Vere-Jones, D & Davies, R.B. (1966) ** A statistical survey of earthquakes
in the main seismic region of New Zealand: Part 2 - Time series analyses.**
* N.Z. J. Geol.
Geophys.* **9** 251-284
(scanned
copy - 2933kb)

Time series analyses are carried out on earthquake data from the main seismic region of New Zealand for the years 1942-61. Origin times only are considered, the energies and exact positions of the shocks being largely ignored. The relevant statistical theory for the first and second order properties of the process is described, and simple probability models for earthquake occurrence are put forward. On the basis of these results, the data are examined for periodic and grouping effects. No significant periodic effects are found, either among the shallow shocks (depths up to 100 km) or among the deep shocks (depths 100 km or greater). Both components show strong evidence of grouping, and several alternative models to describe this effect are put forward and compared.

Mertz, D. B. & Davies, R.B. (1968) ** Cannibalism of the pupal stage by adult
flour beetles: an experiment and a stochastic model.**
* Biometrics* **24** 247-275

This paper describes an experiment involving adult flour beetles (genetic strain
cIV-a of * Tribolium castaneum*) cannibalizing their own pupae and proposes a new
stochastic model for this form of cannibalism. The model differs from earlier
stochastic models of predation in that it is based on the hypothesis of predator
satiation, and
it gives a considerably better fit to the experimental data than do the earlier ones.
In the experiment the percentage survival of pupae increased rapidly as the pupal
density increased. This corresponds to an earlier finding that in self-limiting
populations of cIV-a, large pupal populations are apt to be followed in time by sudden
increases in adult numbers (i.e. outbreaks). The model shows that this would be
expected
for satiable adult predators. However, for insatiable predators, outbreaks
would be unlikely. Apparently, cIV-a adults behave as if their appetites for pupae
were satiable, but for most other *Tribolium* populations 'satiation' seems to be
less
important and outbreaks are uncommon.

Davies, R.B. (1969) ** Beta-optimal tests and an application to the summary evaluation of independent
experiments.**
* J. Roy. Statist. Soc.* B **31** 524-538.

A concept of optimality of a test, based on the speed with which its power function reaches a pre-assigned value, is introduced and conditions for a test to have this property are considered. The concept is applied to the weighted combination of independent normal test statistics; tests for the presence of effects and for the equality of effects are deduced.

Davies, R.B. (1973) ** Asymptotic inference in stationary Gaussian
time-series.**
* Adv. Appl. Prob.*
**5** 469-497.

Conditions are given for the family of distributions of a stationary,
discrete-time, Gaussian, vector-valued time-series with covariance structure
given up to a finite number of parameters to satisfy the asymptotic
differentiability conditions introduced by Le Cam (1969).
(scanned copy - 2156kb)

ASYMPTOTIC INFERENCE; CONTIGUITY; FOURIER SERIES; PERIODOGRAM;
SPECTRUM; STATIONARY GAUSSIAN TIME-SERIES; TOEPLITZ MATRIX

Davies, R.B. (1973) ** Numerical inversion of a characteristic
function.**
* Biometrika*
** 60** 415-417.

A method is described for finding a bound on the error when a version of the usual
characteristic function inversion formula is evaluated by numerical integration.
The method
is applied to the calculation of the distribution function of a quadratic
form in normal
random variables.
(scanned copy
- 317kb)

*
Some key words*: Numerical inversion of characteristic function; Quadratic form in normal variables;
Trapezoidal rule.

Davies, R.B. & Hutton, Bruce (1975) ** The effect of error in the independent variables in linear
regression.**
* Biometrika* ** 62** 383-391.

Suppose that the independent variables in a linear regression are subject to error. This
paper is concerned with the bias introduced into the least squares estimators by these errors,
first when they are regarded as fixed and second when they are regarded as random. Simple
criteria are introduced for deciding whether the bias is likely to be serious. The paper also
considers the effect of the occasional large error in either the dependent or independent
variables. (scanned copy -
1225kb)

*
Some key words*: Errors in data matrix; Least squares; Linear regression; Singularity of
matrix.

Davies, R.B. (1977) ** Testing the hypothesis that a point process is
Poisson.** * Adv. Appl. Prob.*
** 9** 724-746.

The testing of the hypothesis that a point process is Poisson against a one-dimensional
alternative is considered. The locally optimal test statistic is
expressed as an infinite series of uncorrelated terms. These terms are shown to
be asymptotically equivalent to terms based on the various orders of cumulant
spectra. The efficiency of tests based on partial sums of these terms is found.
(scanned copy - 1760kb)

ASYMPTOTIC EFFICIENCY: CUMULANT: CUMULANT SPECTRA; LOCAL EFFICIENCY LOCAL OPTIMALlTY;
PERIODOGRAM: POINT PROCESS: POISSON PROCESS

Davies, R.B. (1977) ** Hypothesis testing when a nuisance parameter is present only under the
alternative.**
* Biometrika* ** 64** 247-254.

Suppose that the distribution of a random variable representing the outcome of an
experiment depends on two parameters \xi and \theta and that we wish to test the hypothesis
\xi = 0 against
the alternative \xi > 0. If the distribution does not depend on \theta when \xi = 0, standard
asymptotic methods such as likelihood ratio testing or *C*(\alpha) testing are not directly
applicable. However, these methods may, under appropriate conditions, be used to reduce the problem to one
involving inference from a Gaussian process. This simplified problem is examined and a test
which may be derived as a likelihood ratio test or from the union-intersection principle is
introduced. Approximate expressions for the significance level and power are obtained.
(scanned copy
- 1067kb)

*
Some key words*: *C*-alpha. test; Hypothesis testing; Likelihood ratio test; Maximum of
Gaussian process ; Simple hypothesis ; Union-intersection principle.

Davies, R.B. (1985) **Asymptotic inference when the amount of information is
random.**
* Proceedings of the Berkeley Conference in Honor of Jerzy Neyman and Jack
Kiefer*, Volume 2. Eds L.M. LeCam & R.A. Olsen. 841-864. Wadsworth, Belmont.

Some results of asymptotic statistical decision theory are extended to allow for the
situation in which the Fisher information
should be treated as random. They are applied to parameter estimation
and hypothesis testing for the supercritical Galton-Watson
process and to sequential analysis.

*
Key Words and Phrases*: asymptotic inference, conditional inference, contiguity, curved exponential family, Galton-Watson process,
information matrix, sequential estimation, stochastic process estimation.
(scanned copy -
2394kb)

AMS 1980 Subject Classifications: primary 62F12; secondary 62FO5,
62L12, 62MO5, 62MO9.

Davies, R.B. & Harte, D.S. (1987) ** Tests for Hurst
effect.** * Biometrika*
** 74** 95-101.

We consider the power of tests for distinguishing between fractional Gaussian noise
and white noise of a first-order autoregressive process. Our tests are based on the
beta-optimal principle (Davies, 1969), local optimality and the rescaled range test.
(scanned copy - 868kb)

*
Some key words*: Autoregressive process; Beta-optimal test; Fractional Gaussian noise;
Hydrological series; Locally optimal test; Long-term dependence; Rescaled range;
Self-similar process; Simulation.

Davies, R.B. (1987) ** Hypothesis testing when a
nuisance parameter is present only under the
alternative.**
* Biometrika* ** 74** 33-43.

We wish to test a simple hypothesis against a family of alternatives indexed by a
one-dimensional parameter, \theta. We use a test derived from the corresponding family of
test statistics appropriate for the case when \theta is given. Davies (1977) introduced this
problem when these test statistics had normal distributions. The present paper considers
the case when their distribution is chi-squared. The results are applied to the detection
of a discrete frequency component of unknown frequency in a time series. In addition
quick methods for finding approximate significance probabilities are given for both the
normal and chi-squared cases and applied to the two-phase regression problem in the
normal case. (scanned copy
- 1313kb)

*
Some key words*: Chi-squared process; Frequency component; Hypothesis test; Maximum; Nuisance
parameter; Quick test; Two-phase regression; Time series; Up crossing.

Davies, R.B. (2001) ** Integrated processes and the discrete cosine
transform.**
*Probability, statistics and seismology, A Festschrift for David Vere-Jones.*
Ed D.J.Daley*. J. Appl. Probab.* Special Volume **38A**. 105-121.

A time-series
consisting of white noise plus Brownian motion sampled at
equal intervals of time is exactly orthogonalised by a discrete cosine
transform (DCT-II). This paper explores the properties of a version of spectral
analysis based on the discrete cosine transform and its use in distinguishing
between a stationary time-series and an integrated (unit root) time-series.
(preprint
- 298kb)

*Keywords*: Beta-optimal test; Brownian motion; DCT-II;
discrete cosine transform; integrated process; random walk; spectrum;
time-series; unit root.

Davies, R.B. (2002) **Hypothesis testing when a nuisance parameter is present only under the
alternative - linear model case.** *Biometrika* **89** 484-489.

The results of Davies (1977, 1987) are extended to a linear model situation with
unknown residual variance. (preprint
- 166kb)

*Some key words*: Change point; *F*-process; Frequency component;
Hypothesis test; Nuisance parameter; *t*-process; Two-phase regression; Up-crossing.

Davies, R.B., Withers, C.S. & Nadarajah, S. (2011) **Confidence intervals in a
regression with both linear and non-linear terms**. *Electron. J. Statist.*
**5**, 603-618.

We present a simple way for calculating confidence intervals for a class of
scalar functions of the parameters in least squares estimation when there are
linear together with a small number of non-linear terms. We do not assume
normality. (Open
access - 1106KB)

*Keywords*: Confidence interval; estimation; optimization; two-phase
regression

Statistics Research Associates Limited, PO Box 12-649, Thorndon, Wellington, New Zealand