Numerous political analysts have invoked the "Bradley Effect", a racially driven phenomenon, to explain the disconnect between pollster predictions and voter turnout for Senator Obama. Mr. Novak's Op-ed entitled, "Obama's Bradley Effect", was disappointing and shortsighted. Such opinions make me wonder if the media will ever be able to analyze election dynamics through a real-time prism rather than simply extrapolating historical anecdotes to explain the events of today.
I also wonder if Mr. Novak explored alternative explanations for the outcome of the California election. First, Clinton won California by 8%, which equates to approximately 350-400,000 votes. Californians cast nearly 1 million absentee ballots and over 50% of them were cast long before the February 5 election and before Obama gained momentum in the state. Were absentee voters included in the polling samples? If so, were they queried about how they voted? It is unlikely that late polling captured the full effect of the absentee vote. Therefore, it would be impossible to accurately predict the election outcome. Second, nearly 200,000 independent or "declined-to-state" (DTS) voters met confusion while casting ballots in California. Consequently, tens of thousands of votes in California are not counted.
While the Bradley effect may be a partially plausible explanation for the California election outcome, there are equally more logical explanations for the disconnect in the polling. Let us consider all the options rather than focus on one based on possibly outdated voter behavior patterns.
