Earl Duncan

Earl DuncanQualifications:

PhD (Statistics), QUT.
B AppSc (Hons: Statistics), QUT.
B Math (Statistics), QUT.
B Bus (Accounting, International Business), QUT.

Research Interests

  • Spatial modelling (e.g. disease mapping)
  • Spatio-temporal modelling
  • Data visualisation
  • Bayesian mixture models and development of relabelling algorithms to reverse label switching

Project Description

Evidence suggests that regular cancer screening is effective at detecting cancer in its early stages and thereby increases patients’ chances of survival.  However, research also suggests that cancer screening services (breast, cervical, and colorectal) are under-utilised.  In an attempt to understand why service utilisation is lower than advised and heterogeneous across geographic regions and years, my research focuses on identifying unusual spatio-temporal trends using Bayesian statistical models.

Recent research interests have been directed at comparing algorithms designed to reverse label switching – an inherent problem to Bayesian mixture models.  Many algorithms have been developed in the last 15 years.  My aim is twofold: firstly to compare and contrast these algorithms in terms of computational efficiency and accuracy, and secondly, extend the recently developed algorithm ‘Zswitch’. This will be achieved through simulation studies.



  1. Pokorny, M. R., M. de Rooij, E. Duncan, F. H. Schröder, R. Parkinson, J. O. Barentsz, and L. C. Thompson. 2014.  Prospective Study of Diagnostic Accuracy Comparing Prostate Cancer Detection by Transrectal Ultrasound–Guided Biopsy Versus Magnetic Resonance (MR) Imaging with Subsequent MR-guided Biopsy in Men Without Previous Prostate Biopsies. European Urology 66 (1): 22-29.
  2. Duncan, E., K. Mengersen, and N. White. 2015. “Comparison of Relabelling Algorithms Applied to Bayesian Mixture Models: A Simulation Study.” Poster presented at Bayes on the Beach 2015, Gold Coast, Australia, December 7-9. doi: 10.6084/m9.figshare.2007702.
  3. Duncan, E. W., N. M. White, and K. Mengersen.  2016.  Bayesian spatiotemporal modelling for identifying unusual and unstable trends in mammography utilisation. BMJ Open 6 (5): p.e010253. doi: 10.1136/bmjopen-2015-010253.
  4. Cramb, S. M., E. W. Duncan, N. M. White, P. D. Baade, and K. L. Mengersen.  2016. Spatial Modelling Methods. Brisbane: Cancer Council Queensland and Queensland University of Technology (QUT).

Under review:

  1. Duncan, E. W., N. M. White, and K. Mengersen. Spatial smoothing in Bayesian models: a comparison of weights matrix specifications and their impact on inference.  Submitted to the International Journal of Health Geographics May 2017.
  2. Duncan, E. W., N. M. White, and K. Mengersen. Improved Bayesian methods for identifying aberrant temporal trends in spatio-temporal data with application to mammography screening services.  Submitted to PLOS ONE, May 2017.
  3. Mengersen, K., E. W. Duncan, N. M. White, J. Arbel, and C. Alston-Knox.  “Applications in Industry”.  In Handbook of Mixture Analysis, edited by G. Celeux, S. Früwirth-Schnatter, and C. P. Robert.  Chapman & Hall/CRC Handbooks of Modern Statistical Methods.  Forthcoming July 2018.

Contact Details:






3 thoughts on “Earl Duncan

  1. Pingback: Brag Meeting 14th May 2015 | Bayesian Research & Applications Group

  2. Pingback: Brag Meeting 28th May 2015 | Bayesian Research & Applications Group

  3. Pingback: Brag Meeting 20th July 2017 | Bayesian Research & Applications Group

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s