# Suppression and Collapsibility for Log-linear Models

• Published : 2004.12.01

#### Abstract

Relationship between the partial likelihood ratio statistics for logisitic models and the partial goodness-of-fit statistics for corresponding log-linear models is discussed. This paper shows how definitions of suppression in logistic model can be adapted for log-linear model and how they are related to confounding in terms of collapsibility for categorical data. Several $2{times}2{times}2$ contingency tables are illustrated.

#### References

1. Agresti, Alan.(1984). Analysis of Ordinary Categoriml Data, New York: John .Wiley and Sons
2. Bishop, Yvonne M. M., Fienberg, Steve E. and Holland, Paul W.(1975). Discrete Multivariate Analysis, Cambridge, Massachusetts: MIT Press
3. Boivin, J-F. and Wacholder, S.(1975). Conditions for Confounding of the Risk Ratio and of the Odds Ratio, Amerimn Journal of the Epidemiology, Vol. 121, pp. 152-158
4. Christensen, Ronaldo'(1990). Log-Linear Models, New York: Springer-Verlag
5. Ducharme, G. R. and Lepage, Y.(1986). Testing Collapsibility in Contingency Tables, Journal of the Royal Statistical Society, B, 48(2), 197-205
6. Freund, R. J.(1988). When is $R^2>r^2_{yx1}+r^2_{yx2}$?(Revisited), The Amerimn Statistician, Vol. 42, pp. 89-90
7. Geng, Z.(1992). Collapsibility of Relative Risk in Contingency Tables with a Response Variable, Journal of the Royal Statistical Society, B, 54(2), 585-593
8. Grayson, D. A.(1987). Confounding Confounding, American Journal of Epidemiology, Vol. 126, pp. 546-553 https://doi.org/10.1093/oxfordjournals.aje.a114687
9. Hamilton, D.(1987). Sometimes $R^2>r^2_{yx1}+r^2_{yx2}$ Correlated Variables are not Always Redundant, The Amerimn Statistician, Vol. 41, pp. 129-132 https://doi.org/10.2307/2684224
10. Hamilton, D.(1988). (Reply to Freund and Mitra), The Amerimn Statistician, Vol. 42, pp. 90-91
11. Horst, P. (1941). The Role of Prediction Variables Which are Independent of the Criterion, in The Prediction Adjustment, ed. P. Horst, New York: Social Science Research Council, pp. 431-736
12. Kleinbaum, D. G., Kupper, L. L. and Morgenstern, H.(1941). Epidemiologic Research: Principles and Quantitative Methods, California: Lifetime Learning Publications
13. Lynn, H. S.(2003). Suppression and Confounding in Action, The Amerimn Statistician, Vol. 57, pp. 58-61 https://doi.org/10.1198/0003130031090
14. Mitra, S.(1988). The Relationship Between the Multiple and the Zero-Order Correlation Coefficients, The Amerimn Statistician, Vol. 42, pp. 89
15. Schey, H. M(1993).The Relationship Between the Magnitudes of SSR $(x_2)$and SSR(X2 I Xl) : A Geometric Description, The Amerimn Statistician, Vol. 47, pp. 26-30 https://doi.org/10.2307/2684778
16. Sharpe, N. R. and Roberts, R. A.(1997). The Relationship Among Sums of Squares, Correlation Coefficients, and Suppression, The Amerimn Statistician, Vol. 51, pp. 46-48 https://doi.org/10.2307/2684693
17. Whittemore, A. S.(1978). Collapsibility of Multidimensional Contingency Tables, Journal of the Royal Statistiml Society, B, 40(3), 328-340