Focusing your search
What are you looking for?
What is your objective: to include a few numbers in a report (borrow), or to analyze data (repurpose)?
What unit of observation do you want to describe?
Do you need aggregated (summarized) data or raw microdata or spatial data or ...?
What software do you intend (or are you willing) to use?
In what format would you like the data delivered? What format(s) will you settle for?
modified from Jake Carlson at ICPSR, 2012
Strategies for finding datasets
Plan A: Article-based
- Find articles that match your topic.
- Look at the full text of relevant articles. Data or statistics may be present in the article.
- Look for a link or URL to "supplementary" items. These may contain more detailed or additional data than you'll find the article text itself. See examples of links to supplementary info.
Plan B: Search article indexes that have thumbnails of images/charts/graphs, such as:
- Academic Search Premier -- shows thumbnails of graphics when full text is available in the service.
- ProQuest -- choose a subject area, then limit your keyword search to "figures and tables" or "data & reports" where available.
Plan C: Data Repositories
- Use the subject tabs above to select multi-disciplinary or subject-based data repositories.
- Browse the datasets available and open any that seem relevant, or keyword-search.
Plan D: Highly specialized search engines may have a filter for datasets, such as:
- ADS (Astrophysics Data System)
Plan E: Grants databases
Tips & Tricks
* Data may be stored in images: photos, charts and tables saved as .gif or .jpeg images, etc.
* Article search engines don't (yet) indicate whether data are present in an article or in supplementary form. You need to look at the article full text to see for yourself if data are present and relevant.
* Relevant data may reside in datasets or articles written in a language that you do not read. For best results, find someone with both language and subject knowledge to assist you.
* Make sure you can track and cite the original source of the data. If you are looking at a databank or handbook or other secondary source, look for evidence of the original data gatherer(s).