The American Statistician, August 1st, 2003
Austin and Brunner examined an interesting issue important to biomedical research: the situation in which an independent variable, [X.sub.1] in a multiple linear regression is "subject to a ceiling effect." This term covers what has also been referred to as a variable being "truncated" or "grouped" that is, that the independent variable used in analysis is [X'.sub.1] which equals [X.sub.1] when [X.sub.1] Consider the simple example in Table 1a. The model used by Austin and Brunner is: Y= [[beta].sub.0] + [[beta].sub.1][X.sub.1] + [[beta].sub.2][X.sub.2] + [[beta].sub.3][X.sub.1][X.sub.2] + [e...
HighBeam Research, Free Preview: 'Austin, P. C., and Brunner, L. J. (2003), "Type I Error Inflation in the Presence of a Ceiling Effect," The American Statistician, 57, 97-104: comment by Kraemer.(Letters to the Editor)'... Full Membership required for unlimited access. Free 7-day trial.
Subscribers: HighBeam content is only available to HighBeam subscribers. Click the link above for more information.
