Small sample confidence intervals

In this article we present a technique for constructing oneor two-sided confidence intervals, which are second-order correct in terms of coverage, for either parametric or nonparametric problems. The construction is valid in the presence of nuisance parameters. The situation we consider is this: there are p parameters and we want a confidence interval for some function of them, possibly one of the parameters itself. The p parameters are estimated by M-estimates, which means they are obtained as a solution of a system of equations. Maximum likelihood estimates are included as a special case. The essential intermediate result, given in Equation (3.4), says that the estimated parameter of interest, 0, can be written as a mean, up to order op(l/ V/ ). The representation (3.4) is attained when 0 = 0(a) is a smooth function of the parameters q and aj is the solution of a well-behaved system of equations. We avoid the use of pivots and strive to obtain accurate coverage. Confidence intervals are constructed from a series of tests for the natural parameter of a one-parameter exponential family. We use the Lugannani and Rice (1980) tail area approximation to calculate bootstrap P values for the test statistic.

[1]  Gary Simon.  Computer Simulation Swindles, with Applications to Estimates of Location and Dispersion , 1976 .

[2]  B. Efron.  Nonparametric standard errors and confidence intervals , 1981 .

[3]  J. Ghosh,et al.  ON THE VALIDITY OF THE FORMAL EDGEWORTH EXPANSION , 1978 .

[4]  S. Kullback,et al.  Information Theory and Statistics , 1960 .

[5]  Robert Tibshirani.  Bootstrap Confidence Intervals , 1984 .

[6]  D. F. Andrews,et al.  Robust Estimates of Location: Survey and Advances. , 1975 .

[7]  M. Tingley.  Robust confidence intervals. , 1986 .

[8]  C. Field,et al.  Small-sample asymptotic distributions of M-estimators of location , 1982 .

[9]  C. J. Lawrence.  Robust estimates of location : survey and advances , 1975 .

[10]  Anthony C. Davison,et al.  Saddlepoint approximations in resampling methods , 1988 .

[11]  J. Kent.  Robust properties of likelihood ratio tests , 1982 .

[12]  P. Hall.  Theoretical Comparison of Bootstrap Confidence Intervals , 1988 .

[13]  H. Daniels.  Saddlepoint Approximations in Statistics , 1954 .

[14]  S. Rice,et al.  Saddle point approximation for the distribution of the sum of independent random variables , 1980, Advances in Applied Probability.

[15]  Chris Field,et al.  Small Sample Asymptotic Expansions for Multivariate $M$-Estimates , 1982 .

[16]  H. E. Daniels,et al.  Tail Probability Approximations , 1987 .

[17]  B. Efron.  Better Bootstrap Confidence Intervals , 1987 .

[18]  I. Good.  THE MULTIVARIATE SADDLEPOINT METHOD AND CHI-SQUARED FOR THE MULTINOMIAL DISTRIBUTION , 1961 .

[19]  B. Gnedenko,et al.  Limit Distributions for Sums of Independent Random Variables , 1955 .

[20]  E. Ronchetti,et al.  A tail area influence function and its application to testing , 1985 .

[21]  H. Daniels.  Saddlepoint approximations for estimating equations , 1983 .

[22]  Solomon Kullback,et al.  Information Theory and Statistics , 1960 .