Wednesday, January 29, 2020

The Unintended Consequences of Diversity & Inclusion Initiatives: Lisa Leslie’s Recent Work & Ideas for Putting her Framework to Use


By Rebecca Temkin and Taimi Itembu


I.         Introduction


Increasingly, evidence shows that organizational diversity and inclusion initiatives (DIIs) are frequently ineffective, or worse, that they lead to worse diversity and inclusion related outcomes.  For example, data shows that standard sexual harassment training may actually worsen behavior among certain employees, leading to increased rates of harassment following trainings (Dobbin & Kalev, 2017).  Professor Lisa Leslie, a thought leader in organizational diversity, seeks to understand and explain the mechanisms by which organizational DIIs are failing. Recently, she has developed a broad theory to explain the unintended consequences of DIIs.

In her 2019 WAPPP presentation, Leslie presented her framework of unintended consequences, which may be used to anticipate, evaluate, and ward against unintended consequences of diversity and inclusion efforts. Specifically, she discussed four types of unintended consequences that result from such initiatives, four signals (ie. interpretations of DIIs) that give rise to the typology of unintended consequences, and the implications of this soon-to-be-published theory for organizational DIIs (Leslie, 2019).


II.         Leslie’s 4 Types of Unintended Consequences

Leslie posits that well-intended DIIs seek to achieve three main outcomes: increased representation of the target group, especially organizational leadership; reduced gaps in career success between the target and non-target groups; and increased feelings of inclusion among target group members.  However, research shows that DIIs produced mixed results.  Thus, Leslie developed a theory to explain the heterogeneous effects of DIIs.  This theory recognizes four types of unintended consequences - backfire, negative spillover, positive spillover, and false progress - which result from multiple and overlapping signaling pathways.

The four types of unintended consequences can be placed in a matrix (Figure 1), with desirability of outcome along the vertical axis, and the intentionality of the outcome variable itself on the horizontal axis.  A backfire is an undesirable effect on an outcome variable that the policy was trying to impact.   Negative spillover is an undesirable effect on an outcome variable that the policy was not intended to impact. Positive spillover is a desirable effect on an outcome variable that the policy was not intended to impact.  False progress is the appearance of a desirable effect on an outcome variable that the policy was trying to impact, without any true change. 



Unintended Consequences
Outcome variable affected WAS intended to be affected
Outcome variable affected WAS NOT intended to be affected
Undesirable Effect
Backfire
Negative spillover
Desirable Effect
False progress
Positive spillover
Figure 1.


To illustrate these four unintended consequences, imagine an organization put in place a diversity program with the singular goal of increasing representation of African-American employees in leadership positions.  If the result of the program was a decrease in African-American organizational leaders, this would be backfire, because it is the opposite of the intended effect.  If the program led white employees to resent African-American employees, this would be a negative spillover effect, because this is an undesirable outcome on a variable (attitude of non-target group members towards target group members) that the program did not set out to impact.  If the program led employees to gain increased respect for the organization, this would be positive spillover, because this would be a desirable outcome on a variable that the policy was not trying to impact.  Finally, if the program led to alterations in African-American employees’ job titles to appear managerial without accompanying shifts in organizational structure or responsibilities, this would be false progress, because there would appear to be a positive change on the intended outcome variable without any true change.


III.         Leslie’s Signalling Theory

Leslie notes, that even if leaders have good intentions, implementing diversity and inclusion policies may send signals to other people that lead to unintended consequences. The four signals she talks about that might result from implementation of DIIs are: targets are more likely than non-targets to succeed; targets need help to succeed; the organization values morality; the organization values making progress on diversity goals. We provide examples of each signalling pathway below.

Signal 1 Targets are more likely than non- targets to succeed: 
For example if a non-target employee perceived it to be unfair that black employees are more likely to be promoted as a result of the DII, they may resent African-American organizational leaders, which would be a negative spillover effect.

Signal 2 Targets need help to succeed:
This signal may trigger negative stereotypes.  For example, if African-American employees are more likely to ascend to leadership positions based on the DII, this might trigger non-target employees to perceive target employees as threats.

Signal 3 Organization values morality:

A company with great DIIs may signal that they value morality, prompting greater employee motivation and organizational engagement, two positive spillover effects.

Singal 4 Diversity goal progress is valued:
A DII may signal that the objective is to look good, rather than to improve in diversity and inclusion.  This may translate into “for show” behaviors that do little to increase workplace diversity, which is false progress.

Each of these signals might act through multiple psychological or behavioral pathways, leading to multiple responses and consequences (Figure 2).  Further, the type of DII implemented likely impacts the signals sent or received.  For example, resource-based policies, which typically provide more concrete and practical interventions, such as  mentoring programs for target groups, often signal that target groups are most likely to succeed or that the targets are in need of help. Non-discrimination policies sometimes referred to as “identity-blind” practices, seek to ensure that hiring and promotions are based on skills and qualifications and not on demographics. Non-discrimination policies send a strong signal that morality is valued. Accountability initiatives create mechanisms that assign responsibility for and monitor diversity outcomes, and often signal that meeting DII goals is the organization’s focus, rather than improving diversity itself, making these initiatives more likely to lead to false progress.



IV.         Discussion

The identities of organizational leaders implementing DIIs and other setting-specific characteristics are likely to impact the signals at play, and therefore the potential unintended consequences of any given DII.  Leslie acknowledges this, and indicated, much the way Iris Bohnet did when asked how to anticipate what instruments will be most effective for de-biasing organizational systems and procedures (Bohnet, 2019), that every setting is extremely nuanced and must be assessed individually. 

The desire for a more nuanced theory or framework speaks to a tension common to many implementation-focused disciplines and practices.  Namely, there is a paradoxical desire for a framework to be ‘one-size-fits-all’ -- applicable to any case -- and simultaneously detailed and nuanced enough to appropriately address the unique features of each setting in which it is applied.  Unfortunately, such a framework does not exist for organizational diversity, and indeed, may never be possible or appropriate.  However, Leslie’s framework of types and causes of unintended outcomes is a tool that can be used to assess existing and potential DIIs, to anticipate and minimize the undesirable unintended effects, and to maximize the desirable unintended effects.

V.         Recommendations: What Works

For organizations seeking to implement DIIs, we urge leaders to anticipate the unintended effects of potential programs, using Leslie’s framework as a guide.  Consider the signals a given policy might send in the specific organizational context - will race, gender, parental status, employees or other characteristics of program leadership impact how your new policies are viewed?  What strategies might you use to mitigate or counteract anticipated negative interpretations of your new policy?  Attitudes, communication, and behaviors from organizational leaders will have tremendous impacts on the success or failure of DIIs, because as Dobbin and Kalev say, “culture is shaped from the top” (Dobbin & Kalev, 2017). 

Thus, CEOs and other leaders need to take a strong public stand - they should be the first in line for DII trainings and programs, and should chair the committees tasked with solving organizational challenges. In the U.S. Armed Forces, in response to alarmingly high rates of sexual harassment experienced by women service members, major anti-harassment efforts were implemented, with increased leadership involvement through anti-harassment messaging, regular trainings, and the establishment of formal and informal reporting mechanisms. Women who reported that their commanders supported these measures also reported less harassment, and greater satisfaction with the organisation’s responses to their complaints (Dobbin & Kalev, 2017).

Next, rather than assuming a new initiative will work, Leslie notes that organisations need to implement a holistic approach to evaluation. It is important to measure not only diversity goal progress, but other mechanisms and outcomes, such as ethical climate perception, 
employee organizational engagement, and employee behaviors.  Essentially, DII leaders must assess progress in ways that guard against “false progress” effects, and allow for discovery of negative outcomes of the initiative. Leslie’s framework of unintended consequences would be a useful guide in developing an evaluation strategy that measures both intended and unintended consequences of DIIs.

Finally, in developing DIIs, representation is a key piece to the puzzle.  As Dobbin and Kalev explain, “harassment flourishes in organizations where few women hold the ‘core’ jobs. Fixing this is about finding power in numbers, not just in authority and hierarchy” (Dobbin & Kalev, 2017).  As explained by Monica Ramirez, a champion for farmworker, Latina, and immigrant women, and a strong supporter of worker-led movements, we must make space for and pass the mic to those who are underrepresented to tell their own stories and become leaders in change (Ramirez, 2019). Organizations seeking to improve in diversity and inclusivity should consider these points seriously, and ensure that the members of groups they hope to lift up hold meaningful positions within the organization and have opportunities to shape new organizational DIIs.


VI.         Conclusion

The impact of any given policy or program will undoubtedly result in heterogeneous effects across different contexts, and while good intentions about increasing diversity and inclusivity are admirable, good intentions are not enough.  With frameworks such as Leslie’s in hand, we now have tools to assess the impacts of those good intentions, and it is time for every organizational leader to ensure that DIIs do not cause more harm than good.