Abstract
We consider the ability to detect interaction structure from data in a regression context. We derive an asymptotic power function for a likelihood-based test for interaction in a regression model, with possibly misspecified alternative distribution. This allows a general investigation of different types of interactions which are poorly or well detected via data. Principally we contrast pairwise-interaction models with ‘diffuse interaction models’ as introduced in Gustafson et al. (Stat Med 24:2089–2104, 2005).
Similar content being viewed by others
References
Greenland S (1983) Tests for interaction in epidemiologic studies: a review and a study of power. Stat Med 2: 243–251
Gustafson P (2001) On measuring sensitivity to parametric model misspecification. J Roy Stat Soc Ser B (Statistical Methodology) 63: 81–94
Gustafson P, Kazi AMR, Levy AR (2005) Extending logistic regression to model diffuse interactions. Stat Med 24: 2089–2104
Gustafson P, MacNab Y, Wen S (2004) On the value of derivative evaluations and random walk suppression in Markov Chain Monte Carlo algorithms. Stat Comput 14(1): 23–38
Le Cam L (1960) Locally asymptotically normal families of distributions. Univ Calif Public Stat 3: 37–98
Lehmann EL, Romano JP (2005) Testing statistical hypotheses. 3rd edn. Springer, New York
Liu J (2007) Average effects for regression models with misspecifications and diffuse interaction models, Ph.D. thesis, Department of Statistics, University of British Columbia
O’Brien SM, Kupper LL, Dunson DB (2006) Performance of tests of association in misspecified generalized linear models. J Stat Plann Inf 136: 3090–3100
Raftery A, Hoeting J, Volinsky C, Painter I, Yeung KY (2009) BMA: Bayesian Model Averaging, R package version 3.12., http://CRAN.R-project.org/pakcage=BMA
Raftery A (1995) Bayesian model selection in social research (with Discussion). Sociological Methodology, 25:111–196, Blackwells, Cambridge, Mass
White H (1982) Maximum likelihood estimation of misspecified models. Econometrica 50: 1–26
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Liu, J., Gustafson, P. On the detectability of different forms of interaction in regression models. Metrika 75, 347–365 (2012). https://doi.org/10.1007/s00184-010-0330-8
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00184-010-0330-8