He further notes several concerns about TREC, and more generally, the emphasis on the laboratory experiment and the use of test collections, which can create bias in what is studied, how it is studied and what is considered as a valid contribution to IR research.