Update: Rudd Labor government censors results from ISP filtering trial

In an earlier post I briefly explained how key results were missing in the Enex TestLab report on the government's live pilot of its proposed ISP-level filtering scheme.

The results of the accuracy testing can be combined into a single table.

No results were published on the percentage of URLs that were blocked on the innocuous list of URLs, as was done for the participants blocking additional content. The results cannot be considered complete without figures providing an indication of the rate of over-blocking that occurred for filters blocking URLs on the ACMA blacklist only.

A filter blocking only those URLs listed on ACMA blacklist should in theory block fewer innocuous sites than a filter blocking additional categories of content as well as the ACMA blacklist. However, the results for filtering the ACMA blacklist only are crucial as ISPs will be required to block the content specified on this list for all Internet users under legislation the federal government intends to introduce into Parliament this year.

While the rate of over-blocking for filters blocking the ACMA blacklist alone is likely to be better, publishing just the results for trials of filters that block both the ACMA blacklist and additional categories of content puts the government in a convenient position when challenged by criticism of the filters' performance. Enex stated in the report that the percentages for the blocking of URLs on the innocuous list of URLs were high:

In terms of over-blocking the results of this trial show that, while an improvement on previous testing levels, this is still considered high.

The government might respond to criticism by stating that less over-blocking will occur with the proposed mandatory filter which will block URLs on the ACMA blacklist without providing any actual figures obtained from tests.

The case is similar with the feedback survey that was carried out for customers of the ISPs involved in the pilot. The customer feedback survey was not issued to customers of ISPs that filtered the ACMA blacklist only, as stated in the report.