Search

Guidelines for using an expert monitor to monitor a plaintiff's remedy to monitor the remedy the court imposes for an employment discrimination lawsuit are given.

Employment discrimination cases can be tried with hypergeometric or binomial statistics. As both produce a standard deviation and a probability to determine statistical significance, both can be used in disparate impact analyses. Disparate impact can be analyzed by threshold, by barrier and by selection and these statistical distributions can be useful with all three. After adverse impact is demonstrated, the employer can prove validity, and the plaintiff has a chance to show if there is another practice with less adverse impact that will meet the employer's legitimate needs.

Disparate impact for job discrimination is defined statistically using three references. Various federal employee selection guidelines use an 80% rule though allow for discrimination occurring when this rule is met if separate hiring practices have a lower diversity rate. US Supreme Court cases state that individual practices can be discriminatory even when the overall hiring shows no disparate impact. The 1991 Civil Rights Act focuses on specific not overall processes. Discrimination can be determined through a threshold test and a barriers test using binomial statistics or applicant data depending on the findings.