project? Microdata? Say what? TRY Conference May 5, 2008 Suzette Giles, Ryerson University Laine Ruus, University of Toronto

Acronym of “Ontario Data Documentation, Extraction Service and Infrastructure Initiative” A new product delivered through Scholars Portal Collaboration between OCUL and Ontario Buys Web-based resource discovery to a growing collection of Canadian data 3 rd generation statistics and data extraction system Equalizes access to selected statistics and data for all Ontario universities

A web-based extraction system Provides equal access to these resources for all Ontario universities Contains diverse, quality, numeric (microdata) data sets Allows data resource discovery, extraction and analysis

First some background about Why is this resource so important that Ontario Buys and OCUL (Ontario Council of Ontario Libraries) are investing over $1.4 million? How will it support teaching and research in quantitative methods and contribute to statistical literacy? How will it help me at the Reference Desk?

Why is this resource so important that Ontario Buys and OCUL (Ontario Council of Ontario Libraries) are investing over $1.4 million? 23 universities & colleges in Ontario belong to DLI, only 7 have full-time staff ‘doing’ data therefore, large differences in data access by faculty and students Goes some way to making access to data and statistics more easily findable.

How will it support teaching and research in quantitative methods and contribute to statistical literacy? As a resource discovery tool, ability to search across collections at a more finely-grained level than Stats Can provides Access to resources that have not previously been readily accessible, eg Canadian Gallup polls Ability to quickly and easily display descriptive statistics, regardless of whether the resource is aggregate statistics or microdata. Web-based access supports use in classroom as well as 24x7 access for research.

How will it help me at the Reference Desk? Easy access to a collection of statistics and data in a uniform interface Blurs the distinctions between aggregate statistics and microdata Enable the creation, on the fly, of aggregate statistics that have not been published elsewhere

What are microdata and why do I need to know about them? Microdata are the actual responses that survey or census respondents give to a questionnaire, Usually translated into a numeric format so that one can do arithmetic with them For example: Income: (1) high, (2) medium, (3) low –Average = ???? Income: (1) $13,725 (2) $118,297 (3) $63,958 –Average = $65,327

What are microdata? A person’s sex or gender could be recorded as Male or Female, but instead it will be coded 1 or 2. 1 = male 2 = female or the other way 1 and 2 are the VALUE given to the VARIABLE in this case sex (or gender) A microdata file consists of numbers

From microdata are generated descriptive statistics Microdata –A person is working or not working Aggregate statistics –A count of the number of persons not working (in a geographic area) –A count of the number of persons not working divided by the number of persons in the labour force = the unemployment rate

…and more descriptive statistics Microdata –A family has a gross annual income in year 2005 Aggegregate (descriptive) statistics –Families in a geographic area have an average income –50% of families in a geographic area have an income above (or below) the median income –LICO is the % of families in a geographic area that have an income below the low income cut-off for that geographic area and family size

Two systems Available to all Ontario universities Searching metadata across files Descriptive statistics only Download system files (users needs to have software) SDA 10 universities subscribe, incl Ryerson & York More advanced statistical analysis functions Download raw data & syntax files (user needs to create system files)

What’s changed? In the ‘bad old days’ –Statistics were published in books/periodicals, data were ‘published’ mainly as files of microdata or very extensive aggregate statistics –You needed access to a mainframe or PC.Mac –You needed special software (SAS, SPSS, Stata) –You needed training in the production of descriptive statistics (weighting, types of data and appropriate types of descriptive statistics)

But nowadays! Statistics are published as Excel files, Beyond 20/20 files, and microdata are available in 3 rd generation interfaces All you need is a computer and a web browser Generate descriptive statistics with a few mouse clicks – tho’ you still need to know how to interpret them (and knowing about weighting is a good idea too!) Users can download data 24x7 for further processing on their own workstations with appropriate software