Testimator

A testimator is an estimator whose value depends on the result of a test for statistical significance. In the simplest case the value of the final estimator is that of the basic estimator if the test result is significant, and otherwise the value is zero. However more general testimators are possible.^[1]

History

An early use of the term "testimator" way made by Brewster & Zidek (1974).^[2]

References

1. Adke, S.R., Waikar, V.B. & Schurmann, F.J. (1987). "A two stage shrinkage testimator for the mean of an exponential distribution". Communications in Statistics - Theory and Methods, 16 (6), 1821-1834. Retrieved April 16, 2009, from http://www.informaworld.com/10.1080/03610928708829474 (restricted access)
2. Brewster, J.F., Zidek, J.V.(1974) "Improving on Equivariant Estimators". Annals of Statistics, 2 (1), 21–38

Further reading

• S. M. Kanbur, C. Ngeow, A. Nanthakumar and R. Stevens (2007). "Investigations of the Nonlinear LMC Cepheid Period‐Luminosity Relation with Testimator and Schwarz Information Criterion Methods", Publications of the Astronomical Society of the Pacific 119, 512–522 online
• Mezbahur Rahman, Gokhale, D.V. (1996) "Testimation in Regression Parameter Estimation", Biometrical Journal, 38 (7), 809–817 online
• Abramovich, F., Grinshtein, V., Pensky, M. (2007) "On optimality of Bayesian testimation in the normal means problem", Annals of Statistics, 35, 2261–2286 online