The Google Subpoenas: In the course of its litigation over the constitutionality of the Child Online Protection Act, aka COPA, the Justice Department has issued a subpoena attempting to compel Google to disclose a) a sample of 1 million Google queries, and b) all of the Google queries that were entered in a one-week period (absent any identifying information). Google has refused to comply, challenging the subpoenas as irrelevant and overbroad, as well as raising other arguments.

  The government needs this information, it alleges, because under Ashcroft v. ACLU, the constitutionality of COPA may hinge on it. To make a long story short, COPA requires commercial providers of adult materials to put pornography behind password gates or other authentication screens to help make sure it not readily available to children. The key issue in the COPA litigation is the effectiveness and restrictiveness of filters versus COPA screens as mechanisms for keeping pornography out of the hands of kids. DOJ's position is that COPA screens are more effective than filters, so the law requiring screens is constitutional under the First Amendment. DOJ's argument is that it needs the information from Google to explain the role of search engines in locating and obtaining pornography, which is then integral to understanding why filters are not an effective alternative to COPA screens.

  I'm hoping VC readers can help me understand and assess these claims a bit better. There are two related issues here, I think. First, how are the results of the Google subpoenas going to help answer whether COPA screens are more effective than filters? In other words, what might the results show that will help DOJ make its case? And second, what other data is out there that might address these questions either as effectively or more effectively than the information requested in the subpoenas being litigated?
Government Expert's Explanation for the Subpoena of Google Records:

The theory strikes me as quite similar to that explained by Mark Seecof in his comments.

There's also some more material pointed to on SearchEngineWatch.