Margaret Donald

Margaret Donald

Postdoctoral fellow at UNSW

Location
Sydney, Australia
Industry
Biotechnology
Current
  1. Unversity of New South Wales,
  2. Anne Clements & Associates
Previous
  1. Simpson Centre for Health Services Research,
  2. Statistical Revelations,
  3. University of Melbourne
Education
  1. Queensland University of Technology
76 connections

Join LinkedIn and access Margaret’s full profile. It’s free!

As a LinkedIn member, you’ll join 400 million other professionals who are sharing connections, ideas, and opportunities.

  • See who you know in common
  • Get introduced
  • Contact Margaret directly

Summary

Margaret is an experienced statistician who has worked on many projects in both health and environmental statistics. She is a highly experienced SAS programmer.

The research position at UNSW opens many possibilities which she is looking forward to exploring.

Experience

Publications

  • Bayesian Network for Risk of Diarrhoea Associated with the Use of Recycled Water

    Risk Analysis

    Estimating potential health risks associated with recycled (reused) water is highly complex given the multiple factors affecting water quality. We take a conceptual model, which represents the factors and pathways by which recycled water may pose a risk of contracting gastroenteritis, convert the conceptual model to a Bayesian net, and quantify the model using one expert's opinion. This allows us to make various predictions as to the risks posed under various scenarios. Bayesian nets provide an additional way of modeling the determinants of recycled water quality and elucidating their relative influence on a given disease outcome. The important contribution to Bayesian net methodology is that all model predictions, whether risk or relative risk estimates, are expressed as credible intervals.

    Authors:

  • Incorporating Parameter uncertainty into Quantitative Microbial Risk Assessment

    Journal of Water and Health

    Modern statistical models and computational methods can now incorporate uncertainty of the parameters used in Quantitative Microbial Risk Assessments (QMRA). Many QMRAs use Monte Carlo methods, but work from fixed estimates for means, variances and other parameters. We illustrate the ease of estimating all parameters contemporaneously with the risk assessment, incorporating all the parameter uncertainty arising from the experiments from which these parameters are estimated. A Bayesian approach is adopted, using Markov Chain Monte Carlo Gibbs sampling (MCMC) via the freely available software, WinBUGS. The method and its ease of implementation are illustrated by a case study that involves incorporating three disparate datasets into an MCMC framework. The probabilities of infection when the uncertainty associated with parameter estimation is incorporated into a QMRA are shown to be considerably more variable over various dose ranges than the analogous probabilities obtained when constants from the literature are simply ‘plugged’ in as is done in most QMRAs. Neglecting these sources of uncertainty may lead to erroneous decisions for public health and risk management.

    Authors:

  • Comparison of three-dimensional profiles over time

    Journal of Applied Statistics

    In this paper, we describe an analysis for data collected on a three-dimensional spatial lattice with treatments applied at the horizontal lattice points. Spatial correlation is accounted for using a conditional autoregressive model. Observations are defined as neighbours only if they are at the same depth. This allows the corresponding variance components to vary by depth. We use the Markov chain Monte Carlo method with block updating, together with Krylov subspace methods, for efficient estimation of the model. The method is applicable to both regular and irregular horizontal lattices and hence to data collected at any set of horizontal sites for a set of depths or heights, for example, water column or soil profile data. The model for the three-dimensional data is applied to agricultural trial data for five separate days taken roughly six months apart in order to determine possible relationships over time. The purpose of the trial is to determine a form of cropping that leads to less moist soils in the root zone and beyond. We estimate moisture for each date, depth and treatment accounting for spatial correlation and determine relationships of these and other parameters over time.

    Authors:

  • 3-D imaging and quantitative comparison of human dentitions and simulated bite marks

    International Journal of Legal Medicine

    This study presents a technique developed for 3-D imaging and quantitative comparison of human
    dentitions and simulated bite marks. A sample of 42 study models and the corresponding bites, made by the mage comparison of a 3-D dentition with a 3-D bite mark, eliminating distortion
    due to perspective as experienced in conventional photography. Cartesian co-ordinates of a series of landmarks were used to describe the dentitions and bite marks, and a matrix was created to compare all possible combinations of matches and non-matches using cross-validation techniques. An algorithm, which estimated the probability of a dentition matching its corresponding bite mark, was developed. A receiver operating characteristic graph illustrated the relationship between values for specificity and sensitivity. This graph also showed for this sample that 15% of non-matches could not be distinguished from the true match, translating to a 15% probability of falsely convicting an innocent person.

    Authors:
    • Margaret Donald,
    • Sherie Blackwell,
    • Ian Gordon,
    • J.G. Clement,
    • M. Yoshino,
    • R.V. Taylor,
    • C. L. Ogleby,
    • T. Tanijiri

  • The Medical Emergency Team system: A two hospital comparison

    Resuscitation

    To compare activity and outcomes of a mature Medical Emergency Team (MET) in two hospitals.

    Authors:

  • Potential for using soil particle size data to infer geological parent material in the Sydney Region

    Soil Research

    Ecological communities are more than assemblages of species. In assessing the presence of many ecological communities, interpretation of soil properties and associated parent material has become a definitive component under environmental legislation worldwide, and particularly in Australia. The hypothesis tested here is that the geological parent material of a soil sample can be determined from particle size fraction data of the Marshall soil texture diagram. Supervised
    statistical classifiers were built from data for four particle-size fractions from four soil landscape publications. These methods were modified by taking into account possible autocorrelation between samples from the same site. The soil samples could not be classified with certainty as being derived from Wianamatta Group Shale or Hawkesbury Sandstone parent material. The classification of alluvial/fluvial-derived soils was no better than chance alone. A good classifier using four-fraction compositional data could not be built to determine geological parent material. Hence, the three size fractions of the Marshall soil texture diagram are insufficient to determine the geological parent material of a soil sample.

    Authors:

  • Model Comparisons for RNA-Seq Data. in Proceedings of the 28th International Workshop on Statistical Modelling

    Istituto Polygrafico Europeo

    Various strategies for finding differentially expressed genes in high throughput genomic studies have been proposed. Using a set of paired patient data, we have explored some of the many possible models. Results differ dependent on whether the data are normalised or not, and on the method of normalisation selected. Models that allow for over-dispersion fitted the data better. There was little consistency in the declaration of differentially expressed genes between the various approaches.

    Authors:

  • METHODS FOR CONSTRUCTING UNCERTAINTY INTERVALS FOR QUERIES OF BAYESIAN NETS

    Australian and New Zealand Journal of Statistics

    In this paper we reconsider the issue of finding confidence or credible intervals for queries in a Bayesian Network. We focus on the situation in which the BN is based on discrete nodes and finite populations, and compare an earlier asymptotic approach with a simulation-based approach, together with two further alternatives based on a single simulate of the BN and using binomial confidence intervals: one based on a single simulate of the BN, and the other on expected population sizes and calculated probabilities. We suggest that any querying of a BN should
    produce a probability embedded in an uncertainty interval. Based on an investigation of two BN structures, the preferred method is the simulation method. However, both the single simulate method and exact probabilities methods may be useful and simpler to compute in many circumstances, and any method at all is more useful than none, when assessing a BN under development, or drawing conclusions in an expert system.

    Authors:

Languages

  • French

Skills

Education

  • Queensland University of Technology

    Doctor of Philosophy (Ph.D.), Statistics
    –

    Thesis: Using Bayesian methods for the estimation of uncertainty in complex statistical models

  • Macquarie University

    M.App. Stats, Applied Statistics
    –

    Project: Use of the Lognormal distribution for left-censored water quality data

  • Melbourne University

    Bachelor of Arts (BA) (Hons), Mathematics & Philosophy
    –

Interests

Groups

View Margaret’s full profile to...

  • See who you know in common
  • Get introduced
  • Contact Margaret directly

Not the Margaret you’re looking for? View more

Find a different Margaret Donald

Example: Margaret Donald

More professionals named Margaret Donald

People Also Viewed

LinkedIn members in Australia:

  1. a
  2. b
  3. c
  4. d
  5. e
  6. f
  7. g
  8. h
  9. i
  10. j
  11. k
  12. l
  13. m
  14. n
  15. o
  16. p
  17. q
  18. r
  19. s
  20. t
  21. u
  22. v
  23. w
  24. x
  25. y
  26. z
  27. more
  28. Browse members by country