These initial studies applying scale analysis to areal units 
of the city led to more intensive investigations of data for Birmingham. 
Although thls particular phase of the research has been reported 
elsewhere,” the findings may be briefly reviewed in the present 
summary paper. Twenty-eight census tracts, comprising about fifty- 
percent of the city area, were chosen to represent the fifty-eight such 
units in Birmingham. Four physical data items, as described above, 
were employed in the development of a scale of residential desirability. 
The eleven "response"! categories contained in the three trichotomous 
items and one dichotomous item were assumed to define varying degrees 
of neighborhood status ascription. Analysis of the resulting total 
of 112 responses (four per tract) followed the procedural steps outlined 
by Guttman's "Cornell technique".ll The findings demonstrated that the 
physical structural categories constitute a single continuum. The 
resulting scale, having a coefficient of reproducibility of .93, revealed 
an internally consistent ranking of the twenty-eight tracts in eight 
scale types ordered from high to low on this dimension of residential 
desirability. | 
The second phase of this Birmingham study hypothesized that the 
several social data categories, describing the same census tracts, and 
referring generally to the social stratification system of the city, 
would also be scalable. Such a scale could be called a continuum of 
socio-economic stytus., Five trichotomous social data items were con- 
structed: 1. Median annual income; 2. Prevalence of within-dwelling 
crowding; 3. Prevalence of home-ownership; 4. Prevalence of social dis- 
organization; 5. Educational achievement. The categories for these 
items were arranged by a system of ranks. To illustrate, the highest 
seven ranks of the tracts on income were classified as "positive" 
(favorable) in terms of socio-economic status, the middle fourteen ranks, 
"neutral", and the lowest seven, "negative", etc, In this case, also, 
scale analysis of the 140 responses demonstrated that the five tri- 
chotomous items represent a single scalable universe of content for 
these census tracts. The resulting scalogram, having a coefficient of 
reproducibility of .92, was defined as a continuum of socio-economic 
status, 
Each of the two scales provides an empirical definition of com- 
monalities among urban spatial patterns, one referring to physical 
structure and the other to social structure. In each case, the scale 
types account for the joint relationships among the all data categories 
  
10. Norman E, Green, "Scale Analysis of Urban Structures: A 
Study of Birmingham, Alabama", American Sociological Review, XXI 
(February 1956), pp. 8-13. 
ll. Louis Guttman, "The Cornell Technique for Scale and Intensity 
Analysis", Educational and Psychological Measurement, VII (Summer. 
1947), Pp. 247 - 280. 
12