Abstract
The small sample distribution of the trimmed mean is usually approximated by a Student’s t distribution. But this approximation is valid only when the observations come from a standard normal model and the sample size is not very small. Moreover, until now, there is only empirical justification for this approximation but no formal proof. Although there are some accurate saddlepoint approximations when the sample size is small and the distribution not normal, these are very difficult to apply and the elements involved in it, difficult to interpret. In this paper we propose a new approximation based on the von Mises expansion for the tail probability functional of the trimmed mean, which improves the usual Student’s t approximation in the normal case and which can be applied for other models. This new approximation allows, for instance, an objective choice of the trimming fraction in a context of hypothesis testing problem using the new tool of the p value line.
Similar content being viewed by others
References
Bickel PJ (1965) On some robust estimates of location. Ann Math Statist 36:847–858
Easton GS, Ronchetti E (1986) General saddlepoint approximations with applications to \(L\) statistics. J Am Stat Assoc 81:420–430
Field CA, Ronchetti EM (1985) A tail area influence function and its application to testing. Commun Stat 4:19–41
García-Pérez A (2003) von Mises approximation of the critical value of a test. Test 12:385–411. doi:10.1007/BF02595721
García-Pérez A (2006a) Chi-Square tests under models close to the normal distribution. Metrika 63:343–354. doi:10.1007/s00184-005-0024-9
García-Pérez A (2006b) \(t\)-tests with models close to the normal distribution. In: Balakrishnan N, Castillo E, Sarabia JM et al (eds) Advances in distribution theory, order statistics, and inference. Birkhäuser-Springer, Boston, pp 363–379. doi:10.1007/0-8176-4487-3_23
García-Pérez A (2008) Approximations for \(F\)-tests which are ratios of sums of squares of independent variables with a model close to the normal. Test 17:350–369. doi:10.1007/s11749-006-0036-4
García-Pérez A (2011a) Hotelling’s \(T^2\)-test with multivariate normal mixture populations: approximations and robustness. In: Pardo Balakrishnan, Gil (eds) Modern mathematical tools and techniques in capturing complexity. Springer, Berlin, pp 437–452. doi:10.1007/978-3-642-20853-9_30
García-Pérez A (2011b) Another look at the Tail Area Influence Function. Metrika 73:77–92. doi:10.1007/s00184-009-0266-z
García-Pérez A (2012) A linear approximation to the power function of a test. Metrika 75:855–875. doi:10.1007/s00184-011-0356-6
García-Pérez A (2014a) The p value line: a way to choose the tuning constant in tests based on the Huber \(M\)-estimator. Test 23, vol 3. doi:10.1007/s11749-014-0361-y
García-Pérez A (2014b) An approximation to the small sample distribution of the trimmed mean for Gaussian mixture models. In: Advances in intelligent systems and computing, Springer
Hampel FR (1974) The influence curve and its role in robust estimation. J Am Stat Assoc 69:383–393
Hampel ER, Ronchetti EM, Rousseeuw PJ, Stahel WA (1986) Robust statistics: the approach based on influence functions. Wiley, New York
Hall P, Padmanabhan AR (1992) On the bootstrap and the trimmed mean. J Multivar Anal 41:132–153
Helmers R, Jing B-Y, Qin G, Zhou W (2004) Saddlepoint approximations to the trimmed mean. Bernoulli 10:465–501
Lehmann EL (1983) Theory of point estimation. Wiley, New York
Maronna RA, Martin RD, Yohai VJ (2006) Robust statistics: theory and methods. Wiley, New York
Mendeleev DI (1895) Course of work on the renewal of prototypes or standard measures of lengths and weights (in russian). Vremennik Glavnoi Palaty Mer i Vesov, 2:157–185. Reedited in 1950: Collected Writings (Socheneniya), Izdat. Akad. Nauk, SSSR, Leningrad-Moscow 22:175–213
R Development Core Team (2013), R: A language and environment for statistical computing. R Foundation for Statistical Computing. Viena, Austria. http://www.R-project.org
Ronchetti E (1997) Robust inference by influence functions. J Stat Plan Infer 57:59–72
Serfling RJ (1980) Approximation theorems of mathematical statistics. Wiley, New York
Staudte RG, Sheather AJ (1990) Robust estimation and testing. Wiley, New York
Stigler SM (1973) The asymptotic distribution of the trimmed mean. Ann Stat 1:472–477
Stigler SM (1976) The anonymous professor Gergonne. Hist Math 3:71–74
Tukey JW (1962) The future of data analysis. Ann Math Stat 33:1–67
Tukey JW, McLaughlin DH (1963) Less vulnerable confidence and significance procedures for location based on a single sample: trimming/winsorization 1. Sankhya A 25:331–352
Withers CS (1983) Expansions for the distribution and quantiles of a regular functional of the empirical distribution with applications to nonparametric confidence intervals. Ann Stat 11:577–587
Acknowledgments
The author is grateful to the Referee for kind and professional remarks. This work is partially supported by Grant MTM 2012-33740 from Ministerio de Economía y Competitividad (Spain).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
García-Pérez, A. A von Mises approximation to the small sample distribution of the trimmed mean. Metrika 79, 369–388 (2016). https://doi.org/10.1007/s00184-015-0559-3
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00184-015-0559-3