i 
, 
} 
: 
: 
J.— PSYCHOLOGY. I a er 
to the competent hands of Mr. Raper. When all was reported ready we 
met, I carrying the ‘ probable error’ of the tetrad-differences—that is, 
the median value that should by theory be expected to arise from sampling 
alone, he bringing the average value of the 3,003 tetrad-differences actually 
derived from the above table. My number was .061. His was .074; 
this, in order to get from the average to the median, had to be multiplied 
by the well-known constant .0845, whereby it came finally to .062 ! 
It may be of interest to survey the entire frequency distribution of the 
values concerned. 
Simpson’s Tetrad-Differences (37 subjects). 
The dotted curve shows the relative frequencies that should be expected 
from the sampling errors alone. About half should lie between a and b ; 
extremely few beyond cord. The continuous rectangles show the relative 
frequencies that actually occurred. 
----------------------- -- 
A better agreement of a theoretical frequency curve with one of actual 
observations does not, I venture to say, exist throughout psychology, 
or perhaps even throughout statistics. 
The preceding result may be instructively compared with another one. 
The doctrine of Two Factors, as is only proper, has had to make its way 
in the face of strong resistance. But the latter has curiously adopted 
two contrary lines of defence. The one is to question whether the mathe- 
matical criterion would really be satisfied by actual observation ; and this 
doubt, I hope, has been met by the facts just cited. But the other opposi- 
tion has instead asserted that any ordinary table of positive correlations 
_ would naturally satisfy it, so that such satisfaction must be devoid of 
_ Peculiar significance. And recently this second line of opposition has 
acquired much greater vitality, in that a table of correlations has now 
_ been brought forward as an actual example ; it is not derived from mental 
but from physical traits, and yet, it is said, exhibits a quite similar 
character. Here is the table : 
1925 N 
