Separating Error, Nuisance Effect and Main Effect: Tangent Models and Third Order Inference
D.A.S. Fraser and N. Reid (1996). Separating Error, Nuisance Effect and Main Effect: Tangent Models and Third Order Inference. Journal of Statistical Research, Vol. 30, No. 1, pp. 1-8.
Abstract
Normal linear regression provides a standardized procedure for separating error, nuisance effect, and main effect. A parallel procedure is discussed for the third order asymptotic context: this provides a resolution of apparently contradictory procedures found with exponential and transformation models. The analysis uses tangent location and tangent exponential models.
