Official Statistics – an introduction and overview Paul Allin, CStat, FRSA Visiting Professor and former director of Measuring National Wellbeing Programme, UK Office for National Statistics We will cover • • • • • What are official statistics and why have them? Key features of official statistics (Very) brief history of official statistics Main areas of official statistics Producing official statistics – statistical value chain • Main sources of data, especially censuses and surveys • Interacting with users – policy makers and public What are official statistics? • For this purpose, all statistics produced by government departments, devolved governments, national statistics office (ONS in UK), but not local authorities, public utilities etc • Constitutional position varies from country to country. UK got statistical legislation relatively late (2007 Statistics and Registration Services Act http://www.legislation.gov.uk/ukpga/2007/18/contents , short guide here http://www.ons.gov.uk/ons/guidemethod/ons-independence/the-statistics-act/index.html) • Examples : population counts, national economic accounts, consumer prices index “Statistics are the mirror through which we view society” David Hand: President of the Royal Statistical Society, 22.02.10 Sometimes contested! • Never mind the figures, the economy is ‘healing’: minister (Observer, 7 Oct 2012) • “I don’t think anybody in America looks at the [unemployment] number ... and thinks that’s a real reflection of reality” (Peggy Noonan, Wall Street Journal, quoted in New Statesman, 12-18 Oct 2012) UK Office for National Statistics ONS: other website features Let’s visit the Afghan Central Statistics Office • “Statistics is a new thing for most people in Afghanistan. They don’t feel it’s a need, a necessity” – head of CSO • “The first step is to admit just how bad and conflicting many of the data now being used really are” – Centre for Strategic and International Studies paper • “We don’t ask for ethnicity or language spoken, this is on purpose” – head of UN Population Fund in Afghanistan, supporting first census since 1979 Afghan CSO – the issue • It is hard to overstate how few reliable numbers there are about population or anything else in Afghanistan, or what a problem this is for those trying to bolster the economy, distribute aid, decide where clinics should be built or how many teachers recruited, or do any other kind of long-term planning. (The Guardian, 04.01.13) Why official statistics? • Statistics for government and for the public good (including to assess performance of government) – Independent and seen to be independent – Observed tendency to focus on government statistical service – Differences between and commonality across the two roles – How to deliver both roles efficiently and effectively? Eg New Public Management and Public Value Theory Towards some fundamental principles • Needs for official statistical information • Need for trust in those statistics • Quality depends on cooperation of data provides and survey respondents • Need for international comparability • Professional ethics UN Fundamental Principles 1. Relevance, impartiality and equal access 2. Professionalism 3. Accountability on sources, methods, procedures 4. Prevention of misuse 5. Cost-effectiveness 6. Confidentiality 7. Legislation 8. National co-ordination 9. International co-ordination 10. International statistical co-operation http://unstats.un.org/unsd/methods/statorg/FP-English.htm UN Fundamental Principle 1 • Relevance, impartiality and equal access Official statistics provide an indispensable element in the information system of a democratic society, serving the government, the economy and the public with data about the economic, demographic, social and environmental situation. To this end, official statistics that meet the test of practical utility are to be compiled and made available on an impartial basis by official statistical agencies to honour citizens' entitlement to public information. http://unstats.un.org/unsd/methods/statorg/FP-English.htm Issues • One of the cornerstones of good government and public confidence in good government • Practical utility – relevant, fit for purpose, accessible and usable • Impartiality in compilation and release (legislative base, open methods) • Make information available widely and impartially Ways of understanding user needs Key to achieving compliance with fundamental principles is “maintaining an understanding of what statistical information users want and how they want it” • • • • Advisory bodies User consultation Responsive organisational planning and operation Dissemination and marketing Key features of official statistics • Quality and standards (definitions, methodologies etc) • UK subset of all official statistics called ‘National Statistics’ with Code of Practice* compliance – Timeliness, accuracy, etc – Accessibility (including pre-release access) – Relevance and comparability • International dimension strong - international standards, ‘regional’ statistical offices (eg Eurostat, OECD) • Organisation, oversight and accountability to Parliament *Code published here: http://www.statisticsauthority.gov.uk/assessment/code -of-practice/index.html (Very) brief history of official statistics • Official statistics before statistics invented (eg linear B tablets, Augustus Caesar’s census) • Age of Enlightenment – political arithmetic, observations on bills of mortality • Early 19th century golden era of official statistics in Europe (and North America) • Board of Trade and Office of the Registrar General foreshadow business statistics and population statistics developments From first edition of proceedings of Statistical Society of London Mr. Benjamin Heywood, the President cf the Statistical Society of Manchester, communicated the results of an inquiry instituted by that Society in 1834, into the condition of 4,102 families of working men in certain districts of the town of Manchester, This information had been procured by personal inquiries at each house, made by an agent of the Society. The following is an abstract of it: One of the first ‘official’ statistics? Some key points from history • Root of the word ‘statistics’ is the Latin word for ‘the state’ • Official statisticians mainly producers and disseminators of (tables of) statistics • The analysis of data about the state – inferences to be drawn from rigorously collected data • Economic ‘vs’ social statistics • National wellbeing as return to the “quantum of happiness” enjoyed by inhabitants of a country Worth following up • Kotz (2005) gives a good overview http://www.jos.nu/Articles/abstract.asp?articl e=212139 • 1874 edition (21st) of Board of Trade Statistical Abstract for the United Kingdom http://archive.org/details/annualabstracts01tr adgoog]. Main areas of official statistics • Population and ‘vital’ statistics, demography, migration, households • Economy, trade, finance, prices, national accounts and balance of payments • Business demography • Labour force, human capital • Health, life expectation • Social statistics – neighbourhoods, crime, education, leisure, media • Measuring national wellbeing and progress of societies Main types of statistical work: the statistical value chain • • • • Identify user requirements Design of data collection, methodology Data collection operations Data processing (often on a large scale), checking, validating • Dissemination • Analysis UK official statistics released on 15 January 2013 at 9:30am Main sources of data for official statistics • Administrative systems (government, banking etc) • Censuses (businesses or people/households) • Sample surveys (ditto) • Specialist data collection and surveys eg time use • New directions eg Beyond 2011 Census; Measuring National Well-being Introduction to censuses and surveys • For systematic collection of data from people, households, businesses • Census = “official enumeration of inhabitants with statistics relating to them” (from Latin for ‘a register’, Chambers Dictionary) • (Sample) survey: “Sample survey is the technique used to study about a population with the help of a sample. Population is the totality all objects about which the study is proposed. Sample is only a portion of this population, which is selected using certain statistical principles called sampling designs (this is for guaranteeing that a representative sample is obtained for the study). Once the sample decided information will be collected from this sample, which process is called sample survey” (Source: Yahoo) Which technique? Depends on which two are most important of ... Cheap Quick Quality Dealing with uncertainty • Census – how do we know we have counted everyone? Cross checks, validation, registers, post-enumeration survey • Surveys: – Aims to quantify the uncertainty of estimates – Balance cost and precision through sample design Survey process • • • • Questionnaire design and cognitive testing Piloting on small sample Drawing main sample Organising fieldwork including strategies to minimise non-response • Data collection now mainly CAPI or CATI • Data checking • (Processing, validation, analysis, dissemination) Disseminating official statistics • Two audiences – policy makers and public • From paper to website • Role of the media • Partnerships and collaboration • Worth listening to – Andrew Dilnot on Numbers and public policy: the power of official statistics and statistical communication in public policy-making http://www.statisticsauthority.gov.uk/news/nu mbers-and-public-policy.html Some concluding comments • How much independence (eg funded by government)? • How radical? • How is public good delivered? – role of media • How accessible? • In short, are official statistics fit for purpose? and how do we know?