Technometrics Highlights
Technometrics Celebrates 50 Years
David M. Steinberg, Editor, Technometrics
T
he February 2008 issue begins vol- bit failure cross section via nonparametric
ume 50 of Technometrics. The journal penalized max imum likelihood.
was launched in 1959 with J. Stuart The next article, by Anna E. Vine, Susan
Hunter as the editor and quickly achieved M. Lewis, Angela M. Dean, and David
standing as a leading journal. We will cele- Brunson, is titled “A Critical Assessment
brate this jubilee year with a number of spe- of Two-Stage Group Screening Through
cial features. This issue begins with a history Industrial Experimentation.” Screening is
of the founding of Technometrics. Of the the process of sifting through a set of fac-
many who contributed to the process, tors via experimentation to determine the
Hunter and George Box stand out for their few important factors that have a substan-
key roles in establishing Technometrics and tial effect on a response. Two-stage group
George Box Stu Hunter
making it one of the top statistics journals. In screening begins with a factorial design in
the process of preparing this article, Søren which some factors are fully confounded, exist, a computer search of regular graph
Bisgaard and I had the pleasure of visiting followed by a second stage in which addi- designs is conducted and finds optimal
Hunter and Box and hearing their accounts tional runs help distinguish which indi- supersaturated designs. The authors give
of how the journal came to be. vidual factors are important. design tables and notes on the construction
We will be taking a forward look in the This paper gives the first description of of saturated designs.
coming special features, highlighting topics practical aspects involved in running a two- A novel approach to multiple infer-
likely to be important in the years ahead. stage group screening experiment for inves- ence in factorial experiments is pre-
The May issue will feature a panel discus- tigating interactions. Issues involved in the sented by Maria Tripolski Kimel, Yoav
sion on the future of industrial statistics. design and analysis of such an experiment Benjamini, and David Steinberg in “The
Subsequent issues will include invited arti- are discussed in the context of a study run False Discovery Rate for Multiple Testing
cles about statistics and computing and the at Jaguar Cars on cold start optimization. in Factorial Experiments.” The false dis-
analysis of massive data sets. The analysis of this experiment gives insight covery rate (FDR) is the expected propor-
In addition to the history of into how group screening works in practice tion of inert effects among those declared
Technometrics, this issue includes nine and how the factorial effects of the indi- active. This measure has proven valuable
research articles. The first is an interesting vidual factors relate to those of the grouped in numerous multiple inference problems,
case study of a problem arising in testing factors. Elicitation of information from and this paper shows it is also a relevant
the reliability of silicon chips at the Los subject specialists, choice of factor groups, quantity to control in moderate and large
Alamos National Laboratory, “Evaluating and selection of designs for two-stage group factorial experiments where the number
Experiments for Estimating the Bit Failure screening are discussed. Through the analy- of potential effects can be large. The paper
Cross Section of Semiconductors Using sis of the experimental data, it is shown that shows how to combine FDR control with
a Colored Spectrum Neutron Beam” by the process of group screening can provide popular methods for estimating contrast
Nicolas W. Hengartner, Sarah E. Michalak, an efficient method of detecting interac- standard error in unreplicated experiments.
Bruce E. Takala, and Stephen A. Wender. tions among large numbers of factors. Simulations of unreplicated 16- and 32-run
The paper considers estimation of the fre- Supersaturated designs are another tool experiments illustrate the improvements in
quency of bit failures, which result when that can be useful in experimental settings power that can be obtained by controlling
cosmic ray–induced neutrons cause bit flips where there are many candidate factors and the FDR. The ideas also are applied to the
in silicon-based electronic devices such as limited resources and only a few of the fac- analysis of a 128-run experiment with rep-
computer memory. The bit failure cross tors are expected to be of major importance. lication, in which the FDR identifies many
section of a silicon-based electronic device In “New E(s²)-Optimal Supersaturated more active effects than other methods.
describes the probability of causing a bit flip Designs Constructed From Incomplete Online process monitoring often
as a function of neutron energy. This paper Block Designs,” Nam-Ky Nguyen and involves autocorrelated data, and changes
discusses estimation of the bit failure cross Ching-Shui Cheng present a new con- in the process can affect the correlation
section based on neutron beam testing. It is struction method for two-level supersatu- structure. Hyunyoung Choi, Hernando
shown that this is a severely ill-posed inverse rated designs. They derive a lower bound Ombao, and Bonnie Ray present two new
problem. The authors present a general of the E(s²) criterion and show it is attained spectral-based methods for detection of
methodology for evaluating, prior to the by supersaturated designs constructed from such changes in “Sequential Change-Point
experiment, the extent to which the experi- balanced incomplete block designs. Where Detection Methods for Nonstationary Time
mental protocol permits estimation of the balanced incomplete block designs do not Series.” Their methods are based on the idea
FEBRUARY 2008 AMSTAT NEWS 11
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64  |  Page 65  |  Page 66  |  Page 67  |  Page 68