In the Begetting

With election day tomorrow, a common question right now pertains to my projection of turnout.

I'm very good at nailing the turnout number the day after the election.

Actually, before the election, fine-tuning each day based on advance voting, I'm generally not that far off with my estimate the night before the election.

Many variables, though, come in to play.  It's trendy right now to discuss predictive modeling techniques to forecast turnout, number of machines needed, and the right potion to ensure that lines at the polls are reasonable.

In fact, I want our staff to become better at logically predicting turnout, because forecasting volume, to me, is the precursor to everything election administer-ey.

I'm thinking the turnout for the August 5 election will be somewhere between 17 and 20 percent, closer to 17 than 20.

(I hate talking about turnout, generally, though because I'm always worried that a public turnout estimate will impact voting behavior.  Some may come or stay at home because they worry that the polls will be busy; others may not want to go to a party that doesn't have anyone there and, counter-intuitively, may not come to vote if they think no one else is.)

(If you haven't picked up this by now by reading this blog or this post you may never pick it up, but coming up with Backup Plans A, B, C, D, E, F, G, Gg, H, H1, H2, Hn, I, etc., induces a constant neurosis and tendency to over-think the most simplistic items).

Anyway, back to the 17 percent.  378,000 voters times .17, divided by 183 polling locations, and voila, there you go, forecasting, right?

No, especially in this election when Republicans cast a Republican ballot, Democrats cast a Democratic ballot, Libertarians cast an Unaffiliated ballot, and Unaffiliated voters can cast an Unaffiliated, or Republican, or Democratic ballot.

Further, while every polling place will feature some competitive countywide races, not every House race, for instance, features two candidates.  Turnout likely will be lower at those locations.  Some polling places will have a combination of precincts with super competitive races and precincts with less competition.

So, toss all that in with your Soup Stone, and what's the turnout again?

Turn to history, some may say.  The key to the future is found in the past.

But, what past?  2010, the last time there was a county commission chair primary?  Turnout in that election was 22.9 percent.  There was an extremely competitive primary for an open U.S. Senate seat that year, with both campaigns pushing advance voting extensively.

About 5,000 voters in that election didn't cast a vote for the county commission chair race, but the "turnout" in that race still was more than 21 percent.

Hmmmm, how about 2012?  Turnout that year was 17.33 percent with no U.S. Senate race and no county commission chair race.

2014 will be higher, right?

Well, look at these numbers through the weekend:

Monday Morning Before Election in:                          2014            2012              2010
Ballots Issued                                                               24,738         26,285           33,118
Ballots Returned
(includes in-person and by mail)                                18,150        18,245           23,805

In-Person Advance (included in numbers above)        9,163           9,254            12,337

So, 2014 is tracking just 95 fewer ballots than in 2012.

You'll see that it's pretty common in an August election to have many ballots mailed out that are not returned.  Many come back undeliverable, not just from the permanent list but also from people who recently requested a ballot.

I've said it before and I'll say it again--people move.

Although we don't have updated numbers through today yet for ballots returned (2012's number jumped to 20,767), the in-person advance numbers from today close our total out to be 9,512--compared to 9,486 in 2012.

So, what does that mean?  Are we surging?  33 percent more voters cast ballots in person today than in the last day of 2012.

The new in-person gain puts 2014 now 22 ballots ahead of 2012.

2012, though, had tons of energy.  Redistricting led to a later filing season and plenty of rookie candidates that made nearly all races competitive.

2012, really, is nothing like 2014.  But, the numbers are nearly identical.

Dismissed as coincidence?

I'm just getting started, kids, but in the interest of time for the three people who have made it this far, I'll stop the post here.  The point is, simply, that all of these variables have to be considered when looking at turnout.

And turnout begets staffing, which begets number of machines, which begets number of locations, which begets the number of printed ballots, which begets the need for advance voting locations.

I'm not exactly sure that I have the "begets" in the correct order, but I am correct that one number begets another, and the best number I can give for August 5 turnout is 18, with homage to today's surge.

We'll see in a week.

What?  The election is tomorrow, you say.

Yes, but tomorrow's turnout percent doesn't include provisional ballots.  Those counted are added to the turnout at the canvass on Monday.  Those not counted aren't in the turnout percentages at all.

And, so, another variable was just beget.

(A new one, by the way, is emerging in terms of facility availability, shown below).



Suitcases staged for supply delivery last week, but....

More and more facilities don't want
deliveries until Monday, a whole
new problem with polling places.  Our
setup teams each delivered to a location
today as well as the delivery company.

It's Not Rocket Science

Election administration may not be rocket science, but it does have one thing in common with the airline industry.

Specifically, that commonality exists with airline security.  In both cases, it seems like we're always focused on the last problem.

We take our shoes off and check our liquids at airports because of incidents that happened.

We train election workers and create processes for election day to address one-offs from the past.

Voters have concerns about voting machines?  Create a streamlined way to allow them to request a paper ballot at the polls.

Voters realize that touch-screen technology in place since the late 1990s can be temperamental?  Create a sign at the registration table to remind them that they aren't voting on modern Android or Apple devices and to review selections before hitting cast ballot.

Election worker fat-fingered the selection of a specific ballot on an encoder?  Send out a finger stylus that likely won't be used but try to create a simple 1-2-3 mental picture of how to check to make sure the proper ballot is encoded.

And so it goes.  A good half of our training time covers things that likely will never happen, but with 200 polling places and 1,000 workers, they will happen to one of them.

Much of the training is, "1 out of every 100 times," "2 out 10," "If you draw the short straw and this happens to you," etc.

April elections are the mostly likely ones where election workers might encounter a locked door when arriving at 6 a.m.  This happens once or twice in an election, and happened once this past week.

That's just a 5 percent of our polling places if you are scoring at home, affecting 0.5 percent of all of our workers.  If we knew where that would happen, training would be more efficient.  Without knowing that, everyone gets the rundown of what to do.

I'm a strong believer than in any industry, the technology is the easiest thing to learn.  Our election workers don't have any problems with voting machine technology or the iPad technology (for our electronic manual, for instance).  It's the procedures, the one-offs, the mental work that is the toughest.

When the mental work intersects with technology, technology looks like it's to blame.  One worker from Tuesday's election suggested that we give out the manual electronically.  We know the iPad was never pulled out of the supply bag at that location.

One of our more alarming pieces of feedback from this election was the suggestion that we separate new election worker training from those who have worked before.  Um, we do that, and even have a bit of a preamble at each training about the type of training they are attending.

In our refresher training, workers will ask questions that often are on others' minds (or should be).  Occasionally, someone goes to an absurd place: "so a dog comes with a photo ID that an owner got as a joke...." (no, not really, that was an absurd example to protect the genuine absurds).

It's common to hear moans when questions are asked.  I have relatively good moaner radar and make mental notes of the moaners.  Generally, the know-it-alls know it less than think they do.

We have lots of things we want to do to modify our training and had hoped to make hay with it and other planning this summer and fall.

Less than 24 hours after Tuesday's election, we found out we will have another special election, a mail-ballot in Overland Park.  Combined with a mail-ballot election in the Olathe School District, the majority of Johnson County voters will be voting again this year.


The special elections and the turnout this year highlight this year's model of "The Last Problem."

With turnout in this election at 8 percent and two summer special elections scheduled, the push to move spring elections to November has legs.

(That's right, by the way--turnout in a full county election on a sunny day in April was in the single digits and not even twice what it was during a February blizzard election that had a 4.5 percent turnout).

I'm not sure if moving these elections to November will increase turnout, but I do think a case can made that it will, just from the predictability factor of November elections.  People are transient, coming from other cities, and most of us just assume that there is an election each November.

Proponents of the move will point no further than to the turnout this year, and we could very well be conducting election worker training in the summer of 2015, explaining the election-day move as the latest "last thing" impacting our training.

We're Number 13!

Last year I posted the possibility of Johnson County elections being among the top 10 attractions in the Kansas City area, based on attendance.

Of course, and especially with this being Easter Sunday, I would argue that church is the number one attraction when adding up those attending masses and services throughout the year.  (It's interesting, in that context, that worship is also distributed a la polling places--although maybe not that interesting because churches often are polling places).

But with some looking at Tuesday's projected turnout of around 10 percent and wishing it were higher, I do think it's worth looking back at the total number of Johnson County voters in 2012.

388,283 voters cast ballots in 2012.  Many more cast provisional ballots that by law could not be counted.

The Kansas City Business Journal's Friday edition lists the top 25 attractions, based on attendance, in the area last year and Johnson County elections placed 13th.

Still, the amount of voters topped visitors to Union Station (364,074), the American Royal (270,000), the T-Bones minor league baseball (260,620) and Starlight Theater (230,474).

I still maintain that the Election Office was the largest attraction that didn't charge admission or a fee to participate.  So, that's something.

It's also interesting that Union Station is now home to our friends at the Kansas City Election Board.  If Missouri gets advance voting, we might just have ourselves an attendance contest in 2016!

(I typed that with a cowboy drawl.  It just seemed appropriate).

We've Got to Fix This

So much has happened in the past week that I have many topics I need to cover here.

One downside to blogging and actually working is that the work gets in the way.  It's been nutty, wall-to-wall, and until Google comes out with telepathy blogging, this old-fashioned method of catch-up will have to do.

Tuesday night, our staff was humming along uploading results to the point I thought I'd be home before the 10 o'clock news was over.  We use several voting machine stations to upload results to our server and sadly in our cross-checks, we determined that about 15 locations didn't upload.

Bummer.  That was about another 125 cards, taking an additional 20 minutes or so.  Still, we posted final results on our website at 10:15.

I then had to complete a reporting sheet to the Secretary of State's office for statewide roll-up of races and as I was faxing that, I heard that news outlets were calling the election.

Bummer, again.  I was hoping to be an Average Joe watching that at home, I thought, with an imaginary clench of the fist, "Oh, Cross Checks!"

Cross-checks are good, of course.  Hence, an imaginary clench of the fist.

So, I didn't watch election coverage when I went home.  I ate my first meal of the day, around 11:30, until I didn't have the strength to raise my fork, and went to sleep.

I missed the acceptance speech and the hub-ub-a-boo line that's got election administrators talking.

Regarding people still in line, President Obama said, "By the way, we have to fix that."

He wasn't talking about lines in Johnson County, but I wasn't happy with the lines we had.  In some locations, we had lines of about 45 minutes, particularly in the early morning.

We had lines in advance voting, too, but our advance voting turnout was lighter than in 2008.  Problem was, more people voted at the polls Tuesday than voted at the polls in 2008.

On Tuesday, more than 154,000 people voted at the polls, compared to about 142,000 in 2008.

Double problem was, we reduced polling locations in response to budget cuts, from 284 in 2008 to 221 this week.

So, nearly 10 percent more voters with 20 percent fewer polling places.  That's a bad combination.

This reduction was encouraged because of the greater acceptance of advance voting.  One county department head, in a budget meeting after the 2010 election, said she thought our voters "got too good of service," in terms of wait time at the polls. 

Really.

Early in my Sprint career, I was responsible for telecommunications relay service, which bridges persons who are deaf or hard of hearing with voice users.  An operator sits between the callers, typing what's spoken to one user and relaying what's typed to the other.

This was a call-center environment, so I'm very familiar with the issues of projecting and handling volumes of traffic.  It's akin to building a highway for thousands of cars but realizing that the entrance ramp can only let one car in at a time.

"Fix this," isn't a simple concept.  Addressing a traffic choke-point by creating more entry points (more polling places or advance locations) also implies we have some control over where voters go. 

Half of our voters voted in advance in 2008, but only 43 percent in 2012.  That's a trend that could reverse in 2016.  We heard anecdotes of voters waiting in advance lines in 2008, only to find friends had no wait at the polls.  I think Tuesday was a result of some 2008 advancers thinking they were chumps for waiting and, instead, felt the stress of waiting on election day.

Traffic and Lines During the Last Hours
of Advance Voting Monday

I talked with a reporter at The New Republic about this and have some views on what "Fix this," might mean.     Doug Chapin also had a great post about this.

The president may not even remember saying it at this point, but the election administrator community will make this topic number one, I'm sure, at the annual legislative meeting conducted by the Election Center in January.

Some of my thoughts to "fix this" were raised in The New Republic post.

A couple more "out there" solutions:

1.  Reduce Voting Options.  This is contradictory to conventional wisdom, but election administrators run three elections in Kansas--advance in person, advance by mail, and at the polls.  If Internet voting were ever allowed, we'd have four elections running.  We have the same staff size we had 20 years ago, with half of the voters we have now and before advance voting was an option.  One way to consider managing traffic is to manage the entrance ramps to highway.

2. Appoint Election Assistance Commission Commissioners.  No election administrator has been more in favor of closing the EAC than me.  Contrived through the Help America Vote Act, I think the mission of the EAC has been accomplished.  But, if  "fix this" is a mantra, this is an agency that exists and should be amped up, again.  Currently, the four commissioner seats (two Republican, two Democratic, appointed by the President and confirmed by the Senate) are vacant.  It's a Dead Agency Walking right now.  Filling the EAC Commissioner posts and adopting a new strategic intent focused on solutions, not advice, would be a potential way to address the "Fix This" objective.

3. Speaking of solutions, how about e-solutions.  Specifically, I'm talking Internet voting, email voting, Bring Your Own Device voting, electronic pollbooks, online registration, and a myriad of ways to leverage technology.

Said more directly, Internet voting is coming.  There are a lot of scientists and other smart people who refute that Internet voting can happen, often citing very valid security concerns.

But here's the thing--it's not up to them.  The number one question I'm asked by voters is, "Why can't we vote on the Internet?"  One day, the people asking the question will come up with their own answer.

We like to think we're in control of our destiny, but often a big bang occurs that causes us to react.  Internet voting will be such a thing.

Social factors will lead to Internet voting.  A legislature will pass a law that Internet voting must be implemented.  A city will decide to conduct its own elections using Survey Monkey.  A natural disaster may occur during a critical voting period (oh, wait, that did happen).

Once society decides it's time for Internet voting, the scientists will no longer have a voice.

We will go from, "why not," to "how." 

Should you disagree, I give you as People's Exhibit A the buzz related to a simple "Fix This" comment by the President.

We already are all asking ourselves, "How?"

Surf's Up!

28 days until Election Day.
7 days until the Close of Registration
8 days until Advance Voting By Mail Begins.
13 days until Advance Voting in Person Begins.

This is a short post that will likely mean more to Johnson County readers than anyone.

Still, our experience right now might reflect the current nationally.

Advance Voting in Person is 13
Days Away.

Specifically, we know several tidal waves are coming.  But we know this mostly because history tells us so.  The waves right now, though, have been manageable.

We've had a lot of interest in advance voting and we're seeing ballots returned from our military and overseas voters.

But right now, we're at about 40 percent of the volume of ballots that we sent out 20 days before the 2008 presidential election. We sent out about 36,000 that day, but we're at around 14,000 right now.

We have a postcard that hits voters late this week as a reminder of the election and I expect at least 5,000 advance applications over the weekend.  By October 17, this year's date that is 20 days before the election, we'll probably be loading up about 26,000 ballots to go to the post office.

That's a significant drop from 2008.

We don't have a U.S. Senate race as we did in 2008.  The U.S. Congress race involves the incumbent Republican but no Democratic challenger.

So, there isn't a large engine driving turnout beyond president and the local state races.  I knew that would impact turnout, initially predicting a drop from 78 percent in 2008 to 75 percent.  Now, I wonder if we'll see 73.

We expected the presidential debate to trigger activity last week, and to some degree it did, and next week will be nutty regardless because of the 12 election worker training classes we have scheduled.  If the first big wave hits Monday, as I'm now expecting, it will have double the impact because of the amped up week ahead.

For now, we sit in this odd place of about 72 hours, knowing things are about to explode.  We're ready.

An Election By Any Other Name

In an election year like no other, it seems only fitting that we are preparing for an election like no other.

That's not good, actually.

In my view, the pathway to properly managing the operations of a business is knowing the key metrics.  And to evaluate the data and pivot properly, I look for patterns.

There is no pattern for this election.  Guessing turnout for the August 7 election is, well, a guess.

It's taking me back nearly 20 years when my gig was product manager for telecommunications relay service at Sprint.  That service bridged persons who were hearing with those who were deaf, hard of hearing, or speech-disabled.  It was a service opportunity created by the Americans with Disabilities Act, grew from start-up to $65 million a year in revenue before I changed roles, and is the only business of any size where Sprint ever was the market-share leader over AT&T.

That was a call-center business and our primary unknown, after winning a state contract, was the day one call volume.  I created some variation of math soup, blending county and state populations with the population of persons who were deaf, to come up with an estimate and we were usually pretty close.  To underestimate would result in long-answer times and to over-estimate would lead to profitability issues.

Predicting election turnout has the same ramifications--long lines or excessive costs are the two bookends.  The primary cost drivers related to turnout are election workers and paper ballots.

In this election, unaffiliated voters can vote unaffiliated, Democratic, or Republican ballots and we have about 1,500 different ballot styles.  Over-building the election can easily result in more than $50,000 of unneeded paper ballots.  Properly building the election still will result in considerable unused ballots.

That's because we have no idea of the distribution of ballots even if we nail the turnout percentage.

If we predict a turnout of 25 percent, the mix of those three political distinctions can vary, so it's not as simple as expecting a flat 25 percent.  It could be a 10/40/5 percent blend or a 20/30/20 blend or any variation based on the actual races.  The blend possibilities literally are nearly infinite.

So, we look for the pattern.  2010 might feel like 2002, 2008 like 2004, and so on.  This year, though,  the pattern is missing.

This year, we're preparing for an August primary in a presidential year:

• President on the ballot?  No, that's not part of the Kansas ballot in August.
• United States Senate race?  Nope.
• Governor or other statewide races?  Not until 2014.
• United States Congress?  Yes!  But, the incumbent is running unopposed in the primary.  No candidates are running in the other party.
• County Commission Primary?  One, affecting 1/6th of the county.
• District Attorney Primary?  No.
• Sheriff Primary?  Yes.  In this year, this becomes the only countywide race, and with all deference to the sheriff, how many people really know what a sheriff does in an urban county canvassed by police from several cities?

We have plenty of state House and Senate primaries, and candidates have been working the neighborhoods.  To some degree, those races are similar to city races, where turnout is driven by door-to-door campaigning.  That said, city races generally have turnouts in the 10 percent range.

I've looked at even-year August elections back to the early days of the Internet and our website ("modern" times, anymore) and there has never been an August election without a competitive race for a national or statewide office.

I'm predicting a 20 percent turnout simply because that's what it feels like.  I can't scientifically define the "20-percent feeling," and that bothers me.  We always build the election for slightly more than my prediction, so we're preparing for 25 percent, or about 75,000 voters.

Our mailroom is starting to get loaded with advance
voting applications.  This is just a smattering.
Ten temporary employees are entering in applications
in another room to get ready for our first mailing
next Wednesday.

Right now, we are entering applications for advance ballots my mail.  I expect we will send out about 8,000 next Wednesday, the first day allowed.  If that represents 5/13 of the advance voting number (the metric from 2008), that would give us about 21,000 advance voters.  

If those voters represent a third of the total, we'd have 63,000 voters.  If half, we'd have 42,000.

That's a turnout range between 11 and 17 percent.  Those are downright spring-like numbers, not August numbers.  

Of course, who knows...I remind everyone that I'm particularly good at predicting turnout the day after the election.

Right now, though, it sure feels like a low turnout and, as a result, many close races.  They number to watch in the next few days will be the number of advance ballots we send out the first day.  If that number jumps to 10,000 or 12,000, all those extractions balloon and we'll be into numbers that look more like August.

Ahead of the T-Bones, Behind the Royals

What will be the top free Kansas City area attraction in 2012?

It could be the Johnson County Election Office.

The Kansas City Business Journal has just released its Top 25 Area Attractions based on 2011 attendance.

If you considered "attendance," to be the same as "number of votes cast," our attendance of 57,550 votes in 2011 didn't even crack the top 25.  The Harry S. Truman Library & Museum is number 25 with 2011 attendance of 98,000.

In 2012, we've already had 48,150 votes cast, still a far cry out of the top 25.

But, if our 2012 August and November election numbers are the same as in 2008, we'd finish with 411,026 votes cast.  A modest 10,000 voter increase because of registration growth from these two elections would put us at approximately 421,000 votes.

That would be good enough for 10th on the area attraction list this year.  If 2012 numbers are the same as in 2011, here's how it would look:

1.   Kansas City Power and Light District                          9,000,000
2.   Ameristar Casino Hotel Kansas City                            3,364,177
3.   Harrah's North Kansas City Casino and Hotel           2,553,000
4.   Kansas City Royals Baseball Club                                1,724,370
5.   Worlds of  Fun and Oceans of Fun                               1,300,000
6.   The City Market                                                                 720,000
7.   Kansas City Zoo                                                                 713,813
8.   Kansas City Chiefs Football Club                                    700,000
9.   Kansas Speedway                                                              600,000
10. JOHNSON COUNTY ELECTIONS                   421,000 (votes cast)
11. Watkins Woolen Mill State Park                                      420,000
12.  Nelson-Atkins Museum of Art                                        410,000
13.  Deanna Rose Children's Farmstead                               400,000
14.  Sporting Kansas City                                                         400,000
15.  Kansas City T-Bones                                                         261,115

Voting in Johnson County is a more popular Kansas City area activity than minor league baseball and major league soccer, but less popular than pro baseball, pro football, and racing.

Of the top 10, only elections are truly free (no cover charge, admission charge, parking charge, or items that are purchased).

Granted, you could consider "church" or "school" to be an attraction and add up all the attendees, but I've only listed the voters involved in Johnson County Elections, put on by one office. We've had quite a schedule in 2012, as many elections as the Chiefs will have regular-season home games.

So, the next time someone criticizes voter turnout or compares the number to those discussing American Idol voting, you have a snappy comeback:

"I'll have you know that all the votes cast in Johnson County in 2012 create a total admission that makes it the number one free area attraction in Kansas City."

Fractions, Onions, and Inactive Voters

Much of this blog thus far has been an intro, the setup to the plot.

You're meeting the characters and getting some of the background so posts over the spring and summer are 10 paragraphs instead of 100.

This is one of those background pieces.

I traded emails with a friend this weekend about turnout this November and reminded him that I've been predicting a drop in the turnout, 2012 from 2008, since the day after the 2008 presidential election.

That's important, because if the turnout is lower, the convenient story line will be that it happened because of Photo ID.  That won't be why--photo ID legislation was barely a gleam in the eye when I began predicting lower turnout in 2012  3 1/2 years ago.

I believe the Johnson County turnout will be 75 percent, down from 78 percent in 2008.  I have some thoughts as to why the turnout will be lower and will present that in future posts.

For now, I wanted to take you back to the third-grade and suggest that turnout percentage is merely the first layer of the onion.  As a third-grader, I did not like onions, but that's also when I hit fractions and division with zeal, and that's the part of the third-grade we're visiting.

You see, turnout percentage is simply the number of voters divided by the number of registered voters.

Unless, in Kansas, if you are looking at a mail-ballot election, where turnout reported to the Secretary of State is the number of voters divided by the number of active voters.  For purposes of our website reporting in those elections, though, we show the percentage "straight"--votes divided by voters.

Active and inactive voters become terminology needed to understand turnout.  That's because a turnout percentage relies on the numerator (the number of voters in this case) as much as it does the denominator (the number of eligible voters).  Year to year, state to state, these simple variables can, um, vary.

Inactive voters would be eligible voters if they came into the polling place to vote.  But it is highly unlikely they will do that.

Inactive is code for, "We don't think you live there anymore."  Once inactive, a voter is included in the denominator (the number of eligible voters) although the likelihood of the inactive voter voting is very small.

How does someone move from active to inactive?

Some communities use the post office's national change of address (NCOA) to get updates.  Johnson County utilized NCOA about 25 years ago and it was essentially a disaster because the county's street guide doesn't always match the post office's and the result can be someone living on a border street being reclassified to the wrong city and getting the wrong ballot.

Postcard Mailed in Roeland Park

The other method of verifying addresses, by federal law, is to send a confirmation mailing to all active voters once a year.  That's what we do.

Every election, we send a postcard to each registered voter who's not inactive.  The postcard tells the voter about the upcoming election but the postcard can't be forwarded.  If returned to us, we flag the voter as inactive and send a second mailing that can be forwarded.  This mailing is a, "Whassup?  Did you move?" kind of notice where we ask the voter to modify his or her registration or inform us of a change.

If nothing changes on the voter's record from that point, and the voter doesn't vote, the voter can be removed from the registration list after two federal elections.  So, someone marked inactive with the upcoming Roeland Park election would be removed after the November 2014 election.  If the voter votes in that time, the voter becomes active again.

So, at any time, the voter registration list has a fair number of people registered who no longer live at the address registered.  And, by law, they can't be removed.

Factors influence the number of inactive voters.  One major factor is the economy--people have moved less frequently, I suspect, over the last two years and that makes trending on percentages alone imperfect.

Another factor is the cards themselves.  We began sending them before every election in 2007.  The frequency of the cards and elections illuminates someone as a potential inactive voter much earlier than an annual mailing only.  Depending on the timing of the inactive tag, someone could remain on the rolls for 2.5 to 4 years.

Because the purging happens after a November election, the number of inactive voters is highest heading into a gubernatorial or presidential election.  Thus, the denominator is larger than it should be if you were looking for an accurate turnout percentage.

Whew, that's a lot of typing, and I've only scratched the surface.  But, in looking at turnout, start with the numerator--voters.

Voters are real and it helps me if I start thinking about how many of them will vote in advance, and how many of those who vote in advance will do so by mail or at one of our in-person locations and then, further, how many would vote at each of those locations.

In many ways, the percentage is meaningless.  Number of actual voters is the true metric.

There will be a lot written (and there has been a lot written) about voter participation and turnout.  It almost always is in terms of percent.  If you're digging into it, take a bite out of the onion and go beyond the first layer of numbers to draw meaningful conclusions.

Return of the Blue Bear

Yesterday at noon was the spring candidate filing deadline and we're in a mad dash to prepare for the Feb. 28 primary.  

We hope to send a ballot order to our ballot printer in Washington state by Friday, getting ballots back by late next week.  That's the best-case scenario, but one we push for so we have time to sort the ballots before sending out those for advance voting 20 days before the election (the 8th would be the first day we can do that).

To develop a ballot order, we first need to determine the turnout, which usually is different precinct-by-precinct, polling place-by-polling place, year-by-year.  We have a mayoral primary in Shawnee, so we expect a higher turnout there than some of the individual council race primaries we have elsewhere.

While I make the call on turnout for planning purposes (and we want to guess high so we have plenty of paper ballots), we have an internal poll taken in our office to see who can, in true "The Price is Right" style, come closest to the actual turnout without going over.

The prize for the staff member who comes
closest to correctly forecasting turnout.

The winner gets the little Blue Bear traveling trophy.  The Blue Bear lives in the winner's office until the next election. 

Once, the Bear lived with me for a year and later with another staff member for two years.  I haven't been a Blue Bear winner in a few years, actually, and that's a good thing.

The point of this is just to get everyone engaged in the numbers of elections.  We want to build future leaders of our office who understand all aspects of our business.  Accurately understanding and predicting turnout runs akin to tracking costs.  We don't want to prepare for a 30 percent turnout if the turnout is 15 percent, or vice versa.

The Blue Bear was a registration giveaway I received at an Election Center conference in 2006.  In a couple weeks, we'll gather everyone's estimates.  For now, we've estimated 15 percent, which is high for a primary.  But that number is used for planning, and I like to plan for about a 5 percent higher turnout than I think we'll have.

One thing that's unique about the spring elections is candidate rotation, so that every candidate has, as much as possible, equal time at the top of the ballot.  I'll explain that more in the coming days.