Beschreibung:
<jats:p> It has been suggested that differences in voter recognition of candidates may cause variations in partisan defection rates, which could account for chamber differences in reelection rates. This defection hypothesis is tested here using survey data from 1956-1980. The rates at which party identifiers defect from their party's Senate and House nominees are compared. Although the pattern for all Senate races parallels that for all House races, there have been differences in the anticipated direction since 1970. Defection patterns in recent Senate elections have not, however, been less pro-incumbent, which is what the high proportion of defeats among incumbent senators would lead us to expect. </jats:p>