finally, participants were asked about the most important or effectual elements of the MLDP. Responses to this question seemed to indicate that the interaction and communication they had with peers was the most valuable part of the program, rather than the emphasis on strategic alignment:

I believe that the discussions we have had, both in the whole group and in the small groups, where you talk about each other's everyday work…I believe it is those [conversations] that have kind of shifted things for me; done things for me; changed something in me.

The qualitative data present only a small sample of results from the MLDP. Although the results are promising in that they provide evidence that participants were accurately receiving the message regarding the need for better strategic orientation, it does not provide clear evidence that this messaging was successful in changing managers' role definitions from operational to strategic. Thus to gain a better picture of manager and subordinate outcomes, we proceeded with hypothesis testing using quantitative survey data.

Hypothesis testing

Means, standard deviations and inter-correlations for all study variables are included in Table 2. Hypothesis 1 stated that there would be a significant increase in manager organizational identification after completion of the MLD program. The results suggest that although there was a significant change in managers' organizational identification, it was in the opposite direction than anticipated (see Table 3). Specifically, organizational identification was significantly higher prior to the MLDP (M = 4.58, SD = 0.62) than after (M = 3.52, SD = 0.56); t (106) = 17.63, p < 0.001.

Hypothesis 2 stated that there would be a significant increase in manager organizational commitment after completion of the MLDP. Again, our results show significant changes in the opposite direction with organizational commitment significantly higher prior to the MLD (M = 4.82, SD = 0.62) than after (M = 3.75, SD = 0.55); t (106) = 20.42, p < 0.001.

Hypothesis 3 stated that subordinate identification with the (a) leader and (c) commitment to the organization would decrease after completion of the MLDP. This hypothesis was fully supported with aggregated subordinate ratings on leader identification falling from time 1 (M = 3.06, SD = 0.46) to time 2 (M = 2.97, SD = 0.56); t (131) = 2.64, p < 0.01. Similar effects were found with aggregated commitment to the organization decreasing from time 1 (M = 3.52, SD = 0.31) to time 2 (M = 3.45, SD = 0.38); t (131) = 2.31, p < 0.05 Low ICC(1) values did not warrant the aggregation of subordinate organizational identification, thus hypothesis 3 (b) could not be tested.

Hypothesis 4 stated that subordinate identification with the work group would increase after completion of the MLDP. This hypothesis was not supported as there were no significant differences in aggregated work unit identification scores from time 1 to time 2 (see Table 3).

Hypothesis 5 stated that changes in subordinate ratings of (a) leader identification, (b) organizational commitment, (c) organizational identification and (d) identification with the work unit would be predicted by changes in their manager's organizational identification and commitment. As mentioned earlier, low ICC(1) values did not warrant aggregation of subordinate organizational identification, thus we were unable to test hypothesis 5c. However, after entering all control variables to account for extraneous variance, our results demonstrate significant effects for Hypotheses 5a and 5b, but no evidence was found for hypothesis 5d (see Table 4).

With regard to leader identification (5a), our results suggest that only changes in manager organizational commitment significantly predicted changes in subordinates' identification with the leader (β = −0.24, t = −2.05, p < 0.05). Similarly, when predicting changes in subordinates' organizational commitment (5b), the only significant predictor was changes in manger's organizational commitment (β = −0.31, t = −3.03, p < 0.01). Thus, hypothesis 5a and hypothesis 5b were partially supported.

Discussion

Previous research suggests MLDPs that promote middle-management strategic orientation are not only more common, but essential for strategic success (McGurk, 2010). Yet, there is a general lack of theory or empirical evidence for the notion that MLDPs affect strategic and organizational outcomes (Bolden, 2016; Ardichvili et al., 2016; Fernandez-Aráoz et al., 2017; Gurdjian et al., 2014), as much of the MLD literature investigates individual level gains in skills and competencies. Thus, our study examines the effects of a MLDP attempting to redefine middle managers' roles from operational actors to strategic creators, motivating them to act in the organizations best interests. We hypothesized that if the strategically oriented MLDP was successful, and organizational identification and commitment were strengthened in this process, that managers may direct their effort toward strategic action.

Our results provide a mysterious and complex picture of the role organizational identification and commitment play as mechanisms related to strategic orientation. Although social identification theory would suggest that making strategic priorities salient to managers role definition may increase organizational identification, and motivate them to act in ways that promote strategic goals (Haslam, 2004; Hogg et al., 2017; McGurk, 2010), our results suggest that the process is much more complex in nature. In this particular study, managers did understand the messaging about integrating strategic orientation into their role and saw the MLDP as an overall success. However, the finding that identification and commitment decreased suggests that there are other mechanisms at work, and that organizational perceptions and attitudes are not linearly linked to MLDP curriculum the way that skill and competency acquisition are (Collins and Holton, 2004; Day and Dragoni, 2015).

Given that MLD research has not studied identification and commitment as outcomes of strategically oriented MLDPs, it is tempting to conclude from our findings that this program was a failure. However, according to participant evaluations and anecdotal data, managers expressed overall satisfaction with the program, as well as the attempt to bring strategic issues into clearer focus. This finding of failure to achieve strategic outcomes while satisfying individual participant expectations is in line with previous research. Specifically, the MLD literature suggests that although MLDPs may be successful at enhancing individual skills and capacities, there is little evidence that MLDPs positively impact strategic or organizational outcomes (Ardichvili et al., 2016; Bolden, 2016; Fernandez-Aráoz et al., 2017; Gurdjian et al., 2014). Our results expand on this finding by suggesting that simply exposing managers to strategic messages and clarifying their role does not directly result in stronger organizational identification or commitment. Although our data preclude us from drawing firm conclusions regarding why identification and commitment decreased, the qualitative pilot data offer several probable suggestions that should be investigated in future research.

First, our qualitative data suggest that although the MLDP attempted to make managers' organizational identification more salient by emphasizing their role in advancing strategic initiatives, managers themselves stated that the most valuable part of the training was the interaction with other managers (i.e. their professional community), and the validation they received from hearing the experiences and stories of others. The social environment thus offered ample possibilities for identity confirmation (Swann, 1987) – when one's personal identity or self-concept as a leader is verified. As managers shared personal experiences with the group, it fostered a stronger “leader” identity and positioned the peer group of other mangers as the most salient feature of the MLD program (Meyer et al., 2006; Smith and Woodworth, 2012). The identification literature notes that attempts to increase organizational identification often fail to consider the diverse perspectives of employees and the individual needs of managers (Haslam, 2004; Tyler and Blader, 2003), and that promoting organizational identification is only successful when it compliments, rather than replaces, professional and work unit identification (Cain et al., 2019; Dovidio et al., 2007). Thus, it is possible that these particular managers needed confirmation and strengthening of their “middle-manager identity” more than their organizational identity.

A second possibility, as noted in our qualitative data, is that managers experienced conflicts between their identification with the immediate work unit and the organization. Indeed, our qualitative data provide some evidence that although managers understood the organization's message regarding strategic orientation, immediate work unit needs often overshadowed their ability to think and act more strategically. In this case, attempts to decouple work unit identification, in favor of organizational identification, were not readily internalized by participating managers (Cain et al., 2019). This finding aligns with previous research suggesting that existing identities (tied to professions and organizational units) might represent significant barriers to programs aimed at increasing organizational identification (Peters et al., 2013).

Although we are unable to conclude exactly why identification decreased following the MLD program, it is clear that the program's strategic messaging about clarifying managers' roles was not sufficient to establish stronger organizational identification and commitment. Future research could investigate identification with the peer group or identity conflicts between the work unit and the organization as outcomes of strategically oriented MLDPs. Furthermore, researchers could investigate how managers' respond to strategic messaging in real time, gathering feedback and reactions to the message as it is disseminated. Researchers might also investigate the longitudinal effects that such messaging has on managers' identification with the work unit and organization. Specifically, it is possible that managers understand and accept the message as it is delivered, but disassociate with the message once they are back in their operational units.

The findings of this study also show that subordinate identification with the leader, and commitment to the organization, was lower after the completion of the manager's MLDP, and at least some of these changes were predicted by changes in managers' organizational commitment. Although we hypothesized that subordinate identification and commitment would decrease after the completion of the MLDP, we did so under the assumption that managers' organizational identification and commitment would increase. Thus, the results regarding subordinate changes in identification and commitment also present a bit of a mystery to which there are several possible explanations.

First, given the complex nature of the organization, it is possible that the simple act of managers enrolling in the program and engaging with the curriculum on strategic orientation was enough to make subordinates dis-identify with their manager. We hypothesized that subordinates who saw their managers becoming more strategically aligned would distance themselves from the leader and the organization because of a perceived threat from the out-group (Haslam, 2004). It is possible that this perception persisted even without specific attitudinal or behavioral change on the part of the manager. Perhaps the manager's appreciation of the strategic nature of the MLDP, even if they did not fully integrate it into their role, was enough for subordinates to perceive the manager aligning with the out group and disassociating with the in-group (Branscomb and Ellemers, 1998; Ellemers et al., 2004). Given our finding that subordinate identification and commitment decreased along with their managers, it is clear that additional research is needed to unpack subordinate perceptions of the strategically oriented MLDP.

Second, dis-identification from the leader may have also contributed to subordinates' reduced commitment to the organization. The previous literature suggests that subordinates who are closely tied to their work unit may see the larger organizational entity as the out-group (Ellemers et al., 2004). If subordinates perceived the MLDP as a threat to in-group functioning, they may have responded by reducing commitment to the organization in favor of the work unit (Klien et al., 2012). Thus, it appears that in this particular case, the diffused or cascading effect of leadership development may have had counterproductive effects on the organization. Although previous research has demonstrated a positive cascading effect for individual skills and competencies (Barling et al., 1996; Dvir et al., 2002), to our knowledge, ours is the first study to demonstrate that this effect may also be negative. If subordinates are adversely affected by their managers engaging in MLDPs, it could have a negative effect on the managers' ability to lead. In essence, the MLDP has reduced their ability to influence as subordinates become dis-identified with their manager. This is an important finding and one that should be examined further by future research.

Implications for theory

Previous theorizing suggests that MLDPs focused on increasing strategic orientation should indeed influence both organizational identification and commitment (McGurk, 2010). Specifically, as managers receive messaging about the importance of increased strategic thinking and action, they should begin to redefine their roles from operational to strategic (McDonald et al., 2018; Peters et al., 2013). Yet, this theory assumes that managers' organizational identification would be more salient than identification with their profession or middle-management cohort (Cain et al., 2019; Ramarajan, 2014). Moreover, it assumes that program participants have the capacity to think and act more strategically in the face of more immediate operational demands. Our results suggest that achieving strategic orientation, and thus increasing organizational identification and commitment, may be more challenging than originally thought. Managers may simply not have the time, space or capability to increase their strategic focus, thus making internalization of the message and role-shifting less likely (Peters et al., 2013). Furthermore, making multiple identities salient (i.e. organization, work unit, professional, etc.) might create the context for managers to choose which identity resonates most strongly with their current needs and desires (e.g. professional identity over organizational identity). Future research should investigate different types of MLDPs to understand when strategic messaging is or is not accepted (Carroll and Nicholson, 2014; Gagnon and Collinson, 2017). It is possible that the complex nature of the organization under study made the messaging less impactful or less relevant to managers who needed professional validation rather than strategic alignment. Thus, future research might study strategically oriented MLDPs in a variety of settings to better understand the types of contexts where such messaging is accepted. This research might also examine how long these effects last, and whether programs have a short or long term effect on one's social identification.

Our study also has implications for theory on the cascading effect, as well as dis-identification as a result of MLDPs. Although the cascading effect has primarily been theorized and studied among programs emphasizing individual skills and competencies (Avolio et al., 2009; Barling et al., 1996; Dvir et al., 2002), as mentioned above, there is reason to believe that strategically oriented messaging may produce different effects (Peters et al., 2013). Future research might study how subordinates react to their managers enrolling in, and being subjected to, MLDPs that emphasize the strategic part of their job. It is possible that subordinates see this as a threat, whereas individual skill and competency acquisition is seen as an opportunity for greater performance. If subordinates see manager participation in a strategically oriented MLDP as a form of “disloyalty” to the work unit (Branscombe and Ellemers, 1998), the organization could be inadvertently creating deeper divides within its sub-units and systems. Future research might collect qualitative data from subordinates regarding their reaction to the organization's emphasis on strategy over operations to better understand how subordinates perceive strategically oriented MLDPs.

Implications for practice

Our research also has important implications for practitioners who design and deliver these programs. As mentioned previously, practitioners should be cognizant of the messages they are sending to middle managers, and ensure that participants have the capacity to integrate strategic thinking into their role. For example, designing curriculum on how managers might integrate one strategic goal or objective, as a starting point, and helping them understand the time and resource requirements involved in perusing this goal, may make the role-redefinition more successful. Likewise, providing specific content on how to link strategic and operational goals, and creating action plans to realize them, may make the task of strategic alignment more realistic to managers balancing multiple demands. Programs that fail to emphasize this balance run the risk of creating role confusion, and inciting managers to reject the strategic orientation being advocated.

The study demonstrates the potential risk associated with development programs that separate managers and employees at different organizational levels, rather than integrating levels in the same training. To avoid this, designers of MLDPs aiming to support strategy implementation might also include managers from different hierarchical levels in the organization in an attempt to better diffuse strategic goals. Allowing multiple levels of management to work together, perhaps in the form of action learning projects, may also spread identification and commitment more broadly. Programs would thus move from focus on individual managers to strategically important social networks, that is, from leader to leadership development (Day, 2001; Day and Harrison, 2007).

Limitations

Although our study had a number of strengths, it also has several weaknesses. First, the lack of a control group in our research makes it impossible to conclude that the changes we observed in manager and subordinate ratings were solely due to the MLDP. Second, we only conducted surveys at two time periods and thus do not have data on changes that might have been occurring as the program was taking place. Future research may want to employ an experimental design in order to rule out extraneous circumstances or variables.

Our study was conducted in a single organization, that was complex in nature, and our results may not be generalizable to other types of organizations in different industries. Moreover, our study was conducted with a specific population of mid-level managers, and thus we might expect different results with different populations. Future research may want to expand the types of organizations and populations under investigation in order to see whether different populations show different effects.

Finally, our unanticipated results clearly call for more mixed methodologies that could help researchers better explain the nature of these findings. Although we presented some qualitative quotes, these were collected from participants who completed this MLD program prior to our study commencing. Having qualitative data from participants would have assisted in answering questions regarding the decline in organizational identification, as well as why managers' change in identification affected subordinate commitment and leader identification.

Conclusion

Complex and fragmented organizations need to develop managers as “strategic linking-pins” to integrate and diffuse strategic objectives throughout the organization. This study takes a fresh perspective on strategically oriented MLDPs by focusing on organizational identification and commitment, rather than individual skills of the leader, as important program outcomes. In doing so, we reveal an intricate pattern of identification and commitment that is perhaps more complex than originally thought. Instead of finding positive effects of leadership development on identification and commitment, our results indicate that strategically oriented MLDPs can have unintended effects. In this case, the very program intended to make middle managers better “leaders” for the organization, has resulted in lower levels of organizational identification and commitment among both leaders and subordinates. Assuming that organizational identification is the mechanism that links middle managers role definition and strategic action, this means that the program may have inadvertently damaged both their ability to engage in successful leadership, and their focus on implementing organizational strategy.

Figures

Figure 1

Conceptual model of MLD strategic orientation effects on managers and subordinates

Table 1

Demographic characteristics of managers and subordinates

	Frequency	Percentage
Manager gender
Male	37	24
Female	115	76
Manager age
30–34	9	6
35–39	24	16
40–44	33	22
45–49	29	19
50–54	33	22
<55	23	15
Manager organizational tenure
1–1 year	22	15
2–3 years	33	22
4–5 years	21	14
6–7 years	16	11
8–9 years	12	8
>10 years	48	32
Manager tenure with work unit
1–1 year	59	31
2–3 years	56	29
4–5 years	12	6
6–7 years	8	4
8–9 years	5	3
>10 years	12	6
Number of direct reports
>10	38	24
11–20	62	38
21–30	34	21
31–40	17	11
41–50	5	3
<50	5	3
Manager organizational work unit
Administration	16	9
Hospital Medical	101	55
Hospital Other	8	4
Laboratory	6	3
Medical (non-hospital)	6	3
Other	12	7
Paramedical	4	2
Psychiatry	5	3
Technical	24	13
Hospital location
Location 1	12	7
Location 2	20	11
Location 3	6	3
Location 4	25	14
Location 5	24	13
Location 6	5	3
Location 7	27	15
Location 8	38	21
Location 9	26	14
Subordinate gender
Male	208	16
Female	1,091	84
Subordinate age
18–29	171	14
30–39	279	22
40–49	340	27
50–59	348	27
<60	131	10
Subordinate organizational tenure
1–1 year	314	24
2–3 years	256	20
4–5 years	121	9
6–7 years	144	11
8–9 years	107	8
>10 years	357	28
Subordinate tenure with supervisor
1–1 year	689	53
2–3 years	416	32
4–5 years	69	5
6–7 years	49	4
8–9 years	26	2
>10 years	49	4

Table 2

Means, standard deviations and inter-correlations of study variables

	1	2	3	4	5	6	7	8	9	10	11	12
1. MGR Org. ID T1	0.76
2. MGR Org. ID T2	0.45	0.80
3. MGR Org. CMT T1	0.76	0.49	0.77
4. MGR Org CMT T2	0.40	0.75	0.58	0.80
5. Subord. Leader ID T1	0.02	−0.15	0.08	−0.10	–
6. Subord. Leader ID T2	−0.05	−0.08	0.01	−0.16	0.68	–
7. Subord. Org. CMT T1	−0.04	−0.11	0.09	−0.04	0.51	0.36	–
8. Subord. Org. CMT T2	0.11	−0.12	0.22	0.03	0.27	0.40	0.46	–
9. Subord. Unit ID T1	0.09	0.04	0.15	0.17	0.60	0.53	0.60	0.42	–
10. Subord. Unit ID T2	−0.02	0.10	0.03	0.03	0.39	0.58	0.38	0.62	0.59	–
Mean	4.53	3.51	4.77	3.75	3.06	2.96	3.54	3.45	3.67	3.68	3.90	3.8
S.D.	0.59	0.58	0.60	0.56	0.46	0.57	0.38	0.39	0.33	0.40	0.30	0.37

Note(s): N = 107. S.D. = standard deviation; MGR = Manager; ID = Identification; CMT = Commitment; Subord = Subordinate. Gender was coded as 0=Female, 1=Male. All correlation coefficients above 0.18 are statistically significant. Two tailed test

Cronbach's alphas appear in italics

Table 3

Results from paired samples T-Tests

Manager measures	Pre-test M (SD)	Post-test M (SD)	t	df	p
Organizational identification	4.58 (0.62)	3.52 (0.56)	17.63	107	0.001
Organizational commitment	4.82 (0.62)	3.75 (0.55)	20.42	107	0.001
Subordinate measures
Leader identification	3.06 (0.46)	2.97 (0.56)	2.64	131	0.009
Work unit identification	3.67 (0.29)	3.68 (0.39)	−0.398	131	0.69
Organizational commitment	3.52 (0.31)	3.45 (0.38)	2.31	131	0.02

Table 4

Regression results predicting change in subordinates' identification and commitment

Model predicting leader identification T2	b	se	β	t	p
Step 1
Location 1	−0.06	0.29	−0.03	−0.19	0.85
Location 2	0.03	0.28	0.02	0.11	0.91
Location 3	0.12	0.39	0.03	0.31	0.76
Location 4	0.05	0.27	0.03	0.18	0.86
Location 5	−0.12	0.27	−0.07	−0.43	0.67
Location 6	1.30	0.68	0.25	1.93	0.06
Location 7	−0.22	0.31	−0.14	−0.72	0.45
Location 8	0.08	0.28	0.05	0.28	0.78
Work unit 1	0.12	0.21	0.06	0.59	0.56
Work unit 2	−0.37	0.18	−0.18	−2.02	0.05
Work unit 3	−0.20	0.29	−0.08	−0.71	0.48
Work unit 4	−1.10	0.47	−0.29	−2.34	0.02
Work unit 5	−0.09	0.28	−0.03	−0.33	0.74
Work unit 6	0.04	0.24	0.02	0.18	0.86
Work unit 7	0.01	0.34	0.01	0.03	0.98
Work unit 8	0.103	0.29	0.06	0.35	0.73
MGR tenure with work unit	−0.04	0.13	−0.12	−1.27	0.21
Aggregated leader identification T1	0.89	0.11	0.65	7.70	0.01
Step 2
MGR organizational commitment	−0.25	0.12	−0.25	−2.05	0.04
MGR organizational identification	0.16	0.11	0.18	1.44	0.15

Model predicting organizational commitment T2	b	se	β	t	p
Step 1
Location 1	0.40	0.21	0.29	1.93	0.06
Location 2	0.24	0.22	0.30	1.08	0.29
Location 3	0.10	0.26	0.07	0.38	0.70
Location 4	−0.13	0.28	−0.07	−0.48	0.63
Location 5	−0.68	0.42	−0.25	−1.63	0.11
Location 6	0.11	0.28	0.06	0.38	0.71
Location 7	0.31	0.29	0.16	1.07	0.29
Location 8	0.21	0.31	0.10	0.67	0.51
Work unit 1	−0.20	0.22	0.14	0.88	0.38
Work unit 2	−0.08	0.22	−0.06	−0.35	0.73
Work unit 3	−0.06	0.30	−0.02	−0.20	0.84
Work unit 4	−0.06	0.21	−0.05	−0.29	0.77
Work unit 5	0.09	0.21	0.08	0.44	0.66
Work unit 6	0.80	0.52	0.21	1.54	0.13
Work unit 7	0.14	0.23	0.13	0.61	0.54
Work unit 8	−0.018	0.21	−0.02	−0.08	0.93
MGR tenure with work unit	−0.05	0.03	−0.20	−2.01	0.05
Aggregated Org commitment T1	0.44	0.12	0.35	3.63	0.01
Step 2
MGR organizational commitment	−0.31	0.09	−0.43	−3.30	0.01
MGR organizational identification	0.14	0.09	0.22	1.65	0.11

Note(s): N = 107; MGR = Manager; Org = Organization; T1 = Time 1; Location and Unit entered as dummy variables. All statistics reported from the final model after computing step 2 in linear regression