Why Batch and User Evaluations Do Not Give the Same Results.

Andrew Turpin, William R. Hersh

Browse the full SIGIR paper archive.