Mara Tableman

Analysis of Employment Termination Time Data with Competing Risks

Models for survival times describe either the survivor function or the hazard rate and their dependence on explanatory variables. As an example, in the USA, the question of whether older employees are fired by a certain company more often than their younger colleagues is investigated in legal claims. Such cases fall under The Age Discrimination in Employment Act of 1967, which federally protects individuals 40 years or older against age discrimination in employment decisions concerning hiring, firing, and promotion.

Employees may leave the company voluntarily for other reasons; such as to join a different company, relocation to a new geographical region for family reasons, retire, or early death. These other risks compete with involuntary termination (firing). Voluntary and involuntary terminations are thus “competing risks”.

In such legal cases flow data are collected to examine a specified observation period. The data observed are subject to right censoring as subjects may still be employed with the firm at the end of the study period. The data may also be subject to left truncation as subjects are hired at various times after the study start date. These subjects are said to have delayed entry into the study period. Some modifications of methods needed for this situation are presented together with the results for the legal case.

To request disability accommodations, contact the Office of the ADA Coordinator ten days in advance of the event. 543-6450 (voice) 685-3885 (FAX) 543-6452 (TDD) access@u.washington.edu