Statistics and the peer review process in construction research

This paper examines, in some detail, two papers published in 2003 and 2004 in respected journals. The papers had been through a typical peer review process, with one of the journals in question using four referees rather than the usual two. The papers under review have fundamental flaws. A number of criticisms of the papers are presented and some unfortunate misuse of statistical methods examined. The use of inappropriate statistical methods suggests that a simple analysis was perceived to be insufficiently "academic" and an attempt was made to give the appearance of more academic rigour than the topic or the data demanded. In conclusion it is argued that journals publishing papers with significant statistical content should ensure that at least one of the referees who examines submitted papers should be a mathematician or statistician, and all referees should be encouraged to decline to comment if they do not possess the statistical expertise to properly review the methods used.