370 
Miscellanea 
Contingency Table for General Intelligence in Sibliiigs. 
Category of Intelligeuce of First Sibling. 
Quick 
Intelligent 
Intelligent 
Slow 
Intelligent 
Slow 
Slow 
Dull 
Very 
Dull 
Totals 
Quick Intelligent 
312 
263-25 
131 -75 
40 
16-25 
4-25 
767-5 
Intelligent 
263-25 
876-5 
564-25 , 
172-5 
36-25 
14-75 
1927-5 
Slow Intelligent 
131-75 
564-25 
697-5 
249-5 
72-5 
27 
1742-5 
Slow 
40 
172-5 
249-5 
219-5 
80 
12 
773-5 
Slow Dull 
16-25 
36-25 
72-5 
80 
68 
18 
291 
Very Dull 
4-25 
14-75 
27 
12 
18 
24 
100 
Totals 
767-5 
1927-5 
1742-5 
773-5 
291 
100 
5602 
1 
The first question that arises is that of the method to be employed in the reduction of this 
table. The answer is fairly straightforward. The only legitimate method is that of corrected 
contingency. The mean square contingency of this 6x6 fold table is 
(/)2= -293,1833. 
Of the two methods of correcting this raw mean square contingency* for numlier of cells one 
leads to 
^•-=■286,9355 
and the other to 
(^- = -288,5659 
giving the nearly equivalent results for the contingency coefficient Co, 
0.2= -4722 and C'.= -4732. 
The class-index correlations are the same in both directions and we find : 
Hence finally we have for the correlation 
/• = -5147 and ?-=-5158, 
and accordingly it is amply adequate to take the correlation of siblings in general intelligence 
be -515, which agrees excellently with the value -508 found from Dr Gordon's data. 
But as the bald figures -508 and -515 convey little to the mind untrained to statistical 
appreciations, I have attempted to provide an illustrative diagram : see Plate VII. 
Assuming normal distribvxtion for the marginal totals and the arrays, I have superposed the 
means of the two systems (General Intelligence and Dr Gordon's Stanford Revision of the 
Binet-Simon tests) and equated their s.d.'s. Using i.q.u. for an intelligence quotient unit, i.e. a 
change of one digit in the intelligence quotient (or 100 mental ago /physical age), we find : 
Mean = 578-909 mentaces = 92-857 i.Q.u.'s, Standard Deviation = 95-5566 mentaces= 15-3215 i.Q.u.'s. 
Thus a mentace=-1604 i.q.u. 
* It is hoped to publish shortly the long-delayed memoir on contingency corrections. The delay 
has largely arisen from the labour involved in reducing adequate material by way of illustration. 
