BRANCH PLANTS AND POVERTY IN THE AMERICAN SOUTH

An alternative to the widely accepted hypothesis that ''big business'' has a negative impact on community welfare is tested with data from 4 45 nonmetropolitan counties in the American South. We argue that the a ppropriate counterhypothesis is not that branch plants have a positive impact on community welfare, because that and similar formulations si mply perpetuate the nonsociological approach of the early hypotheses. An interaction formulation that sees community structure as mediating the impact of branch plants better explains differentials in our three criteria of welfare-per capita income, percent below the poverty line , and infant mortality. This explanation predicts positive welfare whe n pluralism and urbanization are high and branch plants are frequent, and lower welfare when one or both of these is low. Tests using discri minant analysis support the interaction hypothesis while providing lit tle evidence for either the direct positive or negative impact hypothe ses.