See No Evil

Please see the Kaiser Family Foundation site for general information on our study, See No Evil: How Internet Filters Affect the Search for Online Health Information.

Publications and other documents

Richardson, C., P. Resnick, D. Hansen, and V. Rideout, Does Pornography-Blocking Software Block Access to Health Information on the Internet? Journal of the American Medical Association, Dec. 11, 2002. 288(22) 2887-2894.
Resnick, P., D. Hansen, and C. Richardson, Calculating Error Rates for Filtering Software. Forthcoming in Communications of the ACM (preprint version in PDF format)
Richardson, C., H. Derry, D. Hansen, and P. Resnick, Adolescents Searching for Health Information on the Internet: An observational study. J Med Internet Res 2003, October, 17;5(4):e25
Executive Summary for the project available from the Kaiser Family Foundation site
Appendices available from the Kaiser Family Foundation site. Provides all the details about testing dates, configurations, etc., as well as some additional results about Google's filtered search engine.

Selected Media Coverage:

ABC News - Filtered Finds: New Study Shows How Net Porn Filters Block Online Health Info
CBS News - Internet Filters Can Work
Health on the Net Foundation News - Net Filters Don't Necessarily Block Health Sites
Internet News - Study: Net Filtering Blocks Useful Content
New York Times (via Common Dreams News Center) - Internet Filters Block Many Useful Sites, Study Finds
Reuters (via AZ Central) - Web filters block some health sites, study says
San Jose Mercury News - Filters can block health Web sites
Washington Post - Study: Web Filters Block Health Information

Data and Analysis

The main data set

The main data set includes one row for each site that was tested. Some URLs appear multiple times-- duplicates should be removed for certain kinds of analyses, as was done in the analysis files below. The data dictionary file provides some useful information about how to interpret the columns.

Analysis: stata log files

Analysis of search results, filters.log (eliminates duplicate URLs returned from more than one search)
Analysis of search results by search string, filterss.log (the numbers from the JAMA article and the executive summary that are broken down by search string come from here; eliminates duplicate URLs only if they were returned by more than search engine for a single search string, allowing analysis of results by search string)
Analysis of top health sites, filterb.log (eliminates duplicate URLs found in more than one category at Yahoo or Google)
Analysis of search engine performance, filterse.log (not reported as part of the main study; analyses amount of health, porn, and other sites returned by different search engines when given health search terms)
Analysis of Google filtered search, as compared to Google regular search, filtersearchprep.log and filtersearch.log (not reported as part of the main study; reported briefly in an on-line appendix available from the Kaiser site.)

Cached web sites

Cached copies of the first two levels of web sites, as of the time that searches and filtering were done, are available to researchers who wish to do follow-up analysis. The cache is very large (more than 1GB zipped) and contains both copyrighted and pornographic images, so we are not making it available for general download. Please send email if you have a legitimate research use for this data set, and we will arrange to send it to you.

The Study Team

University of Michigan School of Information

Paul Resnick, project director
Derek Hansen

Department of Family Medicine, University of Michigan Medical School and VA Health Services Research and Development Service

Caroline Richardson

University of Michigan Health Media Research Laboratory

Holly Derry
Ian Jones
Mike Nowak
Ed Saunders
Vic Strecher

Department of Biostatistics, University of Michigan School of Public Health

Rod Little
Ying Yuan

Kaiser Family Foundation

Vicky Rideout