IV. The Second Wave: Replications of Shaw and McKay
Shaw and McKay’s (1942) research generated several replications spanning more than a decade. Each added to the knowledge base of ecological literature by examining the relationships first considered by Shaw and McKay in slightly different ways. None of the replications, however, drew substantially different conclusions from those in the original study.
Lander (1954) correlated 8,464 juvenile delinquents tried in the Baltimore Juvenile Court from 1939 through 1942 with demographic variables taken from the 1940 census. Specifically, Lander analyzed juvenile delinquency in terms of the median years of school completed, median monthly rent, homes with 1.51 or more persons per room, homes needing substantial repairs or having no private bath, foreign-born and non-white residents, and owner-occupied homes.
Lander’s (1954) findings followed the concentric ring pattern established by Shaw and McKay (1942). Lander noted, however, that the use of 1-mile increments for the zones oversimplified the spatial distribution of delinquency because it obscured the range of delinquency rates within each zone.
Lander’s (1954) findings did not support Shaw and McKay’s (1942) correlation between high delinquency rates and close proximity to industry. His results indicated that the delinquency rate in census tracts with less than 50% of the area zoned for industrial purposes was lower than the city average. Lander, however, found a more pronounced relationship in Baltimore in areas zoned for commercial use. He concluded from these findings that Shaw and McKay were correct in identifying areas close to the center of the city as the highest in delinquency but that this was primarily due to ecological factors other than proximity to industry. Lander also found no support for the correlation between population change and delinquency. Lander’s conclusions are not wholly contradictory to those of Shaw and McKay, however. His findings showed that the tract with the third-highest delinquency rate had a population increase of 20% and that tracts with population increases of 40% or more and decreases of 20% or more had substantially different delinquency rates than those with little or no population change. Lander found a substantial (r = .69) but nonsignificant relationship between delinquency and substandard housing. He added overcrowding as an additional measure of the physical status of the area and found a substantial (r = .73) but nonsignificant relationship between overcrowding and delinquency.
In Lander’s (1954) analysis, the median rental value of housing units in Baltimore was not significantly related to delinquency. Lander reasoned that economic variables such as rental values were an unreliable predictor because they were merely indicators of where a person might live. Lander did find a significant relationship, however, between homes owned in an area and delinquency. In fact, home ownership was the most highly correlated variable in Lander’s analysis.
Lander’s (1954) analysis of population status followed Shaw and McKay’s (1942). Zero-order correlation of the variables demonstrated that these variables were better predictors of delinquency than physical or economic variables. Although Lander’s conclusions generally supported those of Shaw and McKay, there were some differences in the findings. For example, Lander found a statistically significant, inverse relationship between delinquency and number of foreign-born residents. Lander explained this by noting that many of the foreign-born Chicago residents were recent immigrants, whereas in Baltimore most of the foreign-born residents were well integrated into the community, characterized by a high degree of home ownership. Lander also found that in areas with a moderate proportion of blacks there was a high rate of delinquency. As the percentage of blacks rose above 50%, however, the rate of delinquency dropped proportionately.
Bordua (1959) attempted to replicate part of Lander’s (1954) study in an effort to clarify some of the issues that had drawn criticism. Bordua’s study used delinquency data from the Detroit, Michigan, juvenile court for 1948 through 1952 and census tract data from the 1950 U.S. Census.
Bordua’s (1959) physical status analysis only included substandard housing and overcrowding. His findings were generally supportive of Lander’s (1954) and contradictory to Shaw and McKay’s (1942). Bordua found a weaker but significant relationship between overcrowding and delinquency. Also supporting Lander and counter to the findings of Shaw and McKay, Bordua found a nonsignificant relationship between substandard housing and delinquency.
Bordua’s (1959) findings regarding economic status essentially supported those of Lander (1954). Bordua found the median rental value to be nonsignificant and less substantial than Lander did. Bordua also found a significant but less substantial relationship between the percentage of homes owned and delinquency. Bordua added median income to represent economic status in the analysis and found that income was not a statistically significant indicator of delinquency.
Bordua’s (1959) analysis of population variables was supportive of Shaw and McKay’s (1942) research but contrary to Lander’s (1954). Bordua’s findings revealed that foreign birth was significantly related to delinquency but that number of black heads of households was nonsignificant. On the basis of these contradictions, Bordua chose the ratio of unrelated individuals to the total number of families as an additional measure of population status. Lander found that unrelated individuals was significantly correlated with delinquency.
Chilton (1964) used juvenile court data from Indianapolis, Indiana, from 1948 through 1950 and data from the 1950 U.S. Census to compare the findings of Lander (1954) and Bordua (1959) with those in Indianapolis. The results of Chilton’s analyses of the relationship between physical characteristics and delinquency essentially confirmed the findings of the other replications. Chilton’s findings of the relationship between delinquency and substandard housing showed a substantial but nonsignificant correlation with delinquency. Chilton also found a substantial correlation between overcrowded conditions (more than 1.5 persons per room) and delinquency. Unlike the other two studies, though, the degree of overcrowding in Indianapolis was one of two statistically significant indicators of delinquency. Chilton’s analyses of economic variables essentially confirmed those of Lander’s and Bordua’s studies. The relationship between median rental value and delinquency was found to be nonsignificant and similar to Lander’s. Chilton’s findings concerning home ownership also supported the other replications. His findings related to population characteristics tended to refute both Shaw and McKay (1942) and the other replications. Chilton found both percentage of foreign-born people and percentage of black people to not be significantly related to delinquency in Indianapolis. He concluded that ecological research can identify general conditions associated with delinquency but that differences between cities exist such that they cannot be addressed with traditional social disorganization theory.
The findings of Lander’s (1954), Bordua’s (1959), and Chilton’s (1964) studies suggest that although the relationship between the physical characteristics of an area and delinquency may vary by city there appears to be a sustained relationship at some level. Shaw and McKay’s (1942) findings concerning the relationship between economic characteristics and delinquency were supported by the replications, but not completely. Finally, Shaw and McKay’s findings concerning population characteristics and delinquency were generally not supported by the replications.