\ 
308 The Philippine Journal of Science 1922 
presidentes,? or by the owners of the property. Therefore, al- 
though the figures do not represent actual valuations, they can 
reasonably be accepted as good approximations, accurately com- 
parable among themselves. 
The word “family” is herein used in the sense in which it is 
employed in the United States Census; that is, “a group of per- 
sons living under one roof, or in one suite or apartment.” (2) 
No distinction is made, however, between public and private 
families. A dwelling is considered as one roof, sheltering one 
or more families. 
Certain provinces have been omitted, because of incomplete 
and unreliable records. Manila, being a city, is also omitted 
for the simple reason that conditions obtaining there are not 
comparable with those of the provinces. This is particularly 
true in respect to the family factor, in which are included a 
large number of institutions such as prisons, colleges, hospitals, 
ete. Thirty-nine provinces are included in this work, namely: 
Abra. Masbate. 
Albay. Mindoro. 
Ambos Camarines. Misamis. 
Antique. Nueva Ecija. 
Bataan. Nueva Vizcaya. 
Batangas. Occidental Negros. 
Bohol. Oriental Negros. 
Bulacan. Palawan. 
Cagayan. Pampanga. 
Capiz. Pangasinan. 
Cavite. Rizal. 
Cebu. Romblon. 
Tlocos Norte. Samar. 
Tlocos. Sur. Sorsogon. 
Tloilo. Surigao. 
Isabela. Tarlae. 
Laguna. Tayabas. 
La Union. Zambales. 
Leyte. Zamboanga. 
Marinduque. 
THE CORRELATION METHOD 
A brief explanation of the correlation method may be of advan- 
tage to those not familiar with it. This method was first in- 
troduced on a sound basis by Galton,(4) and later elaborated 
by Edgeworth,(3) Pearson,(9) Yule,(11) and others. “Two 
characteristics are said to be correlated when there is a tend- 
ency for the changes in the value of one to depend on the 
* The presidente is an official corresponding to the mayor in an Amer- 
ican township or city. 
