This volume contains papers presented at the conference "Statistical Anal-
ysis of Measurement Error Models and Applications," held June 10 to June
16, 1989, at Humboldt State University, Arcata, California. The confer-
ence was a Summer Research Conference sponsored jointly by the American
Mathematical Society, the Institute of Mathematical Statistics, and the Soci-
ety for Industrial and Applied Mathematics, and supported by a grant from
the National Science Foundation. The organizing committee was composed
of Raymond J. Carroll, Co-Chair; Wayne A. Fuller, Co-Chair; Leon J. Gieser,
Co-Chair; G. W. Stewart and J. H. Ware.
Measurement error models describe functional relationships among vari-
ables observed subject to random errors of measurement. Examples include
linear and nonlinear errors-in-variables regression models, calibration and in-
verse regression models, factor analysis models, latent structure models, and
simultaneous equation models. Such models are used in medicine, the life
sciences, econometrics, chemometrics, geology, sample surveys, time series,
and other disciplines. Although the problem of estimating the parameters of
such models exists in most scientific fields, there are few sources that treat
measurement error models as an area of statistical methodology. This volume
is designed to add to that small number.
The conference was organized to achieve three main goals. First, it was
desired to bring together researchers working on measurement error models
in a variety of disciplines in order to improve communication. Interaction
between disciplines has been hampered by different model formulations, ter-
minology, and notation. Second, it was felt that a dialogue with numerical
analysts could lead to cooperation in constructing statistical software. Third,
the conference offered a chance to take stock of recent research developments
in nonlinear measurement error models.
To fulfill these goals, a broad mix of theoretical statisticians, applied statis-
ticians, and numerical analysts from academia, government and industry
were invited. Because of the range of participants, a variety of topics were
considered at the conference.
We have organized the papers into four broad groups. The first group
contains papers treating general aspects of the measurement error problem.