Bruce H. Mayhew, J. Miller McPherson, Thomas Rotolo & Lynn Smith-Lovin
Mayhew, B. H. , McPherson, J. M. , Rotolo, T. & Smith-Lovin, L. (1). Sex and Race Homogeneity in Naturally Occurring Groups. Social Forces, 74(1), 15. http://dx.doi.org/10.2307/2580623
We generate a number of hypotheses about face-to-face groups using the energy distribution principle: the frequency of an event is inversely related to the amount of energy expended in that event. The principle predicts that (1) the size of groups will be inversely related to the frequency of their occurrence; (2) at any group size, the composition of social positions will be less heterogeneous than chance; and, (3) as group size increases, observed compositional homogeneity will decline at a slower rate than chance. We test these hypotheses using data on more than 100,000 naturally occurring, public, face-to-face groups gathered in sampling sweeps through two communities over a three-year period. The data support the hypotheses and yield interesting differences in the strength of sex and race heterogeneity. We discuss the findings as they relate to the general energy distribution principle and to other sociological perspectives.
The data supports the first hypothesis which states that the size of a group is inversely related to the frequency of its occurrence according to the energy distribution principle. Larger groups (of two or more individuals) will be less likely to occur than a group of size one because as the group size becomes larger, the larger the energy and time spent to monitor and coordinate action with others. Data supports the second hypothesis which states that the composition of social positions will be less heterogeneous than chance in any size of group. Individuals from different social backgrounds will be less likely to interact with each other because more energy and time must be expended in heterogeneous groups just to check for social signals and codes. Findings suggest that the rate of homogeneity declines slower than predicted by the third hypothesis, which predicts homogeneity to increase at increasing rates with group size. The factor that predicts the most bias towards group homogeneity is race.
Description of method used in the article
Every morning and afternoon, observation sweeps were used every day for one week in Winnsboro and Columbia, South Carolina to gather information on observed groups in public space. This process was repeated four times to produce 56 data sets per town. Observations were made by sweeping the entirety of Winnsboro and a shopping center directly outside of the town. In Columbia, most observations were made in an area encompassing the state capitol, principal business district, a university, the state fairgrounds, and two malls. Observers drove or walked between locations, stopping periodically at street corners to note the number and sexual and racial demographics of groups. Observation error was mitigated by only recording groups within 20 meters of the observer.