Sociologists have increasingly turned to secondary data sources to stu
dy social and economic change.. As our experience in this area increas
es, so should our methodological rigor. Three common methodological si
ns-model mis-specification, inattention to regional influences, and fu
zzy operationalizations-are illustrated through a critique of Barnes a
nd Blevins' (1992) study of farm structure and economic well-being in
nonmetropolitan areas. This paper argues, and demonstrates empirically
, that many of Barnes and Blevins' conclusions are suspect because the
y did not include ethnic and regional variables in their regression st
udies.