You are not currently logged in.
Access JSTOR through your library or other institution:
On Testing for the Degree of a Polynomial
Paul G. Hoel
Vol. 10, No. 4 (Nov., 1968), pp. 757-767
Published by: Taylor & Francis, Ltd. on behalf of American Statistical Association and American Society for Quality
Stable URL: http://www.jstor.org/stable/1267457
Page Count: 11
Preview not available
Two methods are presented for deciding upon the proper degree of a polynomial regression function. One is based on a sequential F test and the other uses a two sample procedure. Both methods assume that the classical regression model is applicable. These methods enable one to realize desired guarantees against choosing either too small or too large a degree, which is not true for current fixed size sample methods.
Technometrics © 1968 American Statistical Association