A Robust Test for Multi-Sample Location Problems with Unequal Group Variances and Nonnormality

✍ Scribed by Dr. M. A. Tabatabai; US W. Y. Tan

• Publisher: John Wiley and Sons
• Year: 1988
• Tongue: English
• Weight: 283 KB
• Volume: 30
• Category: Article
• ISSN: 0323-3847
• DOI: 10.1002/bimj.4710300303

✦ Synopsis

Sunmn y

A robust test (to be referred to as Y* test) is proposed for W i n g equality of several group means without assuming normality> and equality of variances. This test statistic ie obtained by combining Tiku's MML robust procedure with the James statistic. Monte Carlo simulation studies indicate that the M* test is more p6werful than the Welch test, the James test, and the testa baaed on Huber's M-estimators over a wide range of nonnormal universes. It is also more powerful than the .Frown and Forsythe test under most of nonnormal distributions and has substantially the aame power as the Brown and Forsythe test under normal distribution. Comparing with Tan-Tabatabai test, M* is almost aa powerful as Tan-Tabatabai test.

K e y w ~d s

and phrases: Brown-Forsythe teat;: Huber's M-estimators; James test ; Monte Carlo comparisons ; Tan-Tabatabai test ; Tiku's MML estimator;iWelch test.