This is the P2PU Archive. If you want the current site, go to www.p2pu.org!
Managing Election Campaigns
My recent threads
You haven't posted any discussions yet.
Recently updated threads
- Concluding the class Tue, 2010-11-23 00:49
- Week Six: Get Out the Vote (GOTV) Sun, 2010-10-31 03:49
- Week Four-Five: Candidate Statements, Direct Mail, Campaign Organization Fri, 2010-10-22 06:46
- Week One: Understanding the problem Fri, 2010-10-22 06:15
- Miscellaneous Thu, 2010-10-21 16:20
Week Two: How Voter Data Informs Election Strategy
Go back to: General discussion
This week we will look at what voter data is available. Please begin by reading my next posting under this forum and taking a look at the report of an election in San Francisco, California.
Also, a ritual apology. I am attending the OpenCourseWare Consortium's board meeting and the meeting of the Japan OpenCourseWare Consortium, so my timeezone so I haven't posted in the last two days as promised.
In California, county registrars of voters manage election campaigns. Voter data, which includes personally identifying information, is semi-public. You have to be either a candidate or a company or consultant working on behalf of candidates (such as voter data companies) to be able to request voter data. There is a minimal cost for the data, usually well within the reach of any campaign.
When computer storage was more expensive, the data would arrive in two parts - a packed text file in which each line of text corresponded to one voter. The second file was an explanation of the packed format: the first so many characters would represent the voter id, etc...
Here's a brief list of what information is provided:
Voter name (prefix, first, middle, last, suffix)
Voter address (house number, house number suffix (apartment of subdivided house), street name, street name prefix, street name type (boulevard, street, avenue, etc...), street name suffix)
Phone number (optionally provided by the voter, usually on 50% of voter records).
Precinct number (this is a key piece of information)
Voter history! For the past elections, whether (not how) the voter voted and, if he/she voted, how (absentee or at the polls).
Voter absentee stqtus (permanent absentee - a request to always have the ballot sent by mail and returnable by mail or at the polls).
Voter party identification (all recognized parties have codes and each voter is identified by party affiliation).
My next post: why this matters
I am curious about statement "You have to be either a candidate or a company or consultant working on behalf of candidates (such as voter data companies) to be able to request voter data." The federal law Help America Vote Act mandates that every state create a statewide database of all registered voters. Why is this information not available under your state's Freedom of Information law to anybody that wants it? It is in Michigan - Ohio actually has 100% of their registered voter data online!
Here is a report on the Help America Vote Act @ five years which has within it a state by state summary of pertinent info including that status of the voter database:
http://www.pewcenteronthestates.org/uploadedFiles/HAVA.At.5.pdf
You know, I was completely ignorant of this act, probably because I haven't actively managed a campaign in a few years. However, when I tried to get local voter data from the county registrar of voters, I was told that I had to be a candidate, work for a candidate, or have a formal letter from a university specifying the use of the data.
I'm going to do a little research now. :)
In Michigan the legislature moved the primary up to an earlier date for the 2008 election and also added a provision that the voter registration lists for the primary would only be available to the state party chairpersons. After much litigation, our Appeals Court ruled the primary voter registration list were available to the public under our FOIA laws:
Practical Political Consulting v State (COA 291176)
http://coa.courts.mi.gov/documents/opinions/final/coa/20100309_c291176_3...
Dissent
http://coa.courts.mi.gov/documents/opinions/final/coa/20100309_c291176_3...
Great information.
Example Case #1: voter data for an election with 100,000 registered voters.
Okay, let's use what we know about voters to come up with a scenario that shows how to use the data correctly. We will assume that you have already imported all the voter data into your own custom database to be able to analyze the situation.
Step #1: Determine your target "universe." Campaign professionals use the term "universe" all the time. Once they have a "select" from their database, they are concerned only with those voters. What basic criteria go into selecting a universe?
Step 1a. Determine how much money you will spend on your campaign. Until the campaign budget is established, you won't know how much you will spend on voter contact: direct mail, lawn and house signs, phone banks - including donuts and coffee -, etc... Later, we'll get down to specifics about budgeting, but for now, you have to make at least an initial guesstimate about your budget, eg. your capacity to raise funds. In this example, let's assume that you can raise $35,000 - not a lot of money.
Step 2b. Identify likely voters. First drop everyone who voted only 2 out of the previous 5 elections or less. Don't waste your time and budget on someone who only votes in high-turnout elections and even then only occasionally. Okay, we just got ride of at least half our target universe. We are probably down to 50,000 voters who vote at least 3 out of the last 5 elections. However, we only have a budget of $35,000 and we cannot afford to target 50,000 voters. Now we try it again with more selective criteria: voting in 4 of the past 5 elections. This yields us a nice reduction to 25,000 voters. Remember, the goal is to identify the smallest number of voters that you have to contact in order to win the election, unless you have a really big campaign budget.
Step1: ignore the potential voters who won't/didn't vote.... what about finding out why they decided to abstain in the first place to work on an angle to motivate them to become active. there has to be a valid reason why they didn't make any effort to vote to begin with. when talking to them develop a way to contact them directly with electronic media and answer ans many of their questions about your plans for their community and keep it open so others can also ask follow up questions in a forum (if possible or use a system similar to twitter or status.net) if they wish to find a solution to their problems.
Everyone has an opinion of how things need to be fixed. make a list of things that they want done and create a system to explain what you will be attempting to achieve.
avoid making promises. they will fail. set out realistic goals that even the community can chip in to help see through if your campaign fails to even win. don't depend on wining a title to win ;-)
step 2:of the likely voters look for individuals who really understand the goals of the campaign and have them lead discussions with the volunteers to make sure they ask as many questions as possible to prepare them for unexpected questions. worst case scenario they can call/msg/txt/email HQ for answers.
as for the direct marketing i know that as a college student i would collect all the material and dup it straight into the trash. sometimes i'd take it right outta their hands and drop into a pile in a bucket right next to the door-in plain sight! besides creating waste and stress fliers that get left in mailboxes just simply go to waste just like the credit card offers (don't know which i throw out faster). i don't feel they work.
smaller bizz cards with links listed for party discussion forms that ppl can participate in or even a hash tag they can follow on a social networking site to join in with other talking about some of the issues your party is raising.
toughts?
Is Social Networking Useless for Social Change? A Response to The New Yorker's Malcolm Gladwell
What if anything is the potential contribution of web-based "social networking" to social movements and social change? - http://bit.ly/c8O4bz Alternet article/post
I'm going to make a small, almost irrelevant point here and encourage others to jump in.
The goal of campaign literature is to attract the attention of the voter somewhere between their mailbox and the trash. All campaign literature ends up in the trash except in the homes of political junkies. So what message can you devise - what graphical image - that a voter will remember (along with the name and office of your candidate) in the 30 seconds between mailbox and trash.
In a door-to-door campaign, the thirty seconds to two minutes of contact with a voter at the door strikes a much stronger impression than any literature. However, for those houses where the literature is left on the doorstep because the voter wasn't home, it still goes back to the 30 seconds on the way to the trash.
So design your literature carefully!
Step 3: Develop microtargets
Microtargeting is the use of voter data to make small selects, typically in the hundreds or few thousands of voters who may have unique characteristics. For example, I often wrote a special "green" mailer that went to registered voters of the Green Party in the city or county I was concerned with. Now most Green Party voters are young and do not vote as frequently on the average as other voters. So, I selected the very small group of Green Party voters who go to the polls even if both of their legs are broken. Why would I do that? Because you don't have to win over the entire electorate. You are trying to push a percent or two with each tactic.
By the way, I used the Green select both with and without formal Green Party endorsement. In which case, I would take some positive statement from the Green Party mailer and use it as the primary cutout. I positioned this cutout not only in the text of a letter, but also on the outside of the envelope that also bore the candidate's name and large text: "Attention Green Party Voters."
I'm looking forward to your comments. :)
with open projects like this loopy but practical mapping one. a guy set up a map to show the distribution communities in an electorate area. this would be more useful in knowing more about the electorate in an area than just numbers on paper about ballots cast. combined they could make a very interesting tool if used well.
"Any city-dweller knows that most neighborhoods don't have stark boundaries. Yet on maps, neighborhoods are almost always drawn as perfectly bounded areas, miniature territorial states of ethnicity or class. This is especially true for Chicago, where the delimitation of Chicago's official “community areas” in the 1920s was one of the hallmarks of the famous Chicago School of urban sociology. And maps showing perfectly homogeneous neighborhoods are still published today, in both popular and academic contexts alike. " Radicalcartography > http://bit.ly/c5FJ2x
I've done my campaign work in Northern California, and specifically Oakland, which, for the West Coast, had the reputation of being the most segregated city, with black neighborhoods and white neighborhoods. The first campaign consultant suggested that there be no door-to-door campaign in the black neighborhoods, because our candidate, who was white, would be wasting her time. He argued that people, to be friendly, would say that they would vote for her, but that it was unreliable information, unlike results in the white neighborhoods. We ignored the advice and our results in African-American neighborhoods were very positive. And results we got at the door tracked well with voters who turned up to vote and with actual precinct results.
So your point about the composition of neighborhoods is well-taken. Ethnicity is a strong driver in many countries in voting behavior. I guess my point above is to sometimes take the gamble and test your candidate's appeal - even against the common wisdom.
Step 4: Analyze individual voters' likely voting preferences from voter data
Precincts (neighborhoods) tend to vote alike. That's because they are usually homogeneous in their wealth, their ethnic composition, their education, etc. So even when you have a split electorate (50-50), by neighborhood you may find that you have voters who will definitely support your candidate or issue. In that case, you may choose to adjust your targets to so that everyone has voted 3 out of 5 elections is included and not just 4 of 5 elections. The "get out the vote" (GOTV) effort then can focus on getting as many people who live in the same neighborhood to the polls. Again, our criteria is finding tactics that give us an extra percentage point. And any marginal vote picked up anywhere is as good as another marginal vote.
Step 5: Using low-turnout elections
What's good about low-turnout elections? Certainly nothing from the standpoint of democracy. However, a low-turnout election shows you which voters "always" vote. Even better, an election that is low-turnout, but has an important issue, like a school bond, on the ballot, then lets you know who the education voters are. You can use that information in high-turnout elections later!
The key is whether the low-turnout election had a lopsided result, like 70-30. Now you know can use your microtargeting techniques to ensure that you are excluding the 30% who might vote against your pro-education candidate.