Sankhya B (November 2011) 73:211–217
DOI 10.1007/s13571-011-0018-8
A new look at search designs
K. Chatterjee · C. Koukouvinos
Received: 30 August 2010 / Revised: 18 November 2010 / Accepted: 7 January 2011 /
Published online: 1 October 2011
© Indian Statistical Institute 2011
Abstract Search designs were first introduced by Srivastava (1975) in order to
address the problem of identifying a set of possibly non-negligible interactions.
In this paper search designs for 2-level symmetric factorials are considered
under the effect hierarchy principle and under a special structure. We provide
designs for searching at most three two-factor interactions. Using Hadamard
matrices, constructions of such designs are presented.
Keywords Two level factorials · Search designs · Tree structure
AMS 2000 Subject Classifications Primary 62K15; Secondary 05B20
1 Introduction
Considering the hierarchy of factorial effects Srivastava and Hveberg (1992)
observed a tree structure in the factorial effects when analyzing data that arises
from behavioral sciences. It is logical that the non-negligibility of interactions
F
1
F
2
and F
1
F
3
implies the non-negligibility of F
2
F
3
for a factorial experiment
involving factors F
1
, F
2
, and F
3
. Chatterjee and Fang (2001) constructed
K. Chatterjee
Department of Statistics, Visva-Bharati University,
Santiniketan, West Bengal, India
C. Koukouvinos (B )
Department of Mathematics, National Technical University of Athens,
Zografou 15773, Athens, Greece
e-mail: ckoukouv@math.ntua.gr