Abstract
A classical statistical problem is the comparison of two normal samples with an emphasis on inference about the difference of their means. When both variances are unknown, this is the notorious Behrens–Fisher problem; however in some applications one sample’s variance can be considered as given. Closed-form confidence intervals for the difference between two means are derived in both cases. These easily computable procedures are based on modern higher-order asymptotic statistical methods and are quite accurate.
Similar content being viewed by others
References
O. E. Barndorff-Nielsen, “Inference on Full or Partial Parameters Based on the Standardized Signed Log Likelihood Ratio”, Biometrika 73, 307–322 (1986).
O. E. Barndorff-Nielsen and D. R. Cox, Inference and Asymptotics (Chapman and Hall, London, 1994).
M. Bédart, D. A. S. Fraser, and A. C. M. Wong, “Higher Accuracy for Bayesian and Frequentist Inference: Large Sample Theory for Small Sample Likelihood”, Statist. Sci. 22, 301–121 (2007).
A. Belloni and G. Didier,. “On the Behrens–Fisher Problem: A Globally Convergent Algorithm and a Finite-Sample Study of the Wald, LR, and LM Tests”, Ann. Statist. 36, 2377–2408 (2008).
M.-L. G. Bout, S. Hosten, and D. St. P. Richards, “Counting and Locating the Solutions of Polynomials Systems of Maximum Likelihood Equations II: The Behrens–Fisher Problem”, Statist. Sinica 17, 1343–1354 (2007).
A. R. Brazzale, A. C. Davison, and N. Reid, Applied Asymptotics: Case Studies in Small Sample Statistics (Cambridge University Press, Cambridge, UK, 2007).
R. J. Cook and V. T. Farewell, “Behrens–Fisher Problem”, in Encyclopedia of Biostatistics, Ed. by P. Armitage and T. Colton (Wiley, Chichester, 1998), Vol. 1, pp. 287–288.
M. Drton, “Multiple Solutions to the Likelihood Equations in the Behrens–Fisher Problem”, Statist. and Probab. Lett. 78, 3288–3293 (2008).
D. A. S. Fraser, N. Reid, and J. Wu, “A Simple General Formula for Tail Probabilities for Frequentist and Bayesian Inference”, Biometrika 86, 249–264 (1999).
D. A. S. Fraser and N. Reid, “StrongMatching of Frequentist and Bayesian Parametric Inference”, J. Statist. Plann. Inf. 103, 263–285 (2002).
D. A. S. Fraser, N. Reid, E. Marras, and G. Y. Yi, “Default Priors for Frequentist and Bayesian Inference”, J. R. Statist. Soc. B 72, 631–654 (2010).
D. A. S. Fraser, A. Wong, and J. Sun, “Three Enigmatic Examples and Inference from Likelihood”, Can. J. Statist. 37, 161–181 (2009).
M. Ghosh and Y.-H. Kim, “The Behrens–Fisher Problem Revisited: A Bayes–Frequentist Synthesis”, Can. J. Statist. 29, 5–17 (2001).
J. L. Jensen, “The Modified Signed Likelihood Statistic and Saddle Point Approximations”, Biometrika 79, 693–703 (1992).
E. L. Lehmann and J. P. Romano, Testing Statistical Hypotheses, 3rd ed. (Springer, New York, 2003).
Yu. V. Linnik, Statistical Problems with Nuisance Parameters (Amer Math Soc: Providence, RI, 1967).
G. K. Robinson, “Behrens–Fisher Problem”, in Encyclopedia of Statistical Sciences, Ed. by S. Kotz and N. L. Johnson (Wiley, New York, 1982), Vol. 1, pp. 205–209.
V. I. Pagurova, “On Comparison of Means of Two Normal Samples”, Theory Probab. Appl. 13, 527–534 (1968).
A. L. Rukhin, “Assessing Compatibility of Two Laboratories: Formulations as a Statistical Testing Problem”, Metrologia 50, 49–59 (2013).
Y. She, A. Wong, and X. Zhou, “Revisit the Two Sample t-Test With a Known Ratio of Variances”, Open J. Statist. 1, 151–156 (2011).
I. Skovgaard, “An Explicit Large-Deviation Approximation to One-Parameter Tests”, Bernoulli 2, 145–165 (1996).
D. A. Sprott and V. T. Farewell, “The Difference Between Two Normal Means”, Amer. Statist. 47, 126–128 (1993).
A. C. M. Wong and Y. Y. Wu, “Likelihood Analysis for the Difference in Means of Two Independent Normal Distributions with One Variance Known”, J. Statist. Research 42, 17–35 (2008).
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Rukhin, A. Higher order accurate procedures to compare two normal populations. Math. Meth. Stat. 24, 292–308 (2015). https://doi.org/10.3103/S1066530715040043
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.3103/S1066530715040043
Keywords
- asymptotic expansions
- Behrens–Fisher problem
- confidence intervals
- higher order inference
- strong matching priors