9

enter image description here

This article is from 2002, so it may be outdated, but it claims that when web filters are set to the more restrictive settings, such as in libraries or schools, they are more likely to block sites with health/medical information than pornography. Is this a common occurrence?

4
  • The answers would be specific to brands of web filters? Also, some web filters are marketed to end-users (with the blacklist chosen by vendor staff), while enterprise web filters might use an entirely different blocking mechanism (such as consensus of site administrators and submitters).
    – user2547
    Commented May 15, 2011 at 21:24
  • 3
    Important question about the metric: do they mean "block more requests actually made by users" or "block more requests from a theoretical space of valid requests" or "block a larger fraction of requests in each category" or... or... or... Commented May 15, 2011 at 21:46
  • 1
    @rwong The thing is, I know virtually next to nothing about the technology involved. I probably don't even have the knowledge to ask this question correctly. I know there are many programmers and developers here, so anyone can feel free to edit the question as needed, or vote to delete if it's unanswerable or too broad. Commented May 15, 2011 at 21:51
  • @dmckee: One possible metric that is easier to evaluate (though not necessary more meaningful) is to count the number of domains blocked/not-blocked according to whether the domain is porn/non-porn/mixed/questionable, and weighting the score by the domain's global traffic (outside the web filter).
    – user2547
    Commented May 15, 2011 at 21:56

1 Answer 1

10

What the article actually says is that increasing the severity of the filter has more effect on the number of health sites blocked than on the number of porn sites.

That's not the same thing as saying that they're more likely to block health sites.

Here's a completely hypothetical example:

Total number of health sites: 100
Total number of porn sites: 100

Number of health sites blocked with less restrictive filtering: 1
Number of porn sites blocked with less restrictive filtering: 90

Number of health sites blocked with more restrictive filtering: 50
Number of porn sites blocked with more restrictive filtering: 92

In this example, increasing the severity has little effect on the number of porn sites blocked (increasing from 90% to 92%, which is not much change) but has a big effect on the number of health sites blocked (increasing from 1% to 50%, which is a big change).

3
  • you could see it anotherway though, the chance of being on a porn site halves from 2 to 1.
    – Andy
    Commented May 16, 2011 at 11:27
  • @Andy You're right; I'll correct the numbers in the example to get rid of that.
    – ChrisW
    Commented May 16, 2011 at 13:16
  • Since most filters today use some form of heuristics, and not just a blacklist only, you'd do better to calculate "requests for X sites", rather than "Total number of existing sites", which is really nonsensical wrt porn sites (which is, for these purposes, unlimited, as it is perpetually changing).
    – AviD
    Commented Jun 19, 2011 at 9:12

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .