Branch Meetings

Branch Meetings for 2017 Poster

Statistical Society of Australia Canberra Branch

Date:         Tuesday 28 February 2017

Times:  

5.15pm        Refreshments : Room 1175A, John Dedman Building, ANU

Campus http://campusmap.anu.edu.au/displaymap.asp?grid=gh32 [Building 27]

 

6.00pm        Presentation by Prof. Ross Maller : College of Business and Economics, ANU [Building 26C]  Allan Barton Forum, Level 2, Room 248 http://www.anu.edu.au/maps?grid=gh32#search=26c&show=11417[Note : This is 15 mins later than normal due to late availability of the room]

7.30pm        After the talk there will be a dinner at Bistro Nguyen’s, 78-80 Alinga Street, Canberra City (on corner of Northbourne Avenue).

Please RSVP Warren Müller ([email protected] or 0407916868) by 1.30pm Monday 27 February if you would like to attend the dinner. 

 

Speaker: 

Prof. Ross Maller

Australian National University

[Co-authors: Max Maller and Rod Broadhurst, WA Health Department]

 

Topic: A recidivism risk prediction instrument based on survival analysis with long-term survivors

 

Abstract:

The West Australian (Adult) Actuarial Risk Instrument (WAARI) is a comprehensive predictive risk assessment system for predicting individual re-offending behaviour among the West Australian population. I’ll describe the construction of the instrument and give an evaluation of its performance, comparing it with another scheme (OGRS) in common current use for recidivism prediction.

Not all prisoners return to prison after release from it, and estimating the proportion who do (or don’t) is of importance in making policy decisions and evaluating effects of programs. The WAARI is used to predict the probability of re-offending of an individual, based on information on his/her history to date. Its construction is based on experience gained over many years with a long term survivor model for recidivism (return to prison or re-arrest of a prisoner after release).

Longitudinal follow-up data on individuals released from prison will always be censored since not all will have returned by a given cut-off census date.  A version of survival analysis which caters for “long term survivors” or “immune” or “cured” individuals is needed. I’ll outline methodology which can be used for this and illustrate it with examples from the criminology and medical areas.

Warren Müller

Secretary

SSA Canberra Branch

Follow

Get the latest posts delivered to your mailbox:

Show Buttons
Hide Buttons