Shelby Steele has an article in today's Wall Street Journal, "Hillary's Plantation." Predictably this has led some conservatives to moan that 90% of all blacks still vote Democrat and naysayers to claim Steele has no factual basis for asserting Hillary won't be able to hold black voters.   
 
My thoughts run on somewhat different lines.  What should we make of a situation in which the ranks of black conservatives seem to grow each year, yet it is reported that the GOP has made almost no inroads among black voters? Despite George W. Bush's huge outreach program, the media reported only a two percentage point increase in Bush's share of the black vote between 2000 and 2004, from 9% to 11%. A scant return indeed for the effort made. 
 
One question has gnawed at me since November, 2004:  How do we know these percentages are, in fact, an accurate reflection of party affiliation?  Since ballots themselves are secret, what sources were used? 
 
I can think of only two sources of data on black voting: polls and Boards of Election.

A quick Google search confirms the largest source in news stories is polling data, particularly exit polling on election day. Exit polling uses extremely small samples, a mere handful of precincts nationwide.  In addition, reports in 2004 showed exit polling to be highly vulnerable to manipulation, as the stringers responsible for approaching voters outside the assigned precincts allegedly did not do so at random. 

As for telephone polling, on a broad range of sensitive issues the numbers reported in the media increasingly seem to vary with the agenda of whatever organization paid for the poll. Neither polling method seems an entirely reliable basis for long term political planning — or for despair.
 
The other source is data from the thousands of Boards of Election across the nation. Many states require voters to register by party, but such voter registration data is often not officially tracked by race. Also, people tend not to re—register when their national party affiliation begins to drift, especially if local offices remain one party affairs fought out in the primaries. 

I live in a county where Democrats far outnumber Republicans in voter registration, yet Bush, the U.S. Senate candidate and our Republican Congressman all won handily. This makes voter registration statistics a lagging indicator of party affiliation in those jurisdictions where they can reliably be tracked by race....
 
As I analyze the situation, there is only one source of hard data on how blacks vote: the actual returns from those precincts in which an overwhelmingly percentage of all voters are black. While reliable, this data cannot be extrapolated, because to do so assumes that all blacks vote the same regardless of where they live. This is a very dangerous assumption, indeed, because in a free society few things tend to say more about a person's values than where they chose to live. 

So while we do know that blacks in all black urban areas overwhelmingly vote Democrat,  we do not know for sure how the many blacks who chose to live in mixed race neighborhoods actually cast their votes. All we have is the media's word on it.   
 
There are attractive black Republican candidates currently campaigning for statewide offices in Ohio, Maryland and Pennsylvania. Woe to us if we dwell on these dubious numbers and say all is hopeless instead of continuing to ask black Americans for their vote.

Rosslyn Smith   1 23 06