Public Policy Analytics: Cases and Issues 6 Unit Course May 2018

Transcription

1 Course Information Public Policy Analytics: Cases and Issues 6 Unit Course May 2018 Instructor: Tim O Loughlin Office: First level Phone: Emil: toloughlin@australia.cmu.edu Office hours: TBA Teaching Assistants: Office Hours: TBA Meeting Times: TBA Course Web Site: CMU Canvas Prerequisites Description None How can an algorithm used to identify children at risk of mistreatment be hailed as a success in Pennsylvania, useless in New Zealand and the subject of legal action by citizens in Scotland? Are data techniques applied to the criminal justice systems prone to racial bias? What rights should citizens have over the dissemination and use of data collected on them by governments? What redress should citizens have when data anonymized by governments is re-identified? We all know that the same big data techniques applied successfully in the commercial world are being turned to use by governments. However, the relationship between governments and their citizens is vastly more complex than the relationship between companies and their existing and prospective customers. This course focusses on the management issues that arise for governments looking to big data applications for better performance. In taking this perspective it traverses the various levers available to governments such as legislation, administrative reform and participative democracy as well as the pitfalls. The ultimate aim is to develop sound understandings of the principles which public policy professionals using big data techniques can apply to achieve successful outcomes for both governments and their citizens.

2 The course concludes with a critical examination of some contemporary issues emerging from the use of dig data in the public policy arena such as perceptions of disempowerments and assertions of accountability deficits. Course Materials For each lecture, the instructor will provide lectures notes and any class material relevant for the lecture. Occasionally additional readings will be provided. The course will include presentations from several practitioners with students being encouraged to assess their practices critically in open discussion. Course Objectives Analyze the policy questions that rise from the use of AI for prediction Research best international practice in both the use of AI for policy purposes and the creation of legal and administrative frameworks that achieve high levels of public support for AI applications in various policy areas Evaluation Method Student s performance in the class will be evaluated based on two 3,000 word essays dealing with each objectives as applied to a specific area of public policy to be selected by individual students. Grading Scale A+ 97% 100 % B+ 85% % C+ 73% % A 93% % B 81% % C 69% % A 89% % B 77% % C 65% % Scores below 65% equate to a failing grade (R) 2

3 Course Outline / Readings Theoretical foundations I: policy analytics, legitimacy and democratic participation Bryson J,. Crosby B and Bloomberg L Public Value Governance: Moving Beyond Traditional Public Administration and the New Public Management Public Administration Review, Vol. 74, Iss. 4, pp Daniell, K, Morton A and Insua D 2016 Policy analysis and policy analytics Annals of Operations Research 236:57-73 Scharaschkin, A and McBride, T 2016 Policy analytics and accountability mechanisms: judging the value for money of policy implementation Annals of Operations Research 236:36-56 De Marchi, G, Lucertini, G and Tsoukias 2016 From evidence-based policy making to policy analytics Annals of Operations Research 236:15-38 Theoretical foundations II; policy analytics, public sector management and governance Australian Government Productivity Commission Inquiry: Data Availability and Use Report Overview No. 82, 31 March Australian Government Commonwealth Ombudsman 2017 Centrelink s automated debt raising and recovery system. April Cleary, D Predictive Analytics in the Public Sector: Using Data Mining to Assist Better Target Selection for Audit Revenue Irish Tax and Customs, Ireland. Self published. dcleary@revenue.ie Donaldson, D 2016 Magnificent Seven: government data uses that actually work The Mandarin 9/12/ combined+list&utm_campaign=b1691aa3b8-&utm_medium= &utm_term=0_d98f7edac0- b1691aa3b Angwin J 2016 Make Algorithms Accountable The New York Times. 1 August Heller, N Letter from Tallinn: The Digital Republic The New Yorker. December 18 & 25 Special topics for policy analytics I: What works best for child protection? Three approaches: predictive risk modelling, actuarial risk assessment and the consensus method Vaithianathan, R 2012 Can administrative data be used to identify children at risk of adverse outcomes? Centre for Applied Research in Economics, University of Auckland. September Katz I, Cortis N, Shlonsky, A and Mildon R 2016 Modernising Child Protection in New Zealand: Learning for system reforms in other jurisdictions Social Policy Evaluation and Research Unit. University of New South Wales. May Scherz, C Protecting Children, Preserving Families: Moral Conflict and Actuarial Science in a Problem of Contemporary Governance Political and Legal Anthropology Review. May pp Special topics for policy analytics II: Predictive modelling s uses for optimising public and private transport efficiency, access and environmental impacts 3

4 Kumar, A, Nguyen, V and Teo K Commuter cycling policy in Singapore: a farecard data analytics based approach. Annals of Operations Research 236:57-73 Special topics for policy analytics III: public health care Dare T, Deceiving Third Parties 2010 Journal of Primary Health Care Vol.2 No.1 March pp Brennan A, Meier P, Purshouse R Rafia R, Meng Y and Hill-Macmanus D 2016 Developing policy analytics for public health strategy and decisions the Sheffield alcohol policy model framework Annals of Operations Research 236: Ola O and Sedig K The Challenge of Big Data in Public Health: An Opportunity for Visual Analytics Online Journal of Public Health Informatics * ISSN * * 5(3):e223, 2014 Australian Government Productivity Commission Inquiry: Data Availability and Use Report No. 82, 31 March Appendix E Case Study: Heath Data Closing topic: Does bias exist in data analytics and what can be done about it? Angwin J, Larson J, Mattu, S and Kirchner L, 2016 Machine Bias ProPublica May Angwin, J and Castaneda, L.1998 The Digital Divide US Black Engineer and Information Technology, 08/1998, Volume 22, Issue 2 Angwin J 2014 Dragnet Nation: A Quest for Privacy, Security and Freedom in a World of Relentless Surveillance Chapter 1: Hacked, 12 Colo. Tech. L.J. 291 Flores A, Bechtel Kristin and Lowenkamp, C False positives, false negatives, and false analyses: a rejoinder to "Machine bias: there's software Federal Probation, 09/2016, Volume 80, Issue 2 4

5 Course Policies & Expectations Lectures: Class participation is encouraged and expected. As research on learning shows, unexpected noises and movement automatically divert and capture people's attention. Arriving in the classroom late (after a lecture begins), using cell phones, pagers, laptops, etc. makes noise, are distracting and affect everyone s learning experience. For this reason, I ask you to show up for each class a few minutes early and refrain from using your mobile devices during class. No student may record or tape any classroom activity without the express written consent of the instructor. If a student believes that he/she has a learning disability and needs to record or tape classroom lectures/activities, he/she should contact the Office of Equal Opportunity Services, Disability Resources to request an appropriate accommodation. Missed Classes: Students are responsible for obtaining class material, which may have been distributed on days when they are absent. This can be done through the class blackboard, contacting a classmate who was present, or by contacting the instructor during his office hours. Academic Honesty and Integrity All CMU students are expected to follow the ethical guidelines and adhere to the policies as defined in your Program s Student Handbook or in any other source describing such policies as they apply to students at Carnegie Mellon University. These policies and guidelines are available on the CMU web site. Please read them carefully! You will be held accountable for any violations of these policies and guidelines. Individual assignments must reflect individual effort. Although I expect you to attempt solving each problem on your own, I encourage you to seek help from the class TAs if you struggle with any assignment. Sharing your assignments with any other student in any form (whether it is a paper document, an electronic document such like a MS Word document, or a document in any other format) is not permitted and will be considered cheating. Any discussion between students that results in a similar HW submission is also not allowed. If you are in possession of any other person's document or file from this or any other semester, you are in jeopardy. Any violations of academic integrity in this class will have the following consequences: (a) at the minimum, no credit for assignment in question and lowering final grade by one letter (e.g., from B to C); (b) in more serious offences, failing the class; (c) cases will be reported to the Dean s office 5