1. INTRODUCTION
Between the 16th and 19th centuries, astronomy and
statistics were closely allied fields. Many of the foundations of
mathematical statistics were laid by astronomers such as Tycho Brahe,
Galileo, Tobias Mayer and Adrien Legendre
Stigler 1986). But this
relationship weakened during the late 19th century, as statistics
turned to applications in the social sciences and industry, astronomy
reaped benefits from mathematical physics. A by-product of this shift
is that most astronomers are trained by physicists and receive little or
no formal education in statistics. Most astronomers are only vaguely
aware of the tremendous advances in statistical theory and practice of
the last few decades. Similarly, with the notable exception of galaxy
clustering studies by Jerzy Neyman and Elizabeth Scott in the
1950-60s, statisticians became unaware of the tremendous developments
in astronomy.
Mutual interest in astrostatistics has reemerged during the past
decade. The comparison of astronomical data to astrophysical
questions is becoming increasingly complex, outpacing the capabilities
of traditional statistical methods. About 500 astronomical papers
annually have `statistics' or `statistical' in their abstracts, yet
they rarely refer to contemporary statistical texts or monographs for
methodological guidance. Statistical procedures implemented in
Numerical Recipes
(Press et al.
1992) are used on a daily basis.
Recent cross-disciplinary efforts in astrostatistics have produced
valuable resources. A number of conferences have been held in Europe
(e.g., Rolfe 1983;
Jaschek & Murtagh 1990;
Subba Rao 1997) and the
U.S. (Feigelson & Babu
1992;
Babu & Feigelson
1997), astrostatistical
sessions at large meetings are being organized, an introductory
monograph on astrostatistics has emerged
(Babu & Feigelson
1996), and
the Statistical Consulting Center for Astronomy is active
(Feigelson et al.
1995:
http://www.stat.psu.edu/scca).
A monograph on
multivariate data analysis, with FORTRAN codes and bibliography of
astronomical applications, is very relevant to the issues discussed
here (Murtagh & Heck
1987).