The Return of Jim Crow

The Blogosphere has a bit of an obsession with Diebold voting machines. Many worry that the corporate control over such machines could lead to a hacker-led, corporate funded conspiracy to create massive election fraud. In contrast to this science fictionesque possibility of a conspiracy, there are real, massive, ongoing purges of the African-American vote around the country. Through "spoiled ballots" and felony disenfranchisement, Republicans (and even some Democrats) have successfully denied voting rights to as many as 2.4 million voting-age African-Americans:

In the entire nation 1.4 million black men with sentences served can't vote, 13 percent of the nation's black male population...

The Voting Rights Act of 1965 guaranteed African-Americans the right to vote--but it did not guarantee the right to have their ballots counted. And in one in seven cases, they aren't...

Take Gadsden County. Of Florida's sixty-seven counties, Gadsden has the highest proportion of black residents: 58 percent. It also has the highest "spoilage" rate, that is, ballots tossed out on technicalities: one in eight votes cast but not counted. Next door to Gadsden is white-majority Leon County, where virtually every vote is counted (a spoilage rate of one in 500).

Harvard law professor Christopher Edley Jr., a member of the Commission on Civil Rights, didn't like the smell of all those spoiled ballots. He dug into the pile of tossed ballots and, deep in the commission's official findings, reported this: 14.4 percent of black votes--one in seven--were "invalidated," i.e., never counted. By contrast, only 1.6 percent of nonblack voters' ballots were spoiled.

In the 2000 election, 1.9 million votes cast were never counted. Spoiled for technical reasons, like writing in Gore's name, machine malfunctions and so on. The reasons for ballot rejection vary, but there's a suspicious shading to the ballots tossed into the dumpster. Edley's team of Harvard experts discovered that just as in Florida, the number of ballots spoiled was--county by county, precinct by precinct--in direct proportion to the local black voting population.

Florida's racial profile mirrors the nation's--both in the percentage of voters who are black and the racial profile of the voters whose ballots don't count. "In 2000, a black voter in Florida was ten times as likely to have their vote spoiled--not counted--as a white voter," explains political scientist Philip Klinkner, co-author of Edley's Harvard report. "National figures indicate that Florida is, surprisingly, typical. Given the proportion of nonwhite to white voters in America, then, it appears that about half of all ballots spoiled in the USA, as many as 1 million votes, were cast by nonwhite voters."

With laws such as the "Help America Vote Act," conservatives continue to score victories when it comes to disenfranchising African-Americans. With as many as 2.4 million African-Americans being denied the right to vote, one of the most important triumphs of the civil rights movement is in danger of being reversed. When Jim Crow collapsed in the sixties, its sympathizers went on to find another way.

Do Not Give Up on the South

Back in late October and early November when Dean was taking flak for earlier comments about confederate flags, pickup trucks and the South, large elements in the Dean-leaning blogosphere vented some rather bigoted anti-Southern feelings. Even before and since then, many Democrats have expressed frustration about Democratic prospects in the South, and suggested not even trying to win Southern states in the general. There were even rumors back in February that Kerry had no plans to win any of the old eleven states, and would instead focus on pickups in the Midwest, especially Missouri and Ohio.

Well, in case you have ever thought about giving up on the South, take a look at this breakdown of the Democratic and Republican coalitions. These are the percentages of where the voters for each party live. The numbers are derived from 2000 exit polls:

	     DNC     GOP
Northeast     26.4    18.5
Midwest       25.6    26.3
South	     27.3    35.2
West	     20.7    20.0
There are just as many Democratic voters in the South are there are in the Northeast or Midwest. While this is somewhat due to the South being the largest of these four regions and to the large numbers of African-Americans in the South, it does not change the fact that the Southern wing of the Democratic Party should be just as influential and just as focused upon as party members from other regions of the country. Arkansas, Florida, Louisiana, North Carolina, Tennessee and Virginia are all states Democrats can win in a general election without a blowout, and even without a Southerner on the ticket. Kerry is on the air and competitive with Bush in all six of these states right now.

This is also reason number 10,257 that Zell Miller can ram it. Clearly, the Democratic Party is the national party, and the Republican Party is a national party no more. Democrats are well balanced across the country, while Republicans have told the Northeast to go to hell.

Green Clovers, Blue Diamonds and Purple Horseshoes

Democracy Corps has produced a sprawling, complicated, quirky, and cloying snapshot of where the campaign currently stands. It features many graphs, analysis of future directions of the campaign, and twenty-two bizarrely named demographic groups that make up the country. Here are those demographic groups (I removed most of the annoying nicknames):

GOP World			       2004 Party ID
White, Protestant, Evangelical	 GOP +49	       
Young, married, no college and male   GOP +31		       
White, Married, College and Male      GOP +27		     
Deep South			 GOP +22		
Outer Suburbs			 GOP +19		
Married Blue Collar Women		 GOP +13
White Voters 30 or under		 GOP +13 
White, no college, male and senior	 GOP +10		

Contested World
White Rural			 GOP +9
Devout Catholics			 GOP +8
Aging Blue Collar Males 		 GOP +6
College-educated women		 GOP +5
Post-Graduate Men			 GOP +3
Aging Female Blues			 DEM +5
Golden Girls			 DEM +7

Democrat World
Unmarried, no college women		 DEM +12
Cosmopolitan States		 DEM +15
Union Families			 DEM +16
Hispanic				 DEM +24
Post-graduate women		 DEM +26
Non-church going, non-gun owners	 DEM +38
African-Americans			 DEM +81

I switched the categories of some groups, since it seemed strange to me that many "contested" groups in the survey swung more toward one party or another than some of the "non-contested" groups. Almost every group except "white voters under 30" and "married blue collar women" are turning away from Bush. So far, the biggest swing has been among "secular warriors" (no church, no guns), who have swung more than 13 points against Bush since 2000.

The Democratic base consists primarily of union families, non-whites, women, as well as people who live in diverse areas, do not own guns and do not go to church (no wonder conservatives hate Democrats). The Republican base consists primarily of devout, non-union white people who do not live in cities (no wonder I don't know many Republicans).

Personally, I fit into four of these categories, two of which are "Democrat world," one of which is "contested world," and one of which is "Republican world." I suppose that makes me a post-graduate, secular warrior, union family, young and restless male. Check out the study to see what nicknames have been cooked up for you.

All State Voting Trends Since 1976

I spent literally all day and most of the night on this chart, so please check it out. You will need to scroll down past the General Election Cattle Call stuff to view it.

The partisan index shows the relative standing of Democrats and Republicans in a given state by comparing the state popular vote of the two parties with the national popular vote of the two parties. For example, in New Jersey in 1992, Clinton won with 42.954% to Bush's 40.581%. However, in 1992 the partisan index favored the GOP by 3.2, since nationally Clinton had 43.007% of the vote and Bush had 37.448%.

1992 NJ
DNC  43.007-42.954 =  0.053
GOP  40.581-37.448 =  3.133
Total		    3.186 (3.2)

In theory, this means that Clinton would have won the state had he only won the national vote by 3.2, but lost it if he won by 3.1. In reality, local organization and targeting play a small, unquantifiable, but noticeable role in the partisan index. Further, for most candidates, being a "favorite son" makes a large impact on the race. Also, as a candidate rises of falls in the national polls, s/he does not do so across the board, evenly in all states.

While the primary use of this chart will be to better understand this election, it also has clear historical uses, and even goes a long way to answering some common questions. For example, I think it shows that broad trends are more powerful than local targeting. Also, I think the chart conclusively shows that Perot did not cost Bush the 1992 election. Considering that Bush lost nationally by 5.56%, where does he find the necessary 102 electoral votes? Colorado, Georgia, Montana, Nevada, New Hampshire, Tennessee and maybe Kentucky are the only states it appears Perot may have handed to Clinton. Taken together, they are not even half of what Bush needed, as the final total would have been 319-219 instead of 370-168 (for a more detailed discussion on this matter, click here). In the other direction, Perot almost cost Clinton Connecticut and Ohio.

There will be a permanent link to this chart soon in a "Presidency 2004" section that will include lots of other goodies. There is some cool stuff happening at MyDD.

Do Days of the Week Matter in Polls?

Here's one from the improbable research files.

Using the results from the three Bush vs. Gore tracking polls in 2000 (Gallup, Zogby and ABC), I have calculated the simple mean of the total results for each canddiate on every day of the week:

Sunday (18 data points)
Bush: 45.44
Gore: 44.67

Monday (17 data points)
Bush: 45.06
Gore: 44.88

Tuesday (15 data points)
Bush: 44.80
Gore: 44.80

Wednesday (16 data points)
Bush: 45.06
Gore: 44.44

Thursday (18 data points)
Bush: 45.33
Gore: 44.22

Friday (18 data points)
Bush: 46.00
Gore: 44.00

Saturday (18 data points)
Bush: 46.00
Gore: 43.94

Bush's daily mean varied by 1.2, from a high of 46.00 on Friday and Saturday to a low of 44.80 on Tuesday. Gore's daily mean varied by 0.94, from a high of 44.88 on Monday to a low of 43.94 on Saturday. The difference between Bush and Gore varied by 2.06, from a high of 2.06 on Saturday to a low of 0.00 on Tuesday. Overall, the difference is small, but noticeable.

I would chalk this difference up to simple margin of error if it did not follow such a clear pattern. Gore rises in relation to Bush on Sunday, Monday and Tuesday, while Bush rises in relation to Gore on Wednesday, Thursday, Friday and Saturday. The only number keeping this pattern from being perfectly smooth is that Gore drops from 44.88 on Monday to 44.80 on Tuesday.

This pattern seems to go beyond simple margin of error. Do Democrats poll slightly better early in the week, and Republicans slightly better later in the week? If so, why? Do slightly more Democrats go out later in the week so there are fewer Democrats to be polled? Do more Republicans go out to watch Monday Night football?

Frankly, I'm stumped.


Advertise Blogads