The low convergence between survey and institutional measures of organization structure reported in three recent studies challenges assumptions about the interchangeability of these two approaches. Current formulations of convergence appear to be too broad, and this inadequacy has contributed to the low convergence observed in these studies. Several issues are relevant to expectations of congruence between alternative measures of organization structure: the focus (organization, role, or job); the technique (survey or institutional); and the data source. Convergence cannot be expected between measures that focus on different referents or that apply at different levels of analysis.