A few days ago, on my LinkedIn Group, Charles Krugel’s Labor & Employment Law & Human Resources Practices Group, member Stephanie R Thomas, Ph.D., the owner of Thomas Econometrics in the Philadelphia, PA, area, posted an excellent examination of the use and importance of statistical analysis in disparate impact and disparate treatment discrimination cases. I thought that this presentation and our accompanying LI group discussion would be an interesting post.
At the core of Stephanie’s presentation is the below Slideshare/Youtube presentation (click on the link below):
Statistical Analysis as Evidence in Employment Discrimination Litigation
Additionally, there was some good back and forth discussion on my LI Group concerning the importance of statistical analysis. Instead of summarizing the discussion, I’ve posted it in its entirety below:
Charles Krugel test.mielus.ro/chuck • This is great Stephanie! Thanks for posting it. After watching your presentation, my impression is that statistical analysis may be more important for combatting class action certification than actual disparate impact discrimination. Am I wrong?
Stephanie R Thomas, Ph.D. • Statistical analysis has a role in class certification, disparate impact and disparate treatment cases. For class certification, statistics can be used to examine commonality and typicality – I have a separate presentation on this (http://www.slideshare.net/TheProactiveEmployer/examining-class-certification-issues-with-statistical-analysis).
For disparate impact claims, I see statistics as having two roles. First, you can statistically examine your employment practices (hiring, promotion, termination, requirements for a given position, etc.) with respect to various protected group definitions proactively to identify any potential problem areas before you get to the litigation stage. Second, in the litigation context, statistics can be used to examine the challenged policy or practice to determine if there is statistical support for the claim of disparate impact.
With respect to disparate treatment claims, statistics can be used to assess whether there is statistical support for the claim. The analysis for disparate treatment is typically more complex than for a claim of disparate impact, because of the need to carefully define comparator groups and similarly situated employee groupings. But statistics can be very useful in assessing the merits of disparate treatment claims.
Charles Krugel test.mielus.ro/chuck • Thanks again Stephanie! The use of statistical analysis in disparate treatment claims is an interesting idea.
The reason that I wonder if statistical analysis is more beneficial at the class action certification stage of litigation, in disparate impact claims, is because that ruling lies with the judge not a jury. Whereas in a disparate treatment claim, the final determination of guilt or innocence is usually made by a jury.
In my opinion, it requires incredible communication skills to get statistical analysis across to a jury in a disparate treatment case. Obviously, on a case-by-case basis, it has to be determined whether and how to get this analysis across to a jury in a such a case.
I wonder if statistics can be used as a means to assess bias by a teacher against a student. the final class for her graduation was an F that prevented her graduation. she was to graduate in january. the course required for her major is not offered again until the fall of the next year. she loses her full scholarship and comes from a family who cannot afford to support her. her plans to apply to law school are totally disrupted. her confidence is shaken and she is suffering from the consequences of this.
her prior record was 27 A, 7 B, 1 C, 0 D, 0 F. i modeled her grades as a poisson distribution, with 0=A, 1=B etc. the distribution was perfect, with both variance and mean = 0.27. when testing the likelihood of a grade of F being earned by this student is p<0.0002, or less than 1 chance in 5000 it could be her grade.
she has been a star student at a major university in the midwest. she is one of two african-american students in a class of 32 for this course. she has felt hostility from this professor in the past who is also her advisor. the student asked the professor about the grade and he said she did poorly. she asked to meet with dean to discuss, but the dean had the professor waiting in the room when she arrived, the professor took over the meeting without her defending herself and she left in tears.
someone needs to help in this situation. it is unjust and smells of bias, and incredibly damaging to her emotionally and to her career. she wants to give up and forget about ever graduating, has no money and no scholarship with the F.
i didnt know if i could use a statistical argument to help her to get an external review or legal action. (i am a scientist, not a lawyer.) your comments would be welcome. you can reach me at my email: rich32323@yahoo.com.
thank you.
richard
PS:
I should add that she had a B for her mid-term grade and that she did all of the homework assignments for the class, without any evidence she was doing poorly. the professor told her she got a 66 on the final but then in the dean’s office, he lowered it to 57. he has not revealed his grading method, how a grade could go from B to F in this case (this would be hard to understand mathematically), nor has be provided the graded exam.
her GPA is 3.7 without this grade. during the final fall semester when she got this F, in her three other courses she got an A. that is, three A’s and 1 F. the F doesnt make sense. there is something wrong here and i believe it is not the student. she is helpless and distraught.
any help, guidance, advice would be greatly appreciated.
bottom line: this grade is not reflective of this student.