FARMERS* EARNINGS IN SOUTHEASTERN PENNSYLVANIA 47 
records obtained in the survey. This indicated that the eight factors 
measured over 95 per cent of all the causes which affected the acre 
value of these farms. 
The relation between the different factors and the farm value per 
acre was as follows: 
Farm value per acr e 
On the average, for each increase of: showed an increase o f 
10 dollars in dwelling value per acre 89. 75 
10 dollars in dairy building value per acre —.10 
10 dollars in other building value per acre 9. 29 
10 units in crop index 5. 60 
10 per cent in proportion of farm tillable 3. 96 
10 per cent in proportion of farm level 2. 47 
One unit in road type index — ( 73 
One mile in distance to town —5 47 
With the single exception of the value of the dairy buildings, these 
results seem fairly reasonable, although they would indicate that the 
dwelling and farm buildings, other than the dairy buildings, added a 
larger part of their value to the farm value than the previous study 
would have led one to expect. 
The relative importance of the different factors in determining 
farm values per acre in this area, as shown by their coefficients of 
determination (16 .footnote 1, p. 42), are as follows: 
Factor: Percent 
A. Dwelling value 30. 9 
B. Dairy buildings value — . 3 
C. Other buildings value 64. 4 
D. Crop index . 04 
E. Percentage of area tillable — . 4 
F. Percentage of area level . 4 
G. Type of road . 4 
H. Distance to town . 5 
Total 95. 9 
According to this statement, differences in the acre value of the 
farms were owing almost entirely to differences in the acre value of 
"other" buildings and the dwelling, and only in very small part to 
differences in the dairy barns or other factors. On the face of it, this 
seems an unreasonable conclusion, and indicates that there must be 
something seriously wrong with the way the problem was handled or 
the way the factors were "set up" for statistical examination. 
Closer examination of the original data showed this indication to 
be correct. Many of the farms growing mushrooms had only a very 
few acres of land, vet had expensive dwellings and large investments in 
mushroom houses. When these high values were stated as values 
per acre, they gave acre values for the farm, the dwelling, and 
'"other" buildings — principally mushroom houses — very much higher 
than the corresponding values for the general run of farms of the 
region. These extreme observations covered only a small portion 
of the total number of cases, yet by reason of their extreme variation 
were sufficient to overshadow completely the relations of the other 
factors on the more typical farms. 20 Table 40. which shows the 
frequency distribution cf several items for all the 422 farms included 
*° For a fuller discussion of the effect of extreme items of this sort in correlation analysis, see (13), 
pp. 40&-H)8. 
