Variability Due to Technique of the Sedimentation Index

Abstract

The striking variation in sedimentation rate as between individuals has led us to devote considerable attention to an analysis of this phenomenon. Studies of the influence of the technique employed and the interrelationships of the results by those techniques are proving fruitful in revealing important biological measures. Discussion of these results is being reserved until more extensive data can be accumulated, but it seems advisable at this stage to publish a note on the trustworthiness of a single determination of the sedimentation index.

Of fundamental importance at all times in the interpretation of the data of observation or experimentation is a knowledge of the variability to be expected in replicate measures of the same type. Not infrequently the worker in the fields of the so-called “exact sciences” is able to estimate a priori and from theoretical premises the error attaching to any method. Such estimates cannot, of course, be accepted as valid until verified in practical work, a fact which is all too often neglected. The biological worker can rarely follow such a procedure, if for no other reason than that the organism itself contributes an amount to the variability which can only be determined by the critical mathematical analysis of actual results.

Many techniques of making the sedimentation test have been in use during the past 10 years, but since the Cutler, 1 the Linzen-meier, 2 and Westergren 3 methods have had the greatest number of followers, the present discussion will be limited to these methods.

The extent to which any single measure, or average of replicate measures of blood sedimentation is reliable as a basis of generalizations must be dependent basically upon a precise determination of the variability of a large number of replicate determinations.