108 MISC. PUBLICATION 218, U. 8. DEPT. OF AGRICULTURE 
arranged by counties could not be ascribed to differences among the 
constituent towns. This assumption was verified by selecting 6 
counties and 26 towns in those counties so as to obtain a random sample 
of the towns with the largest number of sales best distributed between 
the timber, cut-over, and farm property classes, and by considering 
assessment ratios in each town separately. While the results showed 
some irregularities because of the small number of cases in certain 
eroups, the ratios followed in general the same pattern as assessment 
ratios in similar groups in which all the cases were assembled by 
counties, and the relationships between the assessment ratios of the 
property classes were found to be similar. *® Thus it appeared that the 
summarization by counties did not seriously misrepresent assessment 
practice. 
ELIMINATION OF EXTREME RATIOS 
It has been pointed out that assessment ratios resulting from com- 
parison of appraisals and sale values with assessed values are subject 
to the possiblity of material error, in that there might have been a 
mistake in the appraisal, the selling price might not even approxi- 
mate market value, or there might have been failure properly to 
identify in the assessment books the property sold or appraised. Such 
errors are naturally more likely to be reflected in the extreme assess- 
ment ratios than elsewhere. However, since the assessed value is 
being tested in these studies by reference to the appraisal and sale 
values, it seemed necessary to use caution in making eliminations. 
Therefore these were limited to the extreme and widely separated 
assessment ratios, usually not more than 1 or 2 in a single county or 
town group. The method used is illustrated later in tables 30-38, 
where the discarded ratios are indicated by footnotes. 
APPRAISAL METHODS IN DETAIL 
A more detailed explanation of the method of making extensive 
appraisals may be of interest to some readers. Those used in the 
three States varied somewhat, and each State is therefore taken up 
separately in the following sections. 
MINNESOTA 
COLLECTION OF DATA 
The area, the owner’s name, and the assessor’s ‘‘full and true”’ value of each 
parcel within the selected townships were copied directly from the town assess- 
ment rolls of 1926 in the county auditors’ offices of the several counties. The 
records of the office of register of deeds in each county concerned, except St. 
Louis, were examined to get the location, the date, the area, and the parties to 
all warranty deed transactions since 1921. In St. Louis County these data were 
taken from the transaction cards of the supervisor of assessments. Wherever 
procurable, information concerning the physical condition of the selected towns 
was obtained before entering the field. Thus in St. Louis County the cruise 
records of the county assessor of unorganized territory and of various lumber com- 
panies furnished considerable data on cover, soil, rock, and other characteristics. 
Information from such sources was used to facilitate and supplement field exami- 
nation. 
The limited time allotted to this study and the comparatively large area to be 
covered dictated an extensive rather than an intensive field examination. It 
was found practicable to carry on most of the work from roads, which in the north- 
30 The procedure and results of this test are given in detail in Progress Report 12, pp. 10-11, and tables 12-17. 
See footnote 7, on p. 13. 
