0

Affirmative Action, in the context of race/ethnicity, can cover a variety of measures such as

  1. simple non-discrimination (depending on definitions this might not be classed as "affirmative action" at all)

  2. targeting job advertisements at particular groups

  3. actually favouring target groups in the process of considering applications.

You don't need to have any predefined racial categorisations in order to do (1) because you can have an anti-discrimination law without having legally defined race categories. Anyone bringing a discrimination claim has to show they they are of a identifiable race/ethnicity and that they were unfavourably treated by a named individual (or organisation) because of that race/ethnicity. But the race/ethnicity is identified by the individual complainant in the individual case. There is no need for a predefined list of races/ethnicities.

In order to do (2) you need some very broad conception of the particular group or groups you are targeting but you don't need an actual definition.

However if you do (3) you need a reasonably precise definition of the race/ethnicity which the application process is to favour so that those administering the application process handle applications in the intended way. My question primarily has (3) in mind and is:

Can anyone give any examples of how the relevant race/ethnicity is /was defined in an affirmative action programme that they know of?

0
1

Definitions used by affirmative action policies are provided by other laws and policies which create those definitions and prescribe the manner in which they are assessed. How they are understood, measured, and even categorized changes with time as society collectively learns more.

In The United states, these definitions come from the Census Bureau, whose mission is to count all persons in the United States every ten years and thus to whom demographic information is mission critical.

The Census Bureau provides information, including a synopsis of several categories, here.

Contrary to your claim, however, no degree of precision is actually required, so long as you are comfortable passing that lack of precision through to the final product. Many policy systems are built around the principle of "good enough." Sometimes lampooned as "good enough for government work."

6
  • 1
    I take your general point about "good enough for government work" but what puzzles me is this. The Census Bureau link says racial categories are entirely self-identified and a person may be of more than one race. If there is an Affirmative Action policy in any area it means that those who come within its scope will obtain some kind of benefit - maybe a very valuable benefit (such as free college education for example). If the difficulty of defining race is got over by it being entirely self-defining, what is to prevent anyone identifying as the relevant race. Is there really no objectivity?
    – Nemo
    Nov 23 at 18:15
  • 1
    @Nemo That issue is not new or unique to racial identity. In an ideal world, obviously all data would be validated. However, in the real world, data validation has a cost to it and that cost must be compared to the cost of allowing false-positives enter into the system. The Census does spot-check it's data by performing audits of samples (racial data is cross-checked against birth/immigration records, for example), so it knows it's error rate can be as high as 10%. The cost of getting that error rate to 0% is likely far in excess of the cost of allowing false positives. Nov 23 at 18:29
  • I understand spot checks in general but not with regard to race. If race is entirely self-defined people can change race, add races etc. The fact that the race they primarily identify with now is different from the race they primarily identified with upon immigration (or the race their parents primarily identified with at their birth) would not be an "error".
    – Nemo
    Nov 23 at 18:40
  • 1
    @nemo Perhaps not an error in the sense of "they're wrong" but in the sense that a discrepancy exists which undermines the presumptive validity of the data, yes. If anything, however, the data suggests that the choices being made are mostly false negatives: economics.ucdavis.edu/events/papers/1112Nix.pdf A reason Aff. Act. makes sense is that people will self-select OUT of the beneficiary group, despite the fact that they lose those benefits. It's hard to say that they are overly enriched by the program under those conditions, which is sufficient reason to avoid costly inquiry. Nov 23 at 18:58
  • 1
    OK. So if I understand this correctly you are saying that very general racial categories are used and people self-select whether they fit the target category and the number of people who "game the system" is sufficiently low for the program to be "good enough for government work".
    – Nemo
    Nov 23 at 19:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .