Unofficial Vince Gray for Mayor Blog

Compiled by Joshua Bell of the Unofficial Vince Gray for Mayor website

August 24, 2010

Note:  Because of the length of this blog entry and the formatting challenges of the "statistical analysis section,” I had to break up this blog entry into 2 parts.  The first part you will see below in narrative form.   However, at the end of the posting I have embedded the “full document” with the full "appendix" with all the detailed projection numbers (by ward) for your review.  This embed should also be printable.  You can also find it by clicking here.

Please email me if you have problems accessing the full document.  joshuabelldc@earthlink.net

OVERVIEW

Inspired by a combination of personal curiosity, a question from @EricaVnDC on twitter, way too much free time, and an interesting piece by Mike DeBonis of the Washington Post on speculation surrounding turnout and the “racial mix” of the potential electorate for this year’s primary, I decided to try my hand at a little amateur election prognostication based on information available on the internet and my own supposition.

What follows is my best attempt at a fair and reasoned quantitative and subjective analysis of the upcoming primary results given my limited resources and expertise.  In order to reach my conclusions and my eventual set of “projections” I did a good bit of research, compiling publicly available information from a wide range of internet sources.  Including, but not limited to: information available at the DC Board of Elections and Ethics (DCBOEE) site, polling, surveys, and press reports posted on the internet.

So, after a good bit of research, making some base assumptions, and after finally settling on a rather simple, rudimentary methodology, I crunched the numbers in my excel program and th results I came up with all indicate that unless there is a major shift in current voter sentiment, Vincent Gray is on the road to a modest, but convincing victory on September 14, 2010. 

Under every scenario observed, each “case” indicates that VINCENT GRAY is on a comfortable path to victory.  Barring any significant changes in voter preference, a major gaff, or a game-changing scandal hitting one side or the other, the quantitative analysis below supports the same conclusion that this blogger has observed from a purely anecdotal point of view for months now:   All signs both quantitative and qualitative lead to the conclusion that VINCENT GRAY will likely be the NEXT MAYOR OF THE DISTRICT OF COLUMBIA.

Of the 6 scenarios I considered, Vince Gray’s margin of victory ranged from a high of +9.8% (54.2% to 44.4%) to a low of +4.8% (50.8% to 46.0%).  In every scenario, Gray was in the low to mid 50’s and Fenty was mired in the mid-40’s.

Note:  Though the details of my calculations and methodology are “proprietary”, a careful review of my projections will reveal them to be sound and wholly plausible.  Also, I would be happy to entertain any questions or challenges to the results herein and would be happy to “plug in” any alternative assumptions into my “models” in order to investigate potential alternative outcomes.  You can email me at: joshuabelldc@earthlink.net

PROJECTION MODELS:

For my projections I used 2 basic “models,” and within each model I plugged in numbers based on 3 different “voter turnout” estimates (low turnout case, mid-range turnout case, and a high turnout case).  This provided me with at total of 6 possible outcomes/scenarios.

The first model I created (MODEL A) was based on the recently released Clarus poll/Voter Survey .  I used the “preference” report from this survey as a baseline for how voters in each ward are likely to vote.  The second model (MODEL B) was based on assumptions about how each ward might vote based on consistent “voting patterns” by ward observed over the last 3 primary cycles, with the greater weight given to the primary in 2006.

METHODOLOGY/ASSUMPTIONS:

The most important factors/variables considered while generating the projection models:

1)  CURRENT VOTER REGISTRATION.   Current Democratic Party Registered Voting Statistics (by ward) as of “monthly reporting period” ending 7.31.10.  This is public record that can be found at the DC Board of Elections and Ethics (DCBOEE) site, here.

2)  HISTORIC VOTER REGISTRATION.  Historic Democratic Party Registered Voting statistics (by ward) for each “primary year” beginning in 1994. Again, I found those numbers at the DCBOEE site within the “election results” section of the website.

3)  HISTORIC DEMOCRATIC PARTY PRIMARY RESULTS.  Results of the various Democratic primaries since 1994, with particular attention paid to ward-by-ward results.  Whenever possible pains were taken to not make any “citywide” assumptions as the voting characteristics of each ward (be they results, voter turnout, or number of registered voters) vary greatly by ward.  Again, I got this information from the DCBOEE site.

4)  VOTER TURNOUT.  For the purpose of this analysis I created 3 assumptions for “voter turnout”.  The LOW CASE was computed by calculating the “average” voter turnout over the last 3 primaries.  “THE HIGH CASE” was calculated by taking the voter turnout numbers from the primary election of 1994.  Though this was a record level of turnout that might at first seem an outlier that should be dismissed, after much consideration I decided that the turnout levels for 1994 should be considered at least for the “high case” because I came to the conclusion that the  last time there seemed to be this level of “interest” in a District primary was the Democratic primary of 1994, featuring a very controversial and “polarizing” candidate (like this year’s Adrian Fenty): the “once and future” mayor and “bad boy” Marion Barry. 

In my opinion, there are elements to the 2010 election, particularly the polarizing presence of Adrian Fenty and the passion of the supporters and detractors that could result in a voter turnout level that approaches or exceeds 1994.  Additionally, the unusually high “turnout” at the non-binding Democratic Party forums/straw polls  around the city also indicates higher than normal “interest” in this race, further buttressing the contention that a record or near-record turnout should be taken into consideration.  The MID-CASE was calculated by averaging the difference between the “HIGH CASE” and the “LOW CASE”.

5)  THE CLARUS POLL/VOTER SURVEY (AUGUST 2010).  I relied on the results from the recently released Clarus Voter Survey for the baseline “voter preference trends” for “MODEL A”.  I selected this particular survey because it was the only “publicly released” poll that I could find that provided preference numbers by section of the city, breaking down their numbers into groups: Wards 1, 4, and 5; Wards 2 and 3; and Wards 6, 7, and 8.  I found this breakdown most useful as it allowed the analysis to take into account and use the “variables” of the historic voter turnout history of each ward when formulating a projection and an analysis that seemed more nimble, dynamic, and accurate.

6) UNDECIDED VOTERS.  For the purposes of this analysis, I decided to assume that the current undecided voters (as identified in the Clarus Survey) will likely fall into the category of non-votes, opt-outs, or they will basically break in similar proportion to the “declared” voting preferences for the “wards” in which they reside.   At one point my inclination was to “skew” the break of undecideds toward Gray because of Fenty’s unusually high “unfavorable numbers” in the Clarus Survey, and the fact that historically, those voters that are still “undecided” this late in the game after having had a chance to observe the incumbent for 4 years, likely means that these voters will either not vote or lean toward the “challenger”.  That said, in the interest of erring on the side of an assumption that would not “favor” the candidate that I endorse, I decided to go with the assumption that the undecideds will break proportionally.

7)  RACIAL MAKE UP OF THE ELECTORATE.  Though much has been made of the subject of a “racial divide” characterizing the upcoming primary, it is very hard to get good information from exit polling or other sources that really provide a reliable enough “history” of the actual “racial” make-up of the electorate by ward.  Therefore, I have decided to basically dismiss a direct account for “race” in my model formulation for the purposes of this analysis.  

I came to the conclusion that if a detailed enough analysis were done of each individual ward and the history of turnout and numbers of registered voters for each ward since 1994 that any “racial” dynamic would be reflected implicitly.  Simply put, everyone knows that Wards 3 and 2 are generally considered predominantly “white” wards (and since the 1994 primary have had an average Democratic affiliated “turnout” of 42.9%), and Wards 7 and 8 are considered predominately “African American” (with historic turnout levels since 1994 averaging 33.3%), so my feeling is that the careful analysis of the “voting patterns” of each individual ward would essentially capture any “racial” component of the electorate at large. 

Note:  An interesting “wild card” in this election will be the results from Ward 4, which is an increasingly racially “mixed” (and relatively affluent) ward with one of the highest historic turnout rates, 44.8% (ranking second only to ward 3’s 49.6% number since 1994).  This ward, once a Fenty stronghold, will be interesting to watch as the African American voters in this ward have shown that they are less than thrilled with Adrian Fenty as Vince Gray’s stunning straw poll win in the Ward 4 straw poll (and the nearly 1000 voters that attended that forum) attests to.  That said, if one were to rely on the recent Clarus Survey, it is likely that Ward 4, despite its large projected voting population (ward 4 ranks as the ward with the most registered voters in the District at 47,820), will likely split essentially down the middle, making the real the voting margins (and turnout levels) of Wards 2 and 3 vs. Wards 5, 6, 7, and 8 likely being the determining factor in this primary.

THE SUMMARY OF RESULTS

When all was said and done and my excel program finished crunching, I came to the conclusion that there is no credible or plausible path for Fenty to get over 48% of the vote.  I just don’t see it.  And conversely, I couldn’t find a scenario where Gray fall below 50%.  The critical and “terminal” factors for Fenty is the “east of the Anacostia” factor and his extremely high “unfavorable” numbers.    

Despite all of the snide dismissals from the Fenty camp that the the “east of the Anacostia” folks and the “rabble” that stormed the Ward 4 straw poll are just “angry blacks” that won’t end up turning out to vote, I believe that all evidence points to the contrary.  It is for the very reason that they are “angry” that they are more motivated than they might have otherwise been in a “normal” primary cycle.  If for no other reason, they will turn out just to teach Fenty a lesson for his dismissal of them  as people/constituents, and a perceived lack of attention to the issues that affect them. 

Simply, if the Gray team’s organization and “ground game” is as good as they have shown heretofore (re: the straw polls and forum attendance this summer), Team Gray will likely be able to pull enough support from Wards 7 and 8 (and 5) to make Fenty’s path to re-election nearly insurmountable.

THE PROJECTION RESULTS:

Case 1:   Clarus Model Assuming 35.2% citywide voter turnout

Gray -   59,035  (50.8%)

Fenty -  53,475 (46.0%)

Gray margin of Victory:  +4.8%

Case 2:   Clarus Model, assuming 42.6% citywide voter turnout

Gray -   72,336  (51.6%)

Fenty -  63,619 (45.4%)

Gray margin of Victory:  +6.2%

Case 3:   Clarus Model, assuming 50.1% citywide voter turnout

Gray -   85,637  (52.1%)

Fenty -  73,763 (44.8%)

Gray margin of Victory:  +7.2%

Case 4:   Historical Voting Patterns Model, assuming 35.2% citywide voter turnout

Gray -   60,922  (52.6%)

Fenty -  53,480 (46.2%)

Gray margin of Victory:  +6.4%

Case 5:   Historical Voting Patterns Model, assuming 42.6% citywide voter turnout

Gray -   75,043  (53.5%)

Fenty -  63,271 (45.1%)

Gray margin of Victory:  +8.4%

Case 6:   Historical Voting Patterns Model, assuming 50.1% citywide voter turnout

Gray -   89,164  (54.2%)

Fenty -  73,063 (44.4%)

Gray margin of Victory:  +9.8%

QUALIFICATIONS/CONSIDERATIONS:

In light of the fact that this is a blog which openly endorses the candidacy of Vince Gray, the casual or skeptical reader might look at my results askance.  In response to that I would say that despite the figures below, I took every effort to be as impartial as possible.

That said, do I believe that  there are possible scenarios that would lead to a Fenty victory?  Sure, but my analysis indicates that unless the basic public opinion (and level of Fenty negatives) change substantially in the 3 weeks before the election, every “probable” scenario leads to a Gray victory.  The only chance for a Fenty win lies in either unexpectedly low voter turnout east of the Anacostia, or a sea change in the current polling patterns.  Both are possible, but as for the former, my opinion is that despite historically low voter turnout in Wards 7 and 8, even if the turnout equals the average of the last 3 primaries (2006,2002, and 1998), Gray likely wins.  However, my feeling is that there is more interest in this race because of the level of passion for and against both candidates that this election sets up to be at or near the record turnout of 1994 when Marion Barry beat John Ray and Sharon Pratt Kelly in the Democratic primary.

As the saying goes, endeavors like this are as much “art” (or in this case my case “wishful thinking”) as it is “science”.  But that conceded, I feel that I have given this as fair and as comprehensive a try as I am capable given the resources at my disposal and conceding that I am a decided “non-pro” when it comes to professional level political polling and modeling.  What I do have in my favor, I believe, is decent basic instincts and a passion for the subject.  That said, take this what follows for what it’s worth.

FINAL THOUGHTS:

As mentioned before I welcome questions and challenges to my assumptions and results.  Also, my model is created in a way that I can plug in new numbers and adjust the projection when and if new information or new polling becomes available.  That said, “citywide” polling information will not be as useful as polling information that is specific to each ward. 

Feel free disseminating this information and analysis so long as you give this author full credit. 

THE NUMBERS:

PLEASE FIND THE DETAILED NUMBERS AND QUANTITATIVE ANALYSIS SECTION (AS WELL AS THE COMPLETE PRINTABLE DOCUMENT) AT THE END OF THE DOCUMENT EMBEDDED BELOW:

Vincent Gray Path to Victory the NUMBERS Section 8.24.10

Vincent Gray Path to Victory Result Projection 8.24.10

*************************************************

Note:  For a list of previous posts to the Unofficial Vince Gray for Mayor Blog click here.