On August 26, 1864, having withdrawn from trenches and fortifications outside Atlanta the previous day, U.S. General Sherman sent most of his forces westward around Atlanta and toward the south of the city.
On August 26, 1920, the Nineteenth Amendment to the United States Constitution was adopted. Ratification took place on August 18, 1920, as the Tennessee House of Representatives adopted it, but adoption became official on August 26, when United States Secretary of State Bainbridge Colby certified the Amendment. It reads:
The right of citizens of the United States to vote shall not be denied or abridged by the United States or by any State on account of sex.
Congress shall have power to enforce this article by appropriate legislation.
On August 26, 1939, the first televised major league baseball game aired, as the Brooklyn Dodgers and Cincinnati Reds split a doubleheader in Ebbets Field.
On August 26, 1961, the 718th Engineer Light Equipment Company of Fort Valley and the 210th Signal Base Depot Company of Augusta were called up to take part in the American response to the crisis in Berlin.
President Lyndon B. Johnson was nominated for President by the Democratic National Convention in Atlantic City, New Jersey on August 26, 1964.
On August 26, 1965, Sonny & Cher were at No.1 on the UK singles chart with ‘I Got You Babe’, the duo’s only UK No.1. Sonny Bono was inspired to write the song to capitalize on the popularity of the term “babe,” as heard in Bob Dylan’s ‘It Ain’t Me Babe’.
On August 26, 1996, President Bill Clinton signed a Welfare Reform bill, called the Personal Responsibility and Work Opportunity Reconciliation Act of 1996.
More easy listening and a question for readers
Last week, we had a great discussion of the week’s developments in Georgia Politics. You can listen on Fridays at 3 PM on 88.5 FM in Atlanta or on the GPB radio network statewide. They’ve also started posting the show on the GPB website.
Here’s my question for you: do you listen to podcasts? If so, what are your favorites?
Elaine Boyer resigns from DeKalb Commission
Yesterday, Elaine Boyer resigned her seat on the DeKalb County Commission, where she has served since her election in 1992. From WSB-TV:
Elaine Boyer touted herself as a crusader for fiscal conservatism, railing against government waste and blasting other elected officials’ spending.
But late Monday, with an FBI investigation pending and questions mounting about her own spending of taxpayer money, the longtime DeKalb County commissioner abruptly resigned.
“I’ve betrayed the people and I’ve abused my position of power,” she told Channel 2 Action News in an exclusive interview.
“It’s a very hard decision, and I’m heartbroken and saddened, but I need to resign,” she said.
The federal investigation started with an investigation by the AJC:
The first part of the newspaper’s investigation, published in March, revealed that she had spent nearly $17,000 in taxpayer funds on herself and her family using her county Visa card. Her purchases included airline tickets, a ski resort booking, a rental car and personal cell phone expenses.
Following the March stories, the FBI launched an investigation into the commission’s discretionary spending. Federal prosecutors subpoenaed thousands of documents in June related to purchasing card spending by county commissioners as well as nearly 300 county employees.
The AJC also has continued drilling into questionable office expenses, pressing Boyer for answers on more than $90,000 in payments she authorized to consultants. She has been unable to produce any reports, memos, correspondence or other evidence of the consultants’ work.
Boyer ran up the questionable expenses as she and her husband, John, were having personal financial problems. In 2009, they lost a Stone Mountain strip mall they owned to foreclosure after defaulting on an $886,600 loan, court documents show. Her husband, a chiropractor, had an office there.
In the past three years, she has filed for bankruptcy twice and her husband has filed for bankruptcy five times, subsequently dropping cases each time in an apparent tactic to stall foreclosure of their home in Smoke Rise, near Stone Mountain.
Despite the moves, in April a bank foreclosed and the family was ordered to vacate by this week. The house appeared to be empty on Monday, and a neighbor said a moving truck backed into the driveway on Sunday.
She said she could not explain why she repeatedly pulled out her county P-card to pay for things that had nothing to do with DeKalb County business.
“I don’t have a reason,” Boyer told the AJC in March. “I’ve never had any intent of doing anything (wrong). I have been totally honest and trying to be transparent in returning the funds.”
When serious questions came up about Boyer’s own activities in office, she claimed ignorance of the rules. She said she didn’t think the county’s P-card policy, which forbids using the cards for personal purchases, applied to elected officials.
“It never dawned on me that what I was doing was wrong,” she said in a March interview with WSB.
I’ve known Elaine Boyer for much of her time in office – some 16 years – and considered her a positive influence on the Board of Commissioners for many years. When something like this happens to a political opponent, it’s easy to demonize them, but it’s harder to figure out when it’s someone who has been a friend.
We’ll be discussing weighting of polls in more depth, as it’s been brought up in regards to a poll by Landmark Communications that appears to be an outlier in its assessments of the elections for United States Senate and Governor of Georgia.
First, some perspective from another source. Georgia’s Republican National Committeewoman Linda Herren sent along the following information from the Republican National Committee when she asked them about recent polling in Georgia. It’s excellent information.
Our internal track of the GA polls has Perdue up about 4 pts. The latest poll I have seen shows Purdue up about 8 pts and the general trend of polls for Purdue is good. Having said that the poll you site does have African American turnout #’s at 29%. That is about 4 pts higher than the historic average but I don’t know that it is impossible. (historic #’s below) This is their game plan and this is what they are doing.
We know they have bodies on the ground concentrated in AA areas of ATL. We know that $$ is flowing to Black pastors for this purpose. Fair or unfair I have no doubt the events in Ferguson, MO will be utilized to ‘rally’ AA voters to the polls by the opposition. Remember in 2012 the dem pollsters (PPP) and others kept sharing polls w/them winning the election and we thought they were bogus polls. BUT they turned out those voters that our pollsters didn’t have showing up and their pollsters did. The point being is that this is their game plan and we need to stay diligent on our game plan to turnout our own low-propensity voters as well. We need to turnout lo-propensity Republican voters to the poll.
In Georgia we have identified approximately 594,000 such low propensity Republican targets. The focus of our staff on the ground is primarily talking to THOSE voters. We need to turn them out in order to offset whatever the other side is capable of doing to turnout their lo-propensity voters. They will try to to bump the AA turnout from 25% to 29%. Its hard but not impossible and we need to plan as if they will because they did it in 2012.
African American turnout and election results below:
2006: 16% (Pre-Obama) – Perdue +19%
2008: 30% (Obama 1) – McCain +5%
2010: 25% (Obama-midterm 1) – Deal +10%
2012: 30% (Obama 2) – Romney +8%
2014: 25% (Obama-midterm 2) – Deal +???
Last night, Landmark sent an email discussing the poll:
The difference between our poll and the others is due to demographic weighting: we believe that the model for our poll more accurately matches the demographics of a Georgia general election.
We believe that other firms are aware of the issue and that they will ultimately modify their weighting to account for it.
WHAT HAPPENS IN A GEORGIA GENERAL ELECTION TODAY IF ONE RELIES ON A MODEL BASED UPON 2010 TURNOUT PERCENTAGES?
In 2010 the percentage of the general election voters that was African-American was approximately 28.5% (though it must be taken into context that voters who haven’t provided their ethnicity on voter registration applications are also disproportionately minority voters, meaning that the ‘real’ minority vote would have been about 30%).
In fact, if the 2010 election were held today using today’s voter registration numbers, and if the demographic turnout for white, black, and non-black minority voters were the same percentages as the 2010 election, then about 30% of voters would be African American: about a 1.5% higher than 2010’s general election black turnout.
Many other polling firms have projected a 26% “black vote” turnout in November.
…But the percentage of the black vote in Georgia hasn’t been 26% of the electorate since 2006—more than three elections ago. We do not agree with this model.
The Landmark poll projects a 29% ‘black vote.’ Frankly, anywhere from 28% to 30% would be reasonable, but we believe 26% is too low and has caused other polls to incorrectly display higher support for Republican candidates at this stage of the election.
So, there seems to be a difference in opinion on the historic figures for African-American turnout as a percentage of all voters. The difference between the RNC’s 2010 figure or 25% and Landmark’s 2010 figure of 28.5% is very significant in this setting. We’ll do some research on this issue today and present our findings tomorrow morning.
In October 2012, I wrote about how weighting works and when and under what conditions it should be applied:
Weighting is a statistical process used to compensate for differences between the makeup of a survey sample and known [or predicted] qualities of the electorate.
For example, if our survey yields 65% female respondents and 35% male respondents, while historical turnout figures peg women as around 53-55% of most iterations of the electorate, we know that our sample is gender-biased. We might weight each male respondent by a factor of about 1.28 so that after weighting, males account for about 45% of the weighted sample. We would likewise weight female respondents by a factor of about .846 so that they represent about 55% of the electorate.
My belief and general practice is that it is appropriate to weight survey results when they are out of line with known historical facts and trends. The example above involving gender would clearly qualify.
I think that Landmark’s prediction that African-Americans will constitute roughly 28-30% of the electorate in November 2014 is sensible and weighting to that target is methodologically defensible.
If you’re interested in seeing exactly how weighting works, in August 2013, I took a survey I performed and weighted it in line with another pollster’s preferences to show how weighting moves poll results.
We found in our unweighted results that if Michelle Nunn is the Democratic nominee, she starts today in the range of 31.55 to 34.70 percentage points and the major GOP nominees range from 50.47 to 54.26.
This contrasts strikingly with PPP results showing Nunn in the range of 38% to 41% and GOP contenders in the range of 36% to 41%. I said I think my results are closer to correct, and pointed to PPP’s Party Identification numbers showing self-identified Democrats at 38% to Republicans at 39% as suspect. I believe that they weighted their results to arrive a near-parity between the parties.
So I weighted my results to bring them much closer to what PPP believes is the accurate Party mix and here’s what happens.
After weighting my results to bring them into line with PPP’s party identification numbers, Michelle Nunn’s strength rises about eight points and now ranges from 39% to 42%, very close to PPP’s 38% to 41%. In the weighted scenario, GOP contenders are now in the 44% to 48% range, closer to PPP’s 36% to 41%.
Another issue people like to discuss is the role of IVR or “robo-polls” versus live agent polls, which are considered the “gold standard” by many both within politics and without. A 2011 article in DailyKos discussed the differences and attitudes toward robopolls.
Years ago, critics laughed at the notion that “robo-pollers” could generate good numbers. But that line of attack is finished. SurveyUSA and PPP were ranked the second and fourth most accurate pollsters in 2010 by Nate Silver – above such traditional stalwarts as Mason-Dixon and CNN/Opinion Research. And that accuracy was particularly impressive given that those two IVR pollsters focused extensively on harder-to-poll down-ballot races, while traditional firms stuck with statewide contests like senate and governor.
I asked Jon Cohen, the Washington Post‘s polling director, why he wouldn’t let his writers run IVR polling. He responded via email:
“Our editorial judgments are based on how polls are conducted, not on their results, or apparent accuracy. Now, we flag polls that have really bad track records, but end-of-campaign precision is a necessary, not sufficient condition in our assessments,” said Cohen.
“On the methods front, the exclusion of cellphones is a big – and growing – cause for skepticism about IVRs,” said Cohen.
I asked polling guru Mark Blumenthal of Huffington Post Pollster what he thought about the Washington Post‘s refusal to publish IVR results.
Blumenthal noted several potential problems with IVR polling — that excluding cell phones could provide a more Republican sample since 30 percent of households don’t have landlines (a problem also shared with most traditional horserace polling, since very few of those polls include cell phones). He also noted that the inability to target people specifically in a household makes weighing more important to IVR pollsters (though the much larger sample sizes possible mitigate his problem).
“But that said, short (2-5 minute) public polls that we have examined over the last 4-5 election cycles have generally produced horse-race vote estimates that have been as accurate in comparison to election outcomes — in the final weeks of the campaign — as those obtained using live interviewers,” said Blumenthal.
“I believe those accuracy measures argue against outright bans on reporting of such polls in the final weeks of the campaign,” said Blumenthal.
Here’s how I explain the difference between live agent polling and robopolling. It’s like the difference between a Rolex and a Timex. A Rolex is vastly more expensive than a Timex, though it performs the same basic functions. But it looks better, feels better on the wrist, and provides some bragging rights. A Timex is less expensive but omits braggadoccio, though it’s timekeeping is relentlessly accurate, and your average quartz watch, whether it’s a $2 bargain bin cheapie, or a $2000 plus Omega Seamaster, keeps time more reliably than any Rolex that hasn’t just been factory-calibrated.
But the differences in cost, construction or timekeeping don’t tell you whether any given watch is accurate. As my friend, former Congressman Buddy Darden said last week on Political Rewind, “a broken watch is right twice a day,” and a working watch improperly set will always be wrong.