By David Gopoian
Political performance should be evaluated against a benchmark of realistic expectations. As in a previous post, I use expected partisanship as such a benchmark to evaluate Kerry’s performance relative to expectations for a Democratic candidate in a given jurisdiction.
The expected vote measure is derived from an analysis of the voting behavior of national populations across five previous presidential elections. Estimates of the likelihood of a Democratic vote for specific categories of partisan identifiers are then applied to the geographic regions within state-level exit samples designated by the designers of the 2004 exit surveys.
The geographic designations created by Edison Mitofsky were done for sampling purposes primarily and often do not correspond to the most ideal geo-political boundaries for the purposes of political analysis. Findings may nonetheless prove instructive. Altogether, there were 202 geographic regions for the 50 states plus DC in the state-level exit polls.
The expected outcomes in these 202 regions were evenly split by partisan tendency, with 103 (51%) leaning Democratic and 99 (49%) leaning Republican. In southern and border states, 27 of 62 regions (44%) leaned Democratic. In northern states, 76 of 140 regions (54%) leaned Democratic.
Kerry received a majority of the two-party vote in 85 of the 103 regions that leaned Democratic (or 83% of them). Of the 27 regions in southern and border states that leaned Democratic, Kerry received the majority of the two-party vote in only 9 (or 30% of them). In northern states, where 76 regions leaned Democratic, Kerry actually exceeded that number, carrying 77 regions.
In a nutshell, Kerry was expected to win 51% of all regions nationally, and won in only 42%. In northern states, the Democratic nominee should have carried 54% of all regions, and Kerry captured 55%. In southern and border states, the typical Democrat should have won 30% of the regions. There, Kerry received the majority of the two-party vote in just 13% of them.
Of course in many of these regions, there was no visible Kerry campaign effort. Even such pivotal states as Colorado and Missouri were vacated early by the Democratic nominee. With the exception of Florida, the same may be said for every southern state. So the really intriguing data are those for the contested swing states.
Most of these findings may seem unsurprising, as they should given the stability of partisan tendencies across time. Perhaps the best showing of the Kerry campaign was in the Republican-leaning Philly suburbs, which proved essential to carrying Pennsylvania. In general, Kerry did better than expected in urban centers, pivotal to winning such states as Oregon, Washington, Minnesota, Michigan, and Wisconsin. Even in the uncontested states of Missouri and Colorado, Kerry held his own in urban centers. The Florida data suggest that the Kerry campaign succeeded in the Miami media market, but fell flat in Tampa Bay.
Some Notes on Central Ohio
Central Ohio, a largely rural, very white, very born-again expanse that covered parts of 5 Congressional Districts, each represented by a GOP House member (balanced only by Democratic-leaning Columbus), was the critical region that cost Kerry the state and the election.
In Central Ohio, Kerry gathered only 41% of the vote, three percentage points shy of expectations for the Democratic candidate. This is where Kerry lost the state and the presidential election. So it makes sense to focus a bit on Central Ohio.
Not surprisingly, vanilla is the flavor that best defines Central Ohio. More than 90% were white, and three-fifths were white Protestants. Approximately 40% were rural residents. But nearly half (45%) of the region’s voters were college graduates and their household median income approached $60,000.
The best-fitting model of the presidential vote for the Central Ohio region includes just two predictor variables: party identification and overall Bush approval ratings. Republicans outnumbered Democrats by 45% to 28%. Conservative Republicans outnumbered liberal Democrats by a ratio of three to one. Overall, 60% approved of Bush’s performance in office.
Attitude measurement has never been a forte of exit surveys, but the limited data available do indicate that voters’ perceptions of the candidates best able to handle the economy and terrorism and voters’ outlooks regarding Iraq correlated strongly with both Bush approval and party identification. So, too, did the major social marker and pivotal demographic in the life of Central Ohio.
At the end of the day, the key demographic that explained most of what may be observed about Central Ohio simmered down to religiosity. More than one-third of the region’s voters claimed born-again status, including 41% of all women and 28% of the men. This gender breakdown explains one of the major ironies of the vote preferences of central Ohioans – a reverse gender gap.
Among men, Kerry ran about one percentage point better than expected, and gained 45% of their votes, but among women Kerry ran eight percentage points below expectations and finished with only 37% of their votes. Kerry’s percentage of the vote from men who were less religious was two and one-half times what he received from born-again men; his percentage among less religious women was four times as great as the share of the vote he obtained from born-again women in the region.
Collectively, from voters who were not born-again, Kerry ran even with expectations, and took a slight majority of the vote (51%). From the 35% who were born-again, however, Kerry ran thirteen percentage points below expectations and took only 17% of the vote.
In Central Ohio, as in the nation as a whole, the social issues agenda of the Republican Party did not define the motivations of most voters. But more importantly than anywhere else in 2004, the moralistic agenda of the GOP generated the margin in Central Ohio that provided Bush with four more years.