Peer review not the problem

Although I have not delved into the subtleties of the applied statistics vs. geostatistics debate, I wish to offer an emphatic response to the final paragraphs of Mr. J. Merks’s letter, which appeared in the Dec. 2-8, 2002 issue of The Northern Miner. In his somewhat confusing final sentences, he decries the scientific peer review process as an obstructive mechanism that disregards scientific integrity. I object to this view in the strongest possible terms.

The peer review process is a time-honoured mechanism to have the integrity (logic, data, methods and interpretations) of scientific work evaluated by experts. Although not foolproof (witness the recent scandal about fraudulent published research on organic microchips), it nevertheless goes a long way to ensuring that nonsense is not foisted upon the unsuspecting public.

Problems arise when non-scientific (personal, emotional) issues become involved and begin to overshadow the scientific issues. In these situations, it is the responsibility of the editor of the publication to mediate. For example, on occasion when submitting a manuscript, authors will request that certain individuals not be reviewers if there are grounds for a non-impartial review. The editor can either send it to other experts for review or be prepared to filter the comments if he/she sends it to the named individuals anyway.

In any case, a paper usually receives 2-3 reviews in order to ensure balance. If an author feels that rejection by a journal is unfounded, he/she may choose to submit to a different journal. Repeated rejection would indicate that the problem lies with the submitted material (for reasons of scientific content and/or style of presentation).

The peer review process is the bedrock on which dissemination of responsible scientific results rests. With regard to the dispute over methods of applied statistics vs. geostatistics, an obvious solution would be to find a suitable unclassified data set, analyze it using the two competing methods of analysis, and publish a discussion of the results that include the raw data and a clear exposition of the different methodologies and their assumptions.

David R.M. Pattison

Dept. of Geology and Geophysics

University of Calgary