Special Issue on Alternatives to Multiple Comparisons Journal of Probability and Statistics

Journal of Probability and Statistics
Special Issue on
Alternatives to Multiple Comparisons
Call for Papers
Many modern applications of statistics involve problems that
in the classical (frequentist) framework involve sequences
of tests of hypotheses. A large amount of research in this
area has provided a good understanding of the problems
associated with these types of approaches, but fundamental
issues still exist. These issues include how the sequence of
null and alternative hypotheses should be specified, how the
error rates should be controlled, and what the relationship
is between the specified error rate and the true error rate. In
many modern applications, such as in applications to genetic
studies, the number of tests can be very large, and the issues
discussed above can become more complicated. Because
of these issues, many practitioners find that the results
of these analyses are often confusing and can sometimes
even seem paradoxical. Some practitioners in fields that use
statistics have even advocated ending the use of multiple
comparison methods. This issue calls for resolution since the
interpretation of data varies markedly with whether multiple
comparison procedures are applied and with the choice of
the multiple comparison procedure.
New research in this area has begun to focus on alternative
methods within the classical, Bayesian, evidential, and
information-theoretic frameworks, which avoid the complicated issues associated with multiple testing problems.
Some of these methods avoid pairwise comparisons by using
simultaneous measures of confidence, posterior probabilities, and similar methods. Other approaches use confidence
measures, Bayesian posterior distributions, and informationtheoretic methods that enable pairwise comparisons to avoid
paradoxes by relaxing attempts to control error rates.
We invite authors to present original research articles
and review articles that will stimulate the continuing efforts
in developing new statistical methods. We are particularly
interested in manuscripts that present new methodologies
for problems usually associated with multiple comparisons
that avoid complications due to pairwise comparisons, the
control of error rates, and sequences of hypothesis tests.
Potential topics include, but are not limited to:
• The study of methodologies based on measures of
• Approaches based on evidential and information-
theoretic frameworks
• Bayesian approaches, particularly objective Bayesian
• The connection between Bayesian and alternative
• Issues dealing with false discovery rates
• The application of established alternative methodolo-
gies to large-scale problems such as those found in
marketing and genomics
Before submission authors should carefully read over the
journal’s Author Guidelines, which are located at http://www
.hindawi.com/journals/jps/guidelines.html. Prospective authors should submit an electronic copy of their complete
manuscript through the journal Manuscript Tracking System at http://mts.hindawi.com/ according to the following
Manuscript Due
July 1, 2011
First Round of Reviews
October 1, 2011
Publication Date
January 1, 2012
Lead Guest Editor
Alan M. Polansky, Division of Statistics, Northern Illinois
University, DeKalb, IL, USA; polansky@math.niu.edu
Guest Editors
David R. Bickel, Department of Biochemistry,
Microbiology, and Immunology, Ottawa Institute of
Systems Biology, University of Ottawa, Ottawa, ON,
Canada; dbickel@uottawa.ca
Hidetoshi Shimodaira, Department of Mathematical and
Computing Sciences, Tokyo Institute of Technology, Tokyo,
Japan; shimo@is.titech.ac.jp
confidence such as attained and observed confidence
Hindawi Publishing Corporation