Constructing explained and explanatory variables with strange statistical analysis results

Journal of the Royal Statistical Society Series D: The Statistician

Volumn 47, Issue 2, 1998, Pages 377-383

  Bertrand, Philip V ^a  

^a OXFORD BROOKES UNIVERSITY   (United Kingdom)

Author keywords

Generalized inverse; Gram schmidt procedure; Orthonormalization; Quirk; Singular value decomposition; Spectral decomposition

Indexed keywords

EID: 0346541846     PISSN: 00390526     EISSN: None     Source Type: Journal    
DOI: 10.1111/1467-9884.00138     Document Type: Article

References (9)

1. Belos, A. (1993) Educational research - singled out by the State. Guardian, Oct. 26th, 2.
2. Bertrand, P. V. and Holder, R. L. (1988) A quirk in multiple regression: the whole regression can be greater than the sum of the parts. Statistician, 37, 371-374.
3. Cheng, R. C. H. (1985) Generation of multivanate normal samples with given sample mean and covariance matrix. J. Statist. Comput. Simuln. 21, 39-49.
4. Cuadras, C. M. (1993) Interpreting an inequality in multiple regression. Am. Statistn, 47, 236-258.
5. yx2 (revisited)? Am. Statistn, 42, 89-90.
6. yx2: correlated variables are not always redundant. Am. Statistn, 41, 129-132.
7. Mardia, K. V., Kent, J. T. and Bibby, J. M. (1993) Multivariate Analysis. London: Academic Press.
8. Routledge, R. D. (1990) When stepwise regression fails: correlated variables some of which are redundant. Int. J. Math. Educ. Sci. Technol., 21, 403-410.
9. Styan, G. P. H. (1983) Generalized inverses. In Encyclopedia of Statistical Sciences (eds S. Kotz and N. L. Johnson), vol. 3, pp. 334-337. New York: Wiley.