Braun Tibor, Schubert András (szerk.): Szakértői bírálat (peer review) a tudományos kutatásban : Válogatott tanulmányok a téma szakirodalmából (A MTAK Informatikai És Tudományelemzési Sorozata 7., 1993)

MARTIN RUDERFER: The Fallacy of Peer Review: Judgement without Science and a Case History

188 RUDER

I ER: T

IIK I'

A

I.I ACY O

l PEER REVIEW (a) The derivation is so imprecise and inconsistent it becomes without meaning. As an example, on [p.

389] phase is indicated to be constant because it is given as frequency X time in equation (6), but in equation (10) "frequency suddenly becomes time-dependent". Also T is treated as time dependent in equation (10) but it is later treated as an "inconsistent constant" in equation (11) upon removal from the integral. (b) In the extension of the analysis on [p.

398] to include the results of Earth's orbital motion, the effect of Sun's gravitational motion is neglected. Essential relativistic rate variations at a clock fixed on Earth are thereby omitted, resulting in a "seriously incomplete" analysis. (c) There is insufficient accuracy in the experimental data to detect the effects predicted; the author misconstrued "accuracy, precision and stability" as used in timekeeping. Predicted rate effects are of the order of (3 to 7) X X 10"

1

3, but UTCi cesium clocks have an intrinsic accuracy of only 2 X 10"

12 as stated by Allan, et al, Science, 191, 490 (1976) in rebutting the CannonJensen report. This is about an order of magnitude short of that necessary to measure the effects of the author. The same rebuttal more seriously notes that the UTC; time scales are intermittently adjusted in coordinating them to UTC so that each clock rate is not founded "on any fundamental physical process, as required by theory". Appendix E The Editor, Science. Dear Dr Abelson, Thank you for including the referees' comments with your 6 May return of my MS "One-Way Doppler Effects in Atomic Timekeeping". Both referees have improperly rejected the MS. The one submitting the 6-line comment (which I shall refer to as referee [C]) was grossly in error. The other [D] suggests to me the need for additional clarification in the MS. I have therefore revised the MS accordingly, including some cosmetic improvements, and am herewith resubmitting with the following comments. Except for the two revised pages substituted for page 7 [pp.

390-1] and the added insert for page 19 [p.

400], all other changes are marked in red. The discrepancy in Cannon and Jensen's data handling to which referee [C] appears to be referring is in regard to one station (RGO). I was aware of this discrepancy, which the referee could have ascertained if he had read the MS through, because I: (1) discuss this on page 14 [pp.401-2]; (2) refer to the Greenwich Observatory report which [C] reviewed (ref. 26); (3) explicitly exclude the RGO data from my analysis; and (4) was appraised of the cause of the discrepancy in my discussions with Cannon and Jensen. The "synchronisation procedures" mentioned by [C] consider only the absolute calibration of a station's proper time to the international second (SI); this process is unrelated to the one-way synchronisation process I discuss, which