ANALYSES OF EXPECTED urILITIEE IN A CAUSAL DECISION 'IHEORY*