€ bw 
residential structures, was assumed to be related to additional social 
characteristics. For example, two residential subareas might each 
have approximately sixty-percent single-units, detached, but the average 
density of one might be twice that of the other. Such a situation may 
spell significant differences in population density, ethnic composition 
and socio-economic status in general. Here again, while the photo data 
accuracy test (described below) required interpretation of total numbers 
of dwellings, in the later correlational analyses each sub-area was given 
a qualitative classification on this item, such as "high", "medium" or 
Mow" density. 
All four of these primary interest photo data items were assumed to 
be related to residential desirability and, in turn, to social structural 
characteristics. The first two do not require specialized photo inter- 
pretation skills. Descriptions of residential subareas by zonal location 
or land use are routine products of urban analysis. The accuracy of 
such photographically obtained information is not seriously questioned. 
However, quantitative descriptions of various housing types, estimates 
of percentages, densities, etc. require more detailed study and somewhat 
specialized techniques. Therefore, since these two housing items seemed 
quite crucial to the overall project, a systematic accuracy test was con- 
sidered necessary. The discussion which follows describes the procedures 
and results of this experimental task critically related to the objectives 
of the entire undertaking. 
Photo Data Accuracy Test 
The test centered on a sample of seventeen varied residential sub- 
areas in Birmingham, Alabama, Following the methodology developed in a 
pilot study, its purpose was to evaluate the accuracy of photo interpre- 
tation data describing various types of residential structures, Specifi- 
cally, the evaluative analysis correlated photo and ground observations 
concerning the last two items described above: Item 3, Prevalence of 
single-family homes; and Item 4, Average density of dwelling-units per 
block, 
In the photo data collection phase, several categories of housing 
types were recorded by code symbols on block outline cards; single 
dwelling-unit, detached; double-unit; multi-unit, 3-5; multi-unit, 6-8; 
multi-unit, 9-11, etc.; and Mother" types of mixed occupancy such as 
retail sales with dwellings attached. These residential types were 
identified by stereo-study with reference to previously developed re- 
cognition !'keys'! including such features as form and structure of roofs, 
yards and courts, driveways, entranceways, size, shape, and height of 
structures and their spatial relationships to other buildings. In the 
field data collection phase, surveys were made in the seventeen sub- 
areas, and the housing information was recorded on a duplicate set of 
block outline cards. The data were then organized to permit correlational 
analyses between the photographic observations (experimental data) and 
the ground observations (control data). 
5