Sunday, August 31, 2008

Life is What Happens on The Way to What You Had Planned

Aunt Veronica Was Right

She used to say the “Road to hell is paved with good intentions”.
I intended to compare Delaware state & presidential primary turnout for 2000, 2004 & 2008 to 8 other comparable states, but when I went to their websites they were not as user friendly as Delaware’s Election Commissioner’s website.
California broke down by county,but not by statewide totals for turnout. California has dozens of counties which would have required a lot more number crunching time than I have at the present time since the intention was to post the comparison before the Sept 9,2008 primary. Connecticut (unlike Delaware) did not post a statewide turnout unless there was a statewide race. This means to figure out statewide turnout a person would need to add up turnout from all of the local races throughout the state and then divide statewide registration by that number for each of the three election cycles.
I still plan another post before the Sept 9, 2008 primary, but it will be less ambitious due to the limitations of the time remaining before the primary.

Sunday, August 24, 2008

For the Presidential Primary/State Primary Calendar Few Compare with Delaware

Obama-Clinton Spillover Effect on 2008State Primaries

I have heard several people say that the interest in the Obama-Clinton Presidential primary in Feb 2008 with its increased voter turnout will have spillover effect on the Delaware state Democratic primary in September and increase turnout in that contest. I wanted to quantify that by looking at other states that have the statewide primary after the presidential primary . I looked at delegate selection processes throughout the U.S . and found a lack of uniformity which makes it tough to compare Delaware’s experience relative to other states,but I plan an update within the next week or two for the few contests that appear comparable.

Comparison states with their State primary date and 2008 Presidential Primary date:

California 6-3-2008(state) (1-29-2008presidential)
Connecticut 8-12-2008(state)(2-5-2008presidential)
Missouri 8-5-2008(state)(2-5-2008presidential)
New Jersey 6-3-2008(state)(2-5-2008presidential)
South Carolina 6-10-2008(state)(1-26-2008 presidential)
Tennessee 8-7-2008(state)(2-5-2008 presidential)
Utah 6-24-2008(state)(2-5-2008presidential)
Virginia 6-10-2008(state)(2-12-2008presidential)

The Apples to Oranges Ground Rules for the Obama-Clinton Spillover Effect in 2008 State Primaries

The only contests that will be considered will be contests that are as equivalent to the Delaware Democratic State Primary as possible:
(1) the State primary is held later than the Presidential Primary,but prior to 9-1-2008 so results can be reviewed prior to the Delaware primary
(2) States with a single primary with no runoff election following the primary
(3) only states will be considered (leaving out U.S. territories, Washington,DC, Democrats Abroad, etc)
(4) Only states with DNC approved primaries will be discussed.
(5) Not caucus states

List of contests omitted:

New Mexico
North Dakota

Nonstate contests
American Samoa
Democrats Abroad
District of Columbia
Puerto Rico
Virgin Island

States with state primaries held after September 1, 2008

Arizona 9-2-2008
Massachusetts 9-16-2008
New Hampshire 9-9-2008
New York 9-9-2008
Rhode Island 9-9-2008
Vermont 9-9-2008
Wisconsin 9-9-2008

States with primaries that include runoff elections after the initial primary vote

North Carolina

States that held their primary the same day as their presidential primary

South Dakota
West Virginia

Primaries not sanctioned by the Democratic National Committee due to rules violations


Information on each state was from the elections web page of the state’s web site.
You can access each state’s web site by typing in with the -- being the two letters representing the state (for example, Delaware is
Information about the presidential primary calendar I got from the New York Times at:

Monday, August 11, 2008

I Get By With A Little Help From Other People's Friends

At Large Strategic Voting in Wilmington City Council Primaries

In the upcoming Democratic Primary for the Wilmington City Council At-Large seats, being second favorite might be best. There are four at-large seats on City Council with no more than three at-large members allowed to be from any one party. For decades this has meant 3 Democrats and one Republican due to the more than 4-1 Democratic registration edge (32,343 Democrats to 6947 Republicans) in the city. The Republicans have only two candidates for three ballot slots on the November ballot,so they don’t have a primary.

The Democrats have two incumbents seeking re-election, Charles “Bud” Freel and Loretta Walsh.The third current At-Large council member, Theo Gregory, is running for City Council President. Facing the two incumbents are 6 challengers in the Democratic Primary: Darius Brown, Dwight Davis, Samuel Guy, Gary Hutt,Vincent White and Justen Wright.

In the primary each voter has the option of voting for as many as three candidates,but some people only vote for one or two. In an 8 way race, being the second choice for the supporters of other rivals may be a path to victory. The numbers cited below indicate that the average votes cast per voter in the last 5 city wide primary was between 1.91 and 2.21.

This means that while some voters vote may vote only once ,many voters make two or three choices .

1988 -10,236 voted in mayor’s race , 20,043 voted in the city council at-large,
votes cast per voter in the at-large race 1.96

1992-10,773 voted in mayor’s race, 23,607 voted in the city council at-large,
votes cast per voter in the at-large race 2.19

1996-10,549 voted in mayor’s race, 23,329 voted in the city council at-large,
votes cast per voter in the at-large race 2.21

2000-9353 voted in the mayor’s race , 17,835 voted in the city council at-large,
votes cast per voter in the at-large race 1.91

2004-total Wilmington voter turnout 7,558 (no mayor’s primary),15,915 voted in city council at-large,
votes cast per voter in the at-large race 2.11

The case can be made that the winner could make it with “one choice only voters’ and that is mathematically possible,but not likely. While we can not determine how many people actually vote for only one candidate, we can make some calculations by process of elimination which show that winners tend to garner support from supporters of their rivals.

In 1988 Loretta Walsh and Theo Gregory were the top two finishers in a field of five and got a combined 10,040 votes. There were 10,236 votes cast in the mayoral primary that year,so if Walsh and Gregory both had only single choice voters that would leave only 236 voters times 3 votes per person to be shared by the other three candidates. However, 3 x 236= 708. 708 plus 10,040=17048 potential votes cast. Since 20, 043 votes were actually cast in the at-large race, we can be fairly certain that Walsh and Gregory’s combined total included many second choice votes.

1992,2000 and 2004 present even clearer evidence.
1992 Walsh and Gregory got a combined 10,796 votes in a six way race,but only 10 ,773 votes were cast in the mayor’s race. Unless 23 people voted in the at-large race and skipped the mayor’s race it is guaranteed they got second choice votes.
In 2000 Theo Gregory and Bud Freel were the top two vote getters in a four way race with a combined 9801 votes. Since only 9353 votes were cast in the mayor’s primary, unless over 450 voters skipped the mayor’s race and voted in the at-large race they also mathematically had to get some second choice voters.
In 2004 there was no mayor’s race,but 7,559 city Democrats voted. 7588 of them voted for one of the two top vote getters in a 5 way race, Theo Gregory and Loretta Walsh, indicating they again got second choice votes since the most possible single choice votes they could have gotten would have been 7,559.

Because in 1996 there was 14 way primary for 3 ballot slots after two opened up when Walsh and fellow city council member, Bob Poppiti, both ran for mayor, this type of math is not as easy to establish. Gregory and Freel were the top two vote getters, but garnered a combined 6, 775 votes. 10, 549 votes were cast in the mayoral primary. While it is likely they each got their share of second choice votes, it is not so easy to quantify as when there is a smaller ballot. 10,549 minus 6,775= 3774 which is how many voters would be left over if all of Gregory & Freel’s votes were from single choice voters. 3,774 x 3 votes per person makes 11,322. 11,322 plus 6,775= 18,097. Since 23,329 votes were cast in the at-large race, we can tell that Freel and Gregory must have gotten some second choice votes because 23,329 minus 6,775=16,554. If all of their votes were single choice only votes there would not be 18,097 voters left to cast the votes that were actually cast.

Sometimes the best way to predict the future is to look at the past. While the position expressed in this essay may have been discussed previously I had never seen it quantified. I had heard that candidates ran on the plan of getting their own supporters to vote for them only and it appears that might not be enough.
There are 8 candidates in this year’s at large field, so the math may be closer to the 1996 scenario than the other four elections discussed,but I am betting that whoever wins is going to rack up their share of second choice support from voters that other candidates convince to go to the voting booth.

Sunday, August 3, 2008

Some Thoughts On Relative Turnout

I don’t know who will win the New Castle Countywide Democratic primary in September 2008,but I have an idea where the votes will come from.

Some areas with Democratic registration leads in Wilmington, New Castle and Elsmere that have had local primaries in recent years do not have as many as usual or have none. This may dampen down turnout in these Democratic strongholds. They may still have larger raw numbers of votes. But in a countywide primary their relative impact may be more limited. Some swing districts that are not used to have primaries have a state senate primary in the 6th District that may generate a larger than usual turnout in the Newark-Hockessin area. In Brandywine Hundred there is also a primary in the 4th State Senate District, an area that has been represented by a Republican for many years.The number of actual Democratic voters may be less in these districts than in the more heavily Democratic districts since to have a small percentage of a bigger pie can be larger than a large portion of a much smaller pie mathematically, but the relative impact may be more.

As Bill Clinton might say “The issue is one of proportionality”. The 1st District in Northwest Wilmington has a 9363-1504 . Democrat-Republican voter margin, The 2nd has a 8477-1263 D-R margin. The 16th in the Wilmington/New Castle area has a 8588-1567 D-R margin. The 13th in the Elsmere area has 6798 Dems to 2467 Republicans . The 17th in New Castle has a 7210 to 2924 D-R margin. In theory all of these districts are powerhouses with at least 2-1 Dem-Rep registration rates. I think the only two likely to make a relatively dramatic impact are the 1st and 2nd Rep Districts , due to local primaries giving voters an additional local reason to vote.

While the city of Wilmington is having primaries for mayor ,president of city council, district city council seats and an at large race for 3 of 4 at large seats on council, none of the incumbent state legislators or county council members representing Wilmington has a primary. In the 13th , 16th & 17th there is also no incumbent state legislator or county council person with a primary. The 13th & 16th each has one election district (the basic voting unit of the polling place,sometimes called precincts in other areas) in the city of Wilmington,but not enough to impact relative turnout. In the 3rd ED of the 13thRD only 63 voters of a possible 249 cast votes in 2004 election. In the 6th of the 16th 194 of 776 voters cast ballots in 2004. During 2004 the 16th & 13 th each had a local State rep race and a local county council race, neither of which will take place in 2008.

The 23 RD in Newark has 5384 Democrats. The 22nd RD has 5060 Democrats . The 22nd & 23RDs comprise 13 of the 16 election districts within the 6th state senate district which is having a Democratic primary for the first time in at least decades. In 2004 these two districts had combined voter total of 1656 and 6.4% of the total votes in the county council president’s race. I assume with the additional state senate race, the total turnout will increase,but also the turnout relative to the districts without local primaries.

My brother asked “And how does knowing this help anyone?”. If a candidate has unlimited resources and can campaign effortlessly countywide, this research is unnecessary. If you have limited resources and have a rough guess at where your effort can either generate new voters or help you persuade likely voters to actually vote for you, I think this process has value. It may also help in assessing voter saturation points above which it becomes unrealistic to expect additional votes in a given location. If a candidate can isolate permanent voters (people who vote no matter what is on the ballot – I think those who have repeatedly voted in the primary for the US Representative District when there is no local primary would qualify) from the more casual voters, different messages could be structured to each of these groups. The permanent voters may need less urgency implied and more persuasion. The casual voters may need a sense of urgency to motivate them to even participate.

One area that I think that will potentially produce a high relative turnout is the 1st Representative District along the Baynard Blvd-Washington St corridor in Wilmington. While it usually ranks highest among Representative Districts in actual numbers, it’s relative turnout was lower percentage wise in 2004 than compared to 2002 to 2006. I think this is because the local 1st State Senate seat faced a Democratic primary both 2002 and 2006. One of the things that reduced the relative impact of the turnout in the 1st Representative District in 2004 was that many areas that usually do not have primaries on a regular basis years had primaries. These included the 17th Rep District and 5 of the 6 newly created open seats on County Council. Four years later only two of the 6 county council seats up for election have primaries.

In addition to it’s tradition of producing a lot of votes, the voters in the 1st Representative District have at least six local reasons to vote within the 2nd City Council District in the Baynard Boulevard area . Like other Wilmington residents they have the mayor’s race and president of city council race,but unlike other Wilmington residents they know one of their neighbors will win in the president of City Council contest. Long-time city council members ,Theo Gregory and Norm Griffith, both live in the 2nd City Council District in the First State Representative District. In addition, the race for the local council seat Griffiths is vacating has a four-way race, with one of the candidates being Griffiths ‘ wife, Michelle Griffiths, a former police officer. The other candidates are local Democratic activist Larry Smith, Devon Hynson, who ran unaffiliated in the general election against Norm Griffiths 4 years ago, and Ernest Congo, II, a local funeral director and regular host on Channel 28, the local cable access channel. All six of these candidates have been visible in the community for a number of years ,so name recognition may not be the stumbling block it sometimes is for local candidates.

In the PS Dupont School-Miller Road area in the northern section of the 1st Representative District, Charles Potter, Jr, also a regular on Channel 28 on Sundays, is seeking re-election and faces a primary from another Channel 28 host, Nnamdi Chukwuocha , assistant director of the Kingswood Center and one of the Twin Poets.
These city council primaries should combine to keep the 1st Representative District as the leader in votes produced and highest relative turnout.

Relative Turnout guess for now:
1st Representative District will produce 10-12% of the countywide vote.
The 13th Representative District will produce 3.5% of the countywide turnout.
The 16th Representative District will produce 6% of the countywide turnout.
The 17th Representative District will produce 4% of the countywide turnout.

The 22nd & 23rd Representative Districts will produce 7-9% of the countywide turnout
If these predictions pan out or if the 22nd & 23rd produce a higher relative turnout and possibly more actual voters than the 16th & 17th Rep Districts, it would be notable since the 16th & 17th have over 5000 more registered Democrats than the 22nd & 23rd.