ABSTRACT
Stephen It: Sanderson and Tatl{ J anhallen In their studies of the impact of ethnic heterogeneity on puhlic spending ill this "olume, Sanderson and Vanhanen hoth support the notion that ethnic heterogeneity inhibits \\"dl~lre spending. Howt'\Tr, Sanderson's data shO\\" a moderate to strong dlect of heterogeneity on spending, wherea~ Vanhanen's data show only a 'Try weak dltTt. In "iew of this discrepancy. an attempt "'as made to see "'hether til(> din~rgent findings could he reconciled. i\ third set of analyses was carried out in which Vanhanen's sample was used hut an attempt "'as made to replicate Sanderson's multiple regressions as closely as possible. TI1('re are three dint'rencl's between Sanderson's and Vanhanen's studies that could account Ii»' tht' dilkrential strength of their results. First, Sanderson used two measures of ethnic heterogeneity not used hy Vanhanen, and these measures appear to be more sensitin'. Second, Sanderson used li)l1r control "ariabks to preciict welfare spending, whereas Vanhanen lIsed only two (and Sanderson measured economic dewlopl1lt'nt hy Gl\;P per capita, whereas Van han en used the Human Dewlopment Index). Finally, Vanhanen used, in three of Ic)Ur cases, measures of \yell~l1'e spending dilkrt'nt from thost' used hy Sanderson.
