Testing for transformations
In Section 1.8 we outlined briefly the point that difficulties can arise from using multiple regression analysis due to the failure of assumptions made in the ordinary least squares (OLS) analysis. The assumptions can fail because of lack of normality, homoscedasticity, independence, or because the underlying model is not linear in the unknown parameters. In this chapter we examine the use of transformations, either of the response variable or of the explanatory variables, to attempt to satisfy the assumptions of an OLS analysis.
Unable to display preview. Download preview PDF.
- Atkinson, A. C. (1982) Regression diagnostics, transformations and constructed variables. J. Roy. Statist. Soc. B, 44, 1–13.Google Scholar
- Box, G. E. P. and Cox, D. R. (1964) An analysis of transformations. J. Roy. Statist. Soc. B, 26, 211–252.Google Scholar
- Cook, R. D. and Weisberg, S. (1982) Residuals and Influence in Regression. Chapman and Hall, London.Google Scholar
- Ramsey, J. B. (1969) Tests for specification errors in classical linear least squares regression analysis. J. Roy. Statist. Soc. B, 31, 350–371.Google Scholar