Azzah Al-Maskari Mark Sanderson Paul Clough Eija Airio The good and the bad system: does the test collection predict users' effectiveness? 59-66 2008 SIGIR http://doi.acm.org/10.1145/1390334.1390347 conf/sigir/2008 db/conf/sigir/sigir2008.html#Al-MaskariSCA08