In case you missed the earlier installments in the series:

How the Iowa caucuses work, part 1 (basic elements of the caucus system)
How the Iowa caucuses work, part 2 (corrects an error in part 1 and discusses who is over-represented and who is under-represented when delegates are counted)
How the Iowa caucuses work, part 3 (why it's hard to turn out caucus-goers)
How the Iowa caucuses work, part 4 (more about why caucus turnout is low)
How the Iowa caucuses work, part 5 (on second choices and caucus math)
How the Iowa caucuses work, part 6 (on how precinct captains help their candidates before caucus night)
How the Iowa caucuses work, part 7 (why it's hard to figure out how well the candidates are doing in Iowa)

If you like political geography, you may want to check out Where the Iowa field offices are. I published a new version of that diary because several of the campaigns added more field offices in the past month.

Back to today's topic. There are many ways to win your precinct, but they all have one thing in common: your candidate must be viable.

The winner of the Iowa caucuses is the candidate who wins a plurality of state delegates. Precinct caucuses elect county delegates. The county delegate totals are converted into estimated state delegate equivalents using a mathematical formula. You want your candidate to win as many county delegates as possible, and to do that, you need your candidate to be viable in as many precincts as possible.

Iowa has 99 counties and 1,781 precincts. Most of those precincts will elect four or more county delegates, and a candidate will need the support of at least 15 percent of the people in the room to be viable.

(Side note: in precincts that assign just one county delegate, the threshold is 50 percent. In precincts that assign two county delegates, the threshold is 25 percent. In precincts assigning three delegates, candidates will need at least one-sixth of the voters in the room to be viable.)

If your candidate is a longshot, crossing the viability threshold may itself feel like victory for you. My friends who caucused for Dennis Kucinich in 2004 were satisfied if he was viable in their precincts. They didn't mind winning fewer delegates than the top-tier candidates did.

In the comments below an earlier diary in this series, Daily Kos user clonecone mentioned that he and other volunteers were almost able to make Wesley Clark viable in his Iowa City precinct. That is an incredible achievement, considering that Johnson County, where Iowa City is located, had the highest turnout in 2004 in terms of caucus-goers per state delegate assigned. Also, Clark ran no ads in Iowa and barely visited the state.

But as they say, close only counts in horseshoes and hand grenades. If your candidate is not viable, you can't win any delegates.

Assuming your candidate has enough support to be viable, here are some approaches to winning your precinct:

1. Have more supporters in the room on caucus night than any other candidate

All of the campaigns have had field organizers working hard to identify supporters and persuade leaners and undecided voters. Ideally, local volunteers will take on these tasks as well. Compared to staffers, neighbors have better chance of getting voters to pick up the phone, as a prominent Democrat from Des Moines recently noted:

"It takes local people with local knowledge," says [Ned] Chiodo, who is backing Clinton. "Everybody has caller ID these days. They don't pick up the phone unless they know you."

I wrote last month about the benefits of planning and hosting a house party for a candidate.

In these final days of the campaign, volunteers need to make choices. How should I spend my time effectively in order to make the Edwards corner the most crowded part of the room at our caucus?

Should I call all known supporters one more time? I could remind them of the caucus time and location, let them know how much we are counting on them and ask them to bring a friend from the neighborhood. But that might irritate people who would probably show up on January 3 even without another call from me.

Should I call or door-knock the experienced caucus-goers we've previously identified as undecided? It's quite likely that they will show up on caucus night, but it's also a safe bet that they have been inundated with calls these past few weeks.

Should I call people who were supporting or leaning toward another candidate to see if they would now be open to supporting Edwards? The Obama campaign is calling some previously identified non-supporters, according to this post by an out-of-state volunteer:

Iowans support for candidates is more like sand than concrete.  The following notes were attached to one voter I called:  6/1/07:undecided; 7/2/07: leaning Obama; 9/4/07: leaning Clinton; 10/5/07: solid Clinton; 10/15/07: undecided  Today when I called her she said she was leaning toward Edwards.  At first I thought I was dealing with a particularly unstable voter, but after talking to 50 voters with similar call records, I realized just how fluid Iowans' candidate preferences are.  This, along with the viability issue, is why no one really knows who will win the Iowa caucuses.

Clearly a lot of voters change their minds, so there is potential to find new supporters this way, but on the other hand, there is also potential to annoy Iowans:

I was contacted by Obama in June and said I was an Edwards prec cap.  Fast forward to October, same phone call.  Fast forward to early Nov, same call at which time I told them they were beginning to resemble the Dean camp where they can't scrub a list and would wind up irritating people.  Two weeks later, same phone call, same conversation.  Last week, Team Obama wanting to know if I had made up my mind who I was going to caucus for.  If you can't clean a phone list, look for the irritant factor to play a part espically [sic] if people get 6 or 7 calls between X-mas and caucus.

Should I contact Democrats who did not attend the caucuses in 2000 or 2004 to try to turn them out for Edwards? Trying to persuade people to attend a caucus for the first time seems like a low-probability endeavor. On the other hand, I have a friend who was concerned that she and her husband might be stuck at the caucus until late in the evening, messing up her children's bedtime routines. I was able to explain the process and assure her that she should be done by 7:30 or 7:45 at the latest, and she is now planning to caucus with her husband.

Should I try to contact people who aren't on my list of Democrats--say, people who have moved into the neighborhood recently, or registered independents? The advantage is that these people may not have been contacted as many times as the others. The disadvantage is that they are unlikely to attend the caucus.

Roger Simon wrote an article on different approaches to the ground game in Iowa, and I was struck by the following passage:

John Norris, who was Kerry's Iowa director in 2004 and is now an Obama volunteer, thinks any campaigning that matters will end about Dec. 20, which is why the ground game is reaching a fever pitch right now.

Norris talks about a woman who supported Edwards in 2004 but who is now supporting Obama. Why?

"Because an Edwards volunteer only knocked on her door once and we knocked on her door several times," Norris says.

As a onetime precinct captain for John Kerry, I have great respect for John Norris, who held that campaign together in the face of terrible polls for months. But as a rule, I would avoid knocking on the same door several times, even if I had unlimited time for voter contacts. For every person who was impressed by my dedication, there would probably be others irritated by the Edwards volunteer who kept bothering them.

A former state legislator, who has knocked on thousands of Iowans' doors, told me that at a certain point you risk alienating voters if you contact them again and again. Nate Willems, a regional director for Howard Dean in Iowa, has also touched on this problem:

A late 2003 Dean focus group produced the comment from a participant, "I'd give anything for those Dean people to just quit calling me."

The bottom line is that it's hard to know where to focus my efforts in order to maximize the number of Edwards supporters who show up on the evening of January 3.

Another sobering fact is this: even if my candidate has the largest number of supporters walk into the room on caucus night, he may or may not win the most county delegates from my precinct. Which brings me to another way to win:

2. Take the lead on second choices

A candidate who trails after the first division into preference groups can sometimes gain ground when caucus-goers make their final choices.  

Dan Conley was an Edwards captain in 2004 and recently described what happened in his precinct:

Raw votes were 18 for Kerry, 15 for Dean, 15 for Edwards, 10 for Gephardt, 2 for Kucinich.  That's 30% Kerry, 25% Dean, 25% Edwards, 12.5% Gephardt, 2.5% Kucinich. But guess what ... none of those Gephardt or Kucinch votes counted because they didn't meant the 15% threshold.

So that 12 people had to redistribute ... and they went 6 Kerry, 6 Edwards for a second round of 24 Kerry, 21 Edwards, 15 Dean.  The eight delegates were split up Kerry 3, Edwards 3, Dean 2 ... which is 37.5 percent Kerry, 37.5 percent Edwards, 25 percent Dean.

The end result in Iowa is a statewide tally of that last number, the distributed delegates, not the raw vote total.  So you could get the raw vote 100 percent right and still be WAY, WAY off on caucus night because you don't know how the pinballs will bounce.

If you enjoy scenario spinning, download a caucus calculator, enter a number of delegates for your pretend precinct, and watch how you can change the delegate allocation by altering the number of viable candidates and the percentage of supporters for each candidate. Daily Kos and MyDD user asahopkins created this Iowa Caucus Calculator, or you can use the caucus math spreadsheet that Drew Miller created earlier in the year.

Let's say 200 people show up in my precinct on January 3. Candidates will need 30 supporters to be viable. Let's say that after the first division into preference groups, we have 60 people for Clinton, 55 for Edwards, 45 for Obama, 20 for Richardson, 15 for Biden and 5 for Dodd. Richardson, Biden and Dodd are below the threshold and will not get any of our precinct's 6 delegates. According to Drew's calculator, Clinton, Edwards and Obama will each get 2 delegates.

There are 40 people supporting non-viable candidates. If I get 20 of them to come over to Edwards, and the rest split evenly between Clinton and Obama, the delegate count remains at 2-2-2. If I can get 30 of them to come over to Edwards, he will probably win 3 delegates, with Clinton staying at 2 delegates and Obama dropping to 1 delegate.

One extra delegate may not sound like much, but its significance is magnified when you consider that any additional delegate my candidate earns is a delegate subtracted from someone else.

This zero-sum game brings me to another way to improve your candidate's chances in your precinct:

3. Keep one or more candidates below the viability threshold

As a general rule, having fewer viable candidates will increase the potential number of delegates my candidate can win. For example, a neighboring precinct in my suburb awards 8 Polk County delegates. In 2004 Dean was not viable there, and the precinct gave 5 delegates to Edwards and 3 to Kerry.

This year, with three candidates very likely to be viable, and an outside shot at viability for Richardson or Biden, it will be almost mathematically impossible for Edwards (or any other candidate) to win 5 of that precinct's 8 delegates.

Last week I spoke with a friend who's a precinct captain for Obama in a very Democratic area of Des Moines. Her precinct will award 9 Polk County delegates, and a month or so ago she was feeling good about the number of Richardson supporters who told her Obama was their second choice. Now she feels almost certain that Richardson will be viable in her precinct. In 2004 her precinct awarded 4 delegates for Kerry, 3 for Edwards and 2 for Dean. But if Obama, Edwards, Clinton and Richardson are all viable, it becomes much harder for her to pick up 4 delegates for Obama and almost impossible for her to get 5 delegates.

If four candidates are viable in my precinct, Edwards would have to be way ahead of anyone else to win 3 out of the 6 delegates, as a few minutes with a caucus calculator shows.

A much more likely outcome would be something like 2 delegates for Edwards, 2 for Clinton, 1 for Obama and 1 for Richardson. But if Richardson stays below 15 percent, I have a much better chance of winning that third delegate for Edwards--especially since the precinct captain for Richardson in my neighborhood has already told me that Edwards is his second choice.

Is there anything a volunteer can do to keep other candidates from being viable in the precinct? Whether a large number of undecideds break to Biden or Richardson is mostly out of my control. However, I may want to make extra calls to voters who have previously told me they are on the fence between Edwards and Richardson. Perhaps I could loan them my videotape of the Des Moines Register debate, if they didn't catch it, or give them a copy of the Edwards policy book, if they haven't received one.

Alternatively, I could work harder to turn out first-time caucus-goers leaning to Edwards. If Richardson has 25 supporters in the precinct, he would be viable with an overall turnout of 150 but would not be viable with an overall turnout of 200. In that sense, having a few extra Obama or Clinton supporters show up on January 3 may not necessarily hurt Edwards, if it keeps Richardson below the threshold.

Keep in mind that there will be precincts in which a top-tier candidate is not viable but a second-tier candidate is viable. An Edwards precinct captain who covers a different suburban neighborhood told me he is sure that Obama, Edwards and Richardson will all be viable in his area, but Clinton may struggle to reach 15 percent.

If I don't have enough people in our corner to win 3 delegates for Edwards, there is one more thing I could try:

4. Take a delegate away from a leading competitor

I am 100 percent confident that Edwards will be viable in my precinct. I am not too worried about having enough supporters to win 2 delegates in my precinct. While my goal is to win 3 delegates for Edwards from my precinct, I would reluctantly settle for 2 delegates so long as no candidate finishes ahead of us.

As I described in part 5 of this series, Paul Simon had the largest number of supporters at my precinct caucus in 1988. After the first division into preference groups, it seemed likely that Simon would win 3 of the 6 delegates in the precinct.

However, enough people shifted from Michael Dukakis to Bruce Babbitt to change the way the mathematical formula would convert raw votes to delegates. Simon, Dukakis and Babbitt all ended up winning 2 delegates from the precinct. Since only delegates matter (not raw votes), the result was a tie, despite the Simon group's initial numerical advantage.

That tie could be considered a big victory for Dukakis, since it kept Simon (who had a good chance of winning Iowa) from earning an extra delegate. No doubt the Babbitt supporters in the room were also quite pleased.

If Edwards has no chance of winning 3 delegates, even after our best efforts to bring over the supporters of non-viable candidates, I may consider encouraging a small number of supporters to caucus for another candidate.

This might benefit the third-place candidate in my precinct, or perhaps a few extra people could bring Bill Richardson or Joe Biden up to that 15 percent threshold. If four candidates are viable, our precinct's delegates would probably be divided 2-2-1-1.

It's even conceivable that Clinton's precinct captain could encourage a few people to join the Edwards group if she realizes that she has no chance to win 3 delegates for Hillary and that bringing Edwards' total to 3 delegates would drop Obama down to 1 delegate from our precinct. She may prefer a three-way tie with 2 delegates for each, or she may prefer to finish behind Edwards while beating Obama.

Conversely, my friend the Obama precinct captain has told me that if a delegate in her precinct is up for grabs and she can't earn it for Obama, she will try to ensure that it goes to Edwards rather than Clinton.

Important note: According to desmoulins, the Nevada Democratic Party adopted slightly different caucus rules and will not allow supporters of viable candidates to realign at the second division into preference groups. Good for them. If only people supporting non-viable candidates can go to their second choice, then there will be far fewer opportunities for precinct captains to play around with the caucus math.

Iowans, don't try this in your precinct unless you are sure you know what you are doing. Bleeding Heartland user corncam posted this cautionary tale:

At my 2004 caucus, the Kerry people shifted some votes to Edwards, thinking that it would cost Dean a delegate, but they miscounted, and the new Edwards delegate came from Kerry, not Dean.

The caucus calculators I linked to above are great tools if you want to explore these scenarios for yourself. Thanks again to asahopkins and Drew Miller!

Please take the poll and comment, especially if you've attended a caucus and have a story to share about how your precinct was won.

If you're coming to Iowa to volunteer, welcome! I appreciate the efforts of people who are walking the walk for their favorite candidate.

Poll What should your candidate's volunteers focus on in the final days before the caucuses? Contacting known supporters to remind them of the caucus location and ask them to bring a friend Contacting previously identified undecideds and leaners to try to persuade them Contacting non-supporters to see if they will change their minds Contacting people who have never caucused before to explain the process and encourage them to attend They can do whatever they want--the Iowa caucuses will still be a travesty of a mockery of a sham of a mockery of two travesties of a mockery of a sham Votes: 22 Results : Vote Link : Polls