how do we know that an organizational culture is discriminatory?

Underlying the debate about the ASA amicus brief is a fascinating theoretical question about the link between organizational culture and discrimination. I don’t want to wade into the specifics of the WalMart case – I’m much less informed about the case and the brief than are many of the commenters – but I think it’s worthwhile to think more generally about how one could establish empirically that an organizational culture leads to discrimination. What is the theoretical link between culture and discrimination and what kind of empirical evidence would you need to show this link? I’m also going to steer away from the legal interpretation of discrimination. My reason for writing this post isn’t to discuss employment law; rather I’m just interested in what we think about this from a social science perspective.

Organizational culture is a broad term that identifies differences between organizations in practices, beliefs, values, and symbols. Organizational culture consists of the unwritten rules of the game that organizational members rely on to get stuff done, make decisions, etc.  Culture shouldn’t be equated with the formal structure or demographic composition of the organization, although the two may co-vary. For example, a formal hiring policy isn’t part of an organization’s culture.  Organizational culture is also distinct from the broader set of institutional forces – norms, logics, etc. – that shape all organizations in a field.  We often imagine that organizational culture consists of distinguishing features of organizations.  I don’t think this is a necessary feature of organizational culture, but identifying the distinguishing features of a culture may be necessary from an empirical perspective if you want to demonstrate a causal link between culture and discrimination.

So what are the possible mechanisms that link culture to discrimination? Here are just a few possible mechanisms:

  • Informal selection criteria – what are the informal, shared criteria that members of the organization use to hire and retain employees (e.g., see Lauren Rivera’s work on hiring in professional firms)? Do these criteria favor some classes of individuals over others?
  • Values – Does the organization have a value system (e.g., collectivist or individualist) that promotes the development of in-group biases? Do these biases lead to unfavorable evaluations of people who are not a part of the in-group?
  • Selection practices  and settings – irrespective of job performance, how and where are decisions about hiring and retention made? Are those decisions made in places or using activities that vary in their appeal to certain classes of people? For example, if retention depends on your liking of golf, the organization would naturally favor retaining or hiring people who play a lot of golf.
  • Beliefs or personality traits – does the organization favor beliefs or personality traits that are disproportionately found among a certain class of people (e.g., beliefs about working overtime)? For example, if the organization favors aggressive behavior, then the organization would tend to hire and retain people who have this personality trait.

Once you’ve identified theoretical mechanisms, I think there are three parts to empirically verifying a link between culture and discrimination.*  You first want to show that these elements of culture actually vary across organizations. If your claim is that a particular kind of culture is especially discriminatory, then you want to be able to show how that kind of culture varies.  If you can’t show variance across organizations, then you really can’t establish that any single cultural element leads to more bias than another. Second, you want to show that individuals with the same characteristics receive differential treatment in hiring and retention in different organizational cultures. Ideally, you’d like some sort of natural experiment where you took the same individual and had them apply for a job at organizations that varied along the different cultural dimensions. Alternatively, you’d show how roughly equivalent individuals are treated differently in hiring and retention, depending on the type of culture. Finally, you need to be able to make the case that cultural bias in hiring/retention decisions is independent of job performance.  It may be that individuals who have certain beliefs are going to perform better in that organization than individuals who do not share those beliefs. If this is the case, than cultural competence may be the cause of outcome differences rather than superficial preferences.

I imagine this last empirical criterion is the most controversial because of feedback processes within the organization. It may be that individuals differ in performance because an organization’s culture inhibits the employees’ development. For example, if women are given fewer leadership training opportunities because they are not given access to the same socialization settings as men, then you could still show that the culture indirectly shapes discrimination.  This would essentially require a fourth step of empirical verification. Does culture mediate discrimination by constraining the kinds of opportunities that employees of a certain class have for advancement? If you can show that this is the case, then you can still empirically demonstrate a causal link between culture and discrimination.

* Demonstrating causality in social science is notoriously difficult to do. At best we can show that X is one of the causes of Y. Usually we are only capable of showing that we have good reasons to think that  X might be one of the causes of Y.

Written by brayden king

May 23, 2011 at 5:49 pm

6 Responses

Subscribe to comments with RSS.

  1. I am actually pretty comfortable with the plaintiff’s position that an organizational culture can create discrimination across decentralized work-sites. (In fact, I am currently working on a theory piece that makes a parallel argument that shared understandings can be equivalent to centralized coordination). This is a scientific point.

    On the other hand, I am skeptical that such an indirect mechanism of coordination qualifies as the kind of centralized control necessary to certify a chain-wide class and the kind of intentionality necessary to claim punitive damages. This is a legal and political point.

    Of course (unlike some of the people who have been contributing here) I have no special expertise in discrimination law and so my opinions on the subject aren’t really relevant. But that’s really my point: you can draw a meaningful distinction between the scientific question of how we establish what happened and the legal question of how such evidence is relevant to the statute and case law.

    This is why I don’t see the defendant’s brief as an attack on sociology. It is entirely possible to accept the validity of our expertise (and bracket judgement as to the validity of a specific expert finding) while still rejecting a legal theory that would make our findings irrelevant to a class of legal questions. Read carefully, that’s pretty much Wal-Mart’s argument. Such a legal theory may make our expertise less relevant, but it’s not an attack on our expertise any more than an insurance policy (or a court case adjudicating it) exempting earthquake damages is an attack on the expertise of geologists.



    May 23, 2011 at 6:37 pm

  2. Aren’t all organizations discriminatory in the sense that the firms are trying to find the more qualified people in the general applicant pool and in their workplace that have the qualities and traits to be successful at their jobs in the firm?

    Does the firm need to reorganize if there are also secondary effects?

    For example, suppose the firm uses overtime and rotates night and weekend shifts with daytime weekday workers because stores are open 7 days, 6 AM to 11 PM, plus inventory checking and restocking after closing.

    If the above firm characteristics results in fewer qualified female applicants and workers, lower female retention rates and fewer female versus male promotions (fewer female night shift supervisors, etc.), is it illegal discrimination?


    Milton Recht

    May 23, 2011 at 8:42 pm

  3. A new (and timely given the discussion) piece out in Social Forces works to clarify the connection between gender discrimination and organizational culture:

    Stainback, Kevin, Thomas Ratliff and Vincent J. Roscigno. 2011. “The Context of Workplace Sex Discrimination: Sex Composition, Workplace Culture and Relative Power,” Social Forces 89(4): 1165-1188.



    May 23, 2011 at 11:32 pm

  4. Milton: see “bona fide occupational qualification.” This is a red herring. If, in fact, more men prefer night work, that is one thing. But employers are not allowed to discrimination categorically on the basis of their assumptions about night work preferences. Acceptance of applications net of employment and promotions net of accepting the job should deal with the issue. If you stop and think about it, you will realize that many female-dominated jobs involve shift work and night work, including nursing, flight attendants, sex workers, and dancers. This particular path was well trodden in the 1970s. There has already been lots of work that seeks to identify the pool of “qualified” applicants in terms of what it takes to actually do that job.

    What employers can’t do is say that it’s important to them for their employees to be able to put up pictures of naked women or express prejudice against women or minorities on the job or to hire people who want to play in the company football league (unless the football league is a part of the job).

    In any event, the Wal-Mart case involves statistical evidence that other comparable employers do not show the gender differences that Wal-Mart shows. Wal-Mart is trying to defend itself in a situation where most sociologists would say there is overwhelming evidence of a problem. The ground it is defending is that it should not have to be treated as a single employer in a class action lawsuit.



    May 24, 2011 at 12:40 am

  5. If physical and personality traits are correlated to gender and performance, then is there any case for discrimination?

    If physical and personality traits (e.g. height, shyness) directly affect performance, in what case the organization is guilty for discrimination, esp. in academic sense?



    May 24, 2011 at 4:44 am

  6. passerby,

    see OW’s comment directly above.

    besides, we’re not talking about firefighters who need to run 50 yards while carrying 200 pounds but retail workers going up for managerial positions. the relevant traits are likely to be only very weakly correlated with gender. furthermore, the plaintiff’s statistical analysis has a reasonably good set of observables.



    May 24, 2011 at 5:28 am

Comments are closed.

%d bloggers like this: