Discrimination

I.  History

  1. 14th Amendment
  2. “Equal protection of the laws”

a)  Clearly applies to state and local governments; may apply to private entities.

  1. Yick Wo v Hopkins (1886)

a)  Each of more than 200 Chinese applicants to operate a laundry in wooden building is denied; all but 1 white successful

  1. Gomillion v Lightfoot (1960)
  2. State legislature redraws boundary of a city from a square to a 28-sided figure, excluding 395/400 black voters
  3. Brown v Board of Education (1954)
  4. Explicit use of social science evidence
  5. Civil Rights Act of 1964
  6. Clear expansion of 14th amendment protections.
  7. Creation of administrative agencies follows.
  8. Title VII

a)  Disparate treatment: employer “simply treats some people less favorably than others because of their race, color, religion or national origin.”

b)  Disparate impact: employment practices that are neutral on their face in their treatment of different groups but that in fact fall more harshly on one group than another.

  1. Legitimate business interest exception.

a)  Griggs v Duke Power Co. (1971): Employment tests that demonstrate disparate impact can be used in tests can be shown to serve a “valid business necessity.”

b)  International Brotherhood of Teamsters v US (1977)

II.  Standards

  1. Village of Arlington Heights v Metopolitan Housing Corp (1977)
  2. Standard of proof: “a clear pattern unexplainable on grounds other than race”
  3. Burden of proof: Plaintiff must provide evidence of conduct that “if otherwise unexplained is more likely than not based on the consideration of impermissible factors”
  4. Burden then shifts to defendant to show “some legitimate nondiscriminatory reason” for the outcome.

III.  Statistical Tests

  1. Binomial probability
  2. Castaneda v Partida (1977)

a)  Jury selection case

b)  USSC accepts binomial test and “2 or 3 standard deviations” standard

  1. Hazelwood School District v US (1977)

a)  USSC extends test to employment cases

  1. 80% Rule
  2. EEOC
  3. Swain v Alabama
  4. Connecticut v Teal (1982)
  5. Statistical problem with the 80% rule

a)  Example

Consider the example of a jurisdiction with 100,000 people -- 80,000 whites and 20,000 blacks -- with a jury pool having 900 whites and 100 blacks. In this case, blacks make up 20% of the population but only 10% of the jury pool. By the 80% rule, this difference would be too small to cause a presumption of discrimination even though whites are selected to be in the pool at a rate more than twice that of blacks (the ratio of the selection rates – 1.125%/0.5%).

The 80% Rule
Population / Pool / Selection rate
Whites / 80,000 / 900 / 1.125%
Blacks / 20,000 / 100 / 0.5%
Absolute rate / 20% / 10%

The problem becomes clear, if we examine the worst case scenario – when no blacks would be selected to be in the pool: the difference in the absolute rate is still not more than 20%, although the ratio of the selection rates is infinite.

The 80% Rule
Population / Pool / Selection rate
Whites / 80,000 / 1000 / 1.25%
Blacks / 20,000 / 0 / 0.0 %
Absolute rate / 20% / 0%

IV.  Typical Statistical Strategy

  1. Plaintiff - must build “prima facia” case
  2. Show disparate impact
  3. Demonstrate that effect cannot be explained by typical legitimate factors
  4. Defense
  5. Data too broad
  6. Data too shallow
  7. Intent not shown
  8. Effect explainable by other factors

V.  Social Science Used to Describe Context

  1. Price Waterhouse v Hopkins (1987)
  2. Empirical research on sex stereotyping is generally accepted
  3. Stereotyping can create discriminatory consequences
  4. Stereotypes about women shape perceptions of acceptable roles
  5. Sex stereotypes have demonstrably negative effects on women in work settings
  6. Conditions that promote stereotyping were present in petitioner's work setting
  7. Problem of "mixed-motive" cases