COKN AND HOG CORRELATIONS 
13 
Table 2. — The mean values of the corn and hog variables for the period 1889 to 
1918, inclusive, and the standard deviations for this period and for the longer 
period 1871 to 1913, inclusive, in those cases in which the data were available 
Standard 
Standard 
Mean, 
1889-1913 
deviation 
deviation 
Item 
irom 
trend, 
from 
trend, 
1889-1913 
1871-1913 
Corn acreage 
million acres.. 
89.8 
3.6 
3.1 
Corn yield per acre . 
bushels.. 
25.6 
2.9 
2.9 
Corn crop 
million bushels.. 
2.310 
299 
260 
Corn price per bushel December 1 
cents.. 
42.6 
8.3 
8.3 
Western pack: 
Summer price per 100 pounds 
.dollars.. 
5.61 
0.89 
10.90 
Winter price per 100 pounds 
....do.... 
5.28 
0.99 
1.03 
Summer live weight. 
Winter live weight... 
pounds.. 
do.... 
227.4 
7.0 
228.5 
7.4 
7.4 
Summer pack 
thousands.. 
12. 591 
1,400 
1,171 
Winter pack 
do.... 
8,875 
1,166 
984 
Weight of summer pack 
... million pounds. . 
2,847 
313 
Weight of winter pack 
do 
2,010 
253 
226 
Annual pack: 
Western 
thousands.. 
21, 466 
2,307 
1,947 
Eastern 
..do.... 
5,709 
469 
1417 
Farm price per head, January 1 
dollars.. 
6.19 
0.96 
0.93 
i 1880-1913. 
Owing to the small number of years available for study the prob- 
able errors of the correlations are high. The formula for the probable 
where r is the correlation in question and n is the 
error is 0.6745 - 
number of entries. For correlations differing only slightly from 
based on 43 years the probable error is ±0.10. If based on only 25 
years, as in the case of summer weight and summer pork, the prob- 
able error is ±0.13. These figures are reduced in the case of larger 
correlations. Correlations of less than ±0.30 or —0.30 have little 
absolute significance. Even smaller correlations, however, have 
some significance relative to each other, so far as they are based 
on the same data. 
In finding the degree one quantity is determined by a number of 
others, use has been made of Pearson's 5 coefficient of multiple corre- 
lation, the square of which measures the degree of determination 
by the group of factors under consideration. 
The deviation from the trend to be expected in the dependent 
variable can be predicted from the known deviation of the other 
variables chosen in preceding years by the formula for multiple 
regression. 
In dealing with the nature of the causal relation among the varia- 
bles in a given correlated system, the method of path coefficients 
is used. 
THE CORN CROP 
The corn crop is necessarily completely determined by the acreage 
and yield. Of the two factors, variations in yield have had a much 
more important determining influence than variations in acreage. 
The correlation between crop and yield is ±0.87, between crop and 
acreage only ±0.49. These coefficients indicate 64 per cent determi- 
* Pearson, Karl, Mathematical Contributions to the Theory of Evolution. Regression, Heredity, and. 
Panmixia. In Phil. Trans. Roy. Soc. London, Series A, v. 187, 1896, p. 253-318. 
