Advanced Statistical Methods in Biometric Research

Advanced Statistical Methods in Biometric Research 86 Reviews of Statistical and Economic Books [Part I, out. Some of these methods are apt to appear unnecessarily laborious and complicated to the biologist, who would often prefer simple calculations giving 90 per cent. efficiency on 100 animals to complicated calculations giving 100 per cent. efficiency on 90 animals. As Dr. Finney points out in this book, the division of effort between observation and computation may depend on what is the most economical way of getting the required precision. Some workers have avoided probits because of the complexity of the calculations, but this complexity is largely unnecessary. Dr. Finney gives several examples of iterative approximations, and in all of these the error which would be introduced by taking the first approximation is less than I per cent. of the inevitable fiducial range of the estimate. Such errors are negligible, especi­ ally as the result depends on the arbitrary assumption that P = 0·95. This is typical of the benefits which come from iterative calculations, except in cases where the results are so odd that the experiment ought to be repeated. Iterative calculations are rarely useful, and should not be recommended for routine work. The use of probits to http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of the Royal Statistical Society Series A (Statistics in Society) Oxford University Press

Advanced Statistical Methods in Biometric Research

, Volume 116 (1): 2 – Dec 5, 2018
2 pages

References (41)

ISSN
0964-1998
eISSN
1467-985X
DOI
10.2307/2980952
Publisher site
See Article on Publisher Site

Abstract

86 Reviews of Statistical and Economic Books [Part I, out. Some of these methods are apt to appear unnecessarily laborious and complicated to the biologist, who would often prefer simple calculations giving 90 per cent. efficiency on 100 animals to complicated calculations giving 100 per cent. efficiency on 90 animals. As Dr. Finney points out in this book, the division of effort between observation and computation may depend on what is the most economical way of getting the required precision. Some workers have avoided probits because of the complexity of the calculations, but this complexity is largely unnecessary. Dr. Finney gives several examples of iterative approximations, and in all of these the error which would be introduced by taking the first approximation is less than I per cent. of the inevitable fiducial range of the estimate. Such errors are negligible, especi­ ally as the result depends on the arbitrary assumption that P = 0·95. This is typical of the benefits which come from iterative calculations, except in cases where the results are so odd that the experiment ought to be repeated. Iterative calculations are rarely useful, and should not be recommended for routine work. The use of probits to

Journal

Journal of the Royal Statistical Society Series A (Statistics in Society)Oxford University Press

Published: Dec 5, 2018