Staff, Students, and Associates

Dr Ken Newman
Principal Researcher in Statistical Methodology

PhD

Biomathematics and Statistics Scotland
JCMB, The King's Buildings,
Peter Guthrie Tait Road,
EDINBURGH, EH9 3FD, Scotland, UK

Tel: +44 (0)7899 675 993
Publications:
Detailed Website:
View detailed website »

Statistical Methodology Theme Group Meetings

  1. Luigi Spezia (31 July 2020): Markov switching model with time varying covariates
  2. Graham Horgan (6 August 2020): Multiple weight time series with missing data
  3. Nick Schurch (20 August 2020): Longterm memory network nodes for time series modelling
  4. Adam Butler (27 August 2020): To impute or not to impute
  5. Jackie Potts (3 September 2020): Multiple testing
  6. Ken Newman (10 September 2020): Questions about "spatial" covariance matrices
  7. Zhou Fang (24 September 2020): Modelling aphids
  8. Adrian Roberts (15 October 2020): Variance estimation with a pooled sample?
  9. Adam Butler ((29 October 2020); TBD

Research Interests and Areas of Application

I coordinate BioSS research in the Statistical Methodology Research Theme. In this Theme, we develop and advance statistical methods for data collection and analysis to help answer challenging and wide-ranging scientific questions related to agriculture and the rural economy, the environment, food and health. Some methods and areas of current focus include:

  • modelling spatial-temporal processes, e.g., concentrations of pollutants in a river network over time, using tools like multivariate hidden Markov models;
  • analysis and modelling of compositional data; that is, data carrying relative information, e.g., concentrations of chemical elements in soil samples, relative species abundances, time use patterns;
  • design and analysis of ordinal data arising from socio-economic applications, e.g., using latent class models and extensions of ordinal regression;
  • developing statistical emulators that can be used as approximations to complex, computationally demanding mechanistic and deterministic models, e.g., predicting nitrate levels in water and soil.

My main area of personal research interest is state-space models in particular, and hierarchical models in general. This includes developing procedures for combining multiple data sets that provide separate information about different processes, such as (in a population dynamics context) survival, reproduction, and movement, the process of combination leading to what is sometimes referred to as an integrated analysis. Given such models, I am also interested in developing decision-support tools for domain-specific researchers and resource managers to use such models to do things like predict the long-term effects of management actions on some resource.

Statistical analyses are limited by the quality of data available, hence I have a long-standing interest in statistical sampling theory and methods. This includes methods for sampling spatially-referenced populations and the design and implementation of long-term monitoring programs for determining status and trends in ecological systems.

In general I'm very interested in collaborative interdisciplinary science projects where sound statistical practices are central and where statistical methodology needs to be tailored or expanded upon to better achieve project aims. My strongest application area is population modelling, including heavily managed populations of economic importance. I am also interested in a much broader sphere of activities including the development of hierarchical models for other dynamic natural systems such as agricultural crop growth and yield, plant growth, and hydrology.