1 / 157

Thursday October 11 th 2007

Welcome to the 2 nd Neighbourhood Statistics Data Supplier Open Day. Thursday October 11 th 2007. Introduction. Andy Teague Head of Admin Sources. Data Supplier Open Day, October 11th 2007. Some basic details you need to know: Fire alarms Toilets Smoking.

ayita
Download Presentation

Thursday October 11 th 2007

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Welcome to the 2nd Neighbourhood Statistics Data Supplier Open Day Thursday October 11th 2007

  2. Introduction Andy Teague Head of Admin Sources

  3. Data Supplier Open Day, October 11th 2007 • Some basic details you need to know: • Fire alarms • Toilets • Smoking

  4. Data Supplier Open Day, October 11th 2007

  5. Data Supplier Open Day, October 11th 2007 A ‘Talking Wall’ for questions….. Jot your question down on a note and stick it on to the flipchart at the back of the room. We’ll do our best to answer it during the final session.

  6. Update on Disclosure Control Issues Philip Lowthian Statistical Disclosure Control Branch Methodology Directorate

  7. Statistical Disclosure Control Branch Aims and objectives of the branch: • Develop disclosure control tools, innovative solutions and methods • Promote, disseminate and provide support for disclosure control tools • Develop, promote and support the implementation of professional standards • Keep up to date with international work on SDC methods • Build up a pool of expertise

  8. Statistical Disclosure Control Branch We work closely with several other parts of ONS • National Statistics and International Division (NSID)   • The Data Stewardship Group (DSG) • IM Strategies • The Microdata Release Unit  • Census • Neighbourhood Statistics

  9. Topics under discussion • Release of Tau-Argus 3.2.0 • GSS Standards for Statistical Disclosure Control • Information Loss • Releasing information for under 18 conceptions

  10. Tau-Argus – A Summary • Can read in both tables and microdata (user creates tables within Tau-Argus) • Can operate in both interactive and batch mode • Can make safe both frequency tables and magnitude tables. • For magnitude tables the user sets safety rules (primary suppressions) such as threshold rule and dominance rule.

  11. Tau-Argus – A Summary • Recoding can be applied to the table • Secondary Suppressions are carried out on magnitude tables to avoid disclosure by differencing. • Controlled rounding is applied to frequency tables so that additivity is preserved.

  12. Tau-Argus 3.2.0 Improvements / Alterations I • The partitioning options for the rounding procedure have been improved. Partitioning is now permitted on hierarchical variables. • For tables with non hierarchical variables, the user will have the option of choosing the number of blocks to break the table into prior to rounding. • The batch mode will offer all the features available in the interactive mode, including the partitioning options described above.

  13. Tau-Argus 3.2.0 Improvements / Alterations II • More information will be given to the user on screen after rounding (both with and without partitions) such as No. of jumps Max. jump Number of rapid subtables Number of feasible subtables Number of optimal subtables • The hypercube method of secondary suppression will offer the option of 'small' or 'large' tables. By choosing the relevant option these should improve processing time.

  14. Tau-Argus 3.2.0 Bug Fixes I • An earlier problem with rounding to base 3 and setting the threshold value to 3 sometimes causing the rapid solution to be automatically chosen now been fixed. • Safety rules applied after reading in tables will always be correct. Currently p% and dominance rules are not disabled when there is not enough information in the table about the number of contributors and their values. • The 'timecheck' window is now fully readable.

  15. Tau-Argus 3.2.0 Bug Fixes II • On the rare occasion when the optimal method of suppression chooses all cells for secondary suppression, a message will state: 'All cells have been suppressed, the problem might have become infeasible'. • The program will be more stable and less likely to crash when suppressing or rounding.

  16. GSS Standards for SDC • Developed to ensure consistent approach to SDC for different outputs • GSS Standard for tables produced from administrative data sources • GSS Standard for tables produced from surveys (not covered here) • Social surveys • Subsamples • Business surveys

  17. GSS Standards for SDC • In development: • Standard for microdata • Standard for analytical outputs • Standard for outputs from Census 2011 • These standards exist alongside specific NeSS standards • Concentrate on the standards for tables produced from administrative data sources

  18. GSS Standards for Frequency tables I Administrative data: disclosure risk • Data provider should consider the risk of different types of disclosure • Risks are related to disclosure likelihood and impact • Identify ‘unsafe’ cells • 3 broad risk categories

  19. GSS Standards for Frequency tables II • Low risk • Likelihood and impact of an identification low • High level of aggregation and limited tables produced from source • Care taken with zeros and marginal totals of size 1 or 2 • Medium risk • Majority of health statistics • All cells of size 1 or 2 unsafe, care taken with zeros • High risk • Likelihood and impact of an identification higher, eg statistics on abortions • All cells of size 1 to 4 unsafe, care taken with zeros

  20. GSS Standards for Frequency tables III Administrative data – Disclosure methods • Table redesign is recommended as a simple method • Controlled rounding recommended for most tables • In some cases, if the number of unsafe cells is low, suppression can be implemented

  21. Information loss software I • Measures ‘information lost’ due to the application of an SDC method for tables • Measures the difference between the original and protected tables • Evaluates different measures, e.g. bias, variance, measures of association • Allows comparison of SDC methods • Developed by the ONS • Programmed in SAS

  22. Information loss software II Details on information loss measures • Basic Statistics: Number of cells and the total information in the table; number of zeros, ones, and twos etc. • For suppressed tables: number and percent suppressed cells and total information lost; choice of imputation method • For random rounded tables: Binomial hypothesis test to check for bias in the rounding scheme • For all other SDC methods: paired sign rank test to check for no change in the location

  23. Information loss software III • Distance metrics - distortions to distributions on internal cells - distortions to distributions on marginal sub-totals / totals • Impact on Tests for Independence • Impact on Rank Correlations

  24. Conceptions Data • PSA target to reduce under-18 conception rate by 50% by 2010 • Requirement for information by small area • Three SDC methods used: • Data for 3 years combined • E&W data divided into quintiles • Suppression for small populations • Maps produced on NeSS at the ward level

  25. Conceptions Data

  26. Summary • SDC is involved in a number of projects • Here I have only given a brief update • We also provide training courses especially an awareness raising course • Please contact the branch for any advice / further information

  27. Neighbourhood Statistics Data Supplier Open Day Thursday October 11th 2007

  28. The Statistics and Registration Service Act and Opportunities for Sharing Administrative Data Kieron Mahony & Minda Phillips

  29. Outline of Presentation This presentation will focus on:- • outlining the scope of the Act; • considering key implications; • data sharing issues; and • use/development of administrative data.

  30. Statistics and Registration Service Act 2007 – Key Provisions • Creation of a new body – the Statistics Board, a non-Ministerial Department reporting directly to Parliament • Remit covers the whole UK statistical system • Overall objective: “To promote and safeguard the quality of official statistics that serve the public good” • ONS to become the ‘executive office’ of the Board

  31. Statistics and Registration Service Act 2007 – Key Provisions • National Statistician the Chief Executive of the ‘executive office’ • Robust confidentiality regime • GRO and NHSCR to move elsewhere within Government ; Statistics Commission abolished 8. Cabinet Office takes over residual Ministerial responsibilities

  32. Main responsibilities of the Statistics Board • Monitoring and reporting on all official statistics • Development of Code of Practice • Assessment of National Statistics • Oversight of ONS (the ‘executive office’) responsible for statistical production • Reporting to Parliament

  33. Code of Practice & Assessment • Act requires the Statistics Board to prepare, adopt and publish a Code of Practice • Existing National Statistics will remain as National Statistics until re-assessed for compliance against this new Code • Candidate National Statistics will be assessed against the Code at the request of Ministers • Act allows for assessment to take place using the existing Code • Statistics Board must prepare and publish a programme for assessment and re-assessment

  34. Data Sharing Act will allow two-way sharing of data between Statistics Board and public authorities/Devolved Administrations for statistical purposes, subject to the following safeguards:- (i) only possible where otherwise prohibited by law or public authority/DA has no power; (ii) subject to bilateral Ministerial agreement; (iii) sharing only for specific purposes and subject to public interest test; and, (iv) Parliamentary approval

  35. Data Sharing - Key Issues • Legal basis; • Access and legal rights; • Confidentiality; • Importance of public approval; • Unified identification code systems; and • Co-operation among administrative authorities.

  36. Preparing a Draft Regulation • Special attention will need to be given to:- • understanding the present legal barriers to sharing; • defining each data item to be shared; • preparing a business case to show why each data item is needed; • explaining the benefits associated with access to each data item; and • demonstrating the public good.

  37. Requirements for Administrative Data - General • To inform policy formulation and review; • To support resource allocation; • To improve service planning and delivery; • To enhance research and analysis.

  38. Requirements for Administrative Data - ONS • To improve population and migration statistics; • To facilitate work on 2011 Census; • To support validation and quality assurance; • To respond to increasing demands for more sophisticated statistics/analysis; • To reduce respondent burden; and • To meet efficiency targets.

  39. Administrative Data Development and Data Sharing • Working to develop a strategic plan or roadmap by:- • understanding needs and benefits; • developing criteria for determining priorities; • setting up procedures to deal with ethical issues; • establishing appropriate technical environment(s); and • researching the feasibility of linking/matching.

  40. Next Steps • Developing coherent work programme; • Undertaking pilot/feasibility work; • Evaluating alternative methods of data linkage and matching; • Considering data processing issues; • Understanding technical/infrastructure requirements; and • Obtaining resources.

  41. THANK YOU

  42. Neighbourhood Statistics Data Supplier Open Day Thursday October 11th 2007 Tea and Coffee

  43. Oct 2007 • Small Area Geography • (OAs & SOAs) • Consultation • Findings • & • approach • Alistair Calder

  44. Oct 2007 background results proposed policy

  45. background Output Areas created in 2002 for Census Adopted as the base of NeSS Geography – Super Output Areas built 2004 ‘Review’ promised Consultation Nov ’06 – Feb ’07 Widely publicised – online consultation & ‘blog’ – great response

  46. results and outcome now published graphs necessarily over-simplified

  47. Strongly supportive of stability (but what kind of stability – at what level?)

More Related