ON INCREASING TREATMENT CONTRAST PRECISION AND THE ESTIMATION OF STRUCTURAL PARAMETERS IN COVARIANCE ANALYSIS
Abstract
Analysis of covariance in randomized and balanced incomplete block designs is reconsidered in terms of structural regression. In practice, the covariable is usually uncontrolled and may follow a linear model as does the variable of interest. If the covariable and treatments are independent, then the covariable model contains no treatment effect, and the treatment contrast precision on the variable of interest is increased, not only asymptotically, but in the finite sample. If the covariable is affected by one more of the treatments, then the estimation of direct and indirect effects is considered. Finally, when structural parameters are underidentified, in which case direct and indirect effects are not estimable, an alternative estimation procedure is discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1964
- Accession Number
- AD0603563
Entities
People
- William S. Mallios
Organizations
- Booz Allen Hamilton