Results Obtained in a Factor VIII Assay

Here the plasma slope indices are not sufficiently similar (ratio 70/111 = 0.64), and it may be seen that the higher index is associated with the higher plasma total. A log transformation would therefore be likely to make the fitted lines more parallel. The data are therefore rewritten in logs, as shown in Table 7, and the calculation repeated. Note that since the final calculations depend entirely upon differences (slope indices, and control total minus patient total), the characteristic can be reduced by 1 to lighten the manipulations. 

By this maneuver, the ratio between the individual slope indices has been raised to 0.92 so that the lines are now sufficiently parallel. Notice also the reduction in the 

From the right-hand part of Table 4, for testing three dilutions of two plasmas at a 6-fold step interval (n = 2, i = 6), Q = 1.04. From Table 7 above, d = 4.89 — 5.84 

Armitage, P., and Ingram, G. I. C., Studies in the variability of two blood-clotting tests. Appl. Statist. 7, 1-14 (1958). 

Baumgarten, W., Sanders, B. E., Belkin, B. D., Pagenkemper, F. E., Albers, W. G., and Ciminera, J. L., Antihemophilic factor its presence in fibrinogen, assay problems and the stability of lyophilized AHF-Fibrinogen preparations. Thromh. Diath. Haemorrhag. 9, 354r-367 (1963). 

---