Iowa Robopoll Report 6/30/2011

rp08orbust

Member
Joined
Dec 20, 2007
Messages
4,803
Why am I helping rp2012.org?

Ron Paul's Iowa campaign team have made it clear that the best way to help Ron Paul win the Ames straw poll right now is to donate to the Iowa projects and volunteer for the campaign's phonebanking among Iowa Republicans. However, whether you like it or not, the grassroots phonebanking project, rp2012.org, which targets Iowa independents, appears determined to continue. Given this, my goal is to make rp2012.org as successful as possible and provide a way for those who have maxed out their donations to the official campaign, but who do not have time to phonebank, to continue supporting Ron Paul's efforts to win the Ames straw poll.


Why robopolling?

First, the purpose of the robopolling is not to measure Ron Paul's support in Iowa. I have deliberately skewed the voter sample to maximize the chances of identifying Iowa independents who support Ron Paul. Rather, the purpose of the robopolling is to do just that: Identify Ron Paul supporters who can be urged to vote in the Ames straw poll.

There are many reasons why I favor robopolling over manual phonebanking for the purpose of identifying Ron Paul supporters.

First, the robopoll messages are completely neutral (if anything, they are biased against Ron Paul by his placement in the list of choices) and, except for my Skype caller ID, they are anonymous, so they do not reflect badly on Ron Paul in any way among Iowa voters. (And yes, the robopoll is legal.)

Second, the robopolling is extremely efficient: With this first poll, I've identified as many Ron Paul supporters as the rp2012.org phonebankers have been identifying in a week or more, and it cost only $100.

Third, by separating the identification of Ron Paul supporters from any Ames straw poll invitations, we can ensure that invitations get handled by the best human callers possible. I'll offer the robopoll results to the official campaign first. They can probably accept the results of robopolls that I paid for myself as personal contributions to the campaign (I am far from maxed out), but I'm expecting any future robopoll results paid for by grassroots donations to be rejected due to campaign finance laws. In that case, rp2012.org folks can have the best callers invite the identified Ron Paul supporters to the straw poll.


Today's Sample

Today I called 2,134 registered Iowa voters who met the following criteria:

1. Registered to vote in Iowa as independent (Non-Partisan).
2. Male
3. Will be younger than 40 years hold on August 13.
4. The phone number provided in the voter database is a cell phone number as determined by phonevalidator.com.
5. Primary ZIP code is 50010, 50012, 50013, 50011, 50201, 50105 or 50014.

All of the ZIP codes are within 9 miles driving distance of the Hilton Coliseum, as determined by zip-codes.com (see http://www.ronpaulforums.com/showth...-by-zip-code&p=3367532&viewfull=1#post3367532). I would normally have a bottom age limit to exclude the flakier end of Ron Paul's younger supporters, but I figure if they live that close to Ames, the teenagers are worth a try.


The question

The exact text of the poll question was as follows:

Good evening. Please tell me which candidate you currently support for president in 2012. Enter: One for Barack Obama. Two for Mitt Romney. Three for Michele Bachmann. Four for Herman Cain. Five for Newt Gingrich. Six for Tim Paw-lenty. Seven for Rick Perry. Eight for Rick Santorum. Nine for Ron Paul. Or zero for other or undecided. Once again, please enter: One for Barack Obama. Two for Mitt Romney. Three for Michele Bachmann. Four for Herman Cain. Five for Newt Gingrich. Six for Tim Paw-lenty. Seven for Rick Perry. Eight for Rick Santorum. Nine for Ron Paul. Or zero for other or undecided.

(The hyphen in Pawlenty's name was necessary to get the TTS system to pronounce his name correctly.) The question was designed so that calls would remain just under 1 minute (CallFire charges in 1 minute increments).


Results

2,134 phone numbers were dialed between 7:30pm and 8:15pm Central time.

1,959 calls were answered, which at $0.05 per minute, cost just under $100.

There were 141 responses to the poll question as follows:

Barack Obama: 52 (37%)
Ron Paul: 26 (18%)
Michele Bachmann: 16 (11%)
Mitt Romney: 9 (6%)
Rick Perry: 5 (4%)
Tim Pawlenty: 4 (3%)
Rick Santorum: 2 (1%)
Herman Cain: 1 (< 1%)
Newt Gingrich: 1 (< 1%)
Other/Undecided: 25 (18%)

So I identified 26 Ron Paul supporters in the Ames area at a rate of about $4 each. Considering that phonebanking is identifying Ron Paul supporters at a rate of about 1 out of 100 manual calls made in about an hour, I consider this very efficient.
 
Now you are talking. Do you know how many target participants there are total? Or was the 2,134 number the population size?

Are you going to expand your search radius? 9 milies is not that far to drive. I would think that maybe you could stretch it to 20 miles.
 
Last edited:
Now you are talking. Do you know how many target participants there are total? Or was the 2,134 number the population size?

That was the population size meeting the criteria. The next robopoll, if there is one, will likely have to reach into Des Moines (at least 30 miles driving distance from Ames) for the same demographic.
 
I'm going through my 46 Skype voicemails now. Funniest so far:

Hello, my Chinese wife received a call from you and didn't understand what it was about. Please give me a call back on XXX-XXX-XXXX.

So far it seems Iowans are a lot more polite than Floridians.
 
Last edited:
Next time, put Ron Paul in the middle of that lineup, say, at #5, and let's see if the results come out similar.
 
Any reason to not use Google Voice? It's free calling with a customizable phone number when first signing up. How are you scripting? I might be able to help scripting in the browser.
 
Any reason to not use Google Voice? It's free calling with a customizable phone number when first signing up. How are you scripting? I might be able to help scripting in the browser.

Google Voice is fine for manual phonebanking, but do you have any idea how much bandwidth and CPU power is required to make 2,134 phone calls in well under an hour? The programming (detecting keybad responses, distinguishing between human and answering machine, etc) would be extremely challenging, which is why existing software (like Voicent) costs thousands of dollars.
 
Any reason to not use Google Voice? It's free calling with a customizable phone number when first signing up. How are you scripting? I might be able to help scripting in the browser.


You can't integrate a robocall into Google Voice. (dont think)



To the O-P


I love your enthusiasm and its great to see folks taking initiative!

Let me make some constructive suggestions / ask some questions though.....

1. Folks utilizing the phone bank and not being paid, thus the cost for IDng a supporter is $0. Granted the big factor is that the amount of calls being made is probably not high enough to mobilize more than 1000 supporters (at THIS point, I assume the rate will steadily increase over time) So my question would be... at a cost of $4 per supporter as compared to $0 a supporter why is the robopoll a better option than 100% manual phonebanking?

2. I believe the campaign is targeting its registered support list. This is not necessarily a list of just R's it will certainly include I's that support Ron as well. However, i'm not 100% sure on this point.

3. Do you plan to use the robopoll in conjunction with a manual GOTV operation? Stated Differently, will you use the robopoll to ID support and then utilize manual callers for the GOTV operation?

4. At a cost of $4 per supporter why should i support this venture over the campaigns?


Again, great work and enthusiasm.
 
Last edited:
A CPU/bandwidth issue can be addressed with multiple users/computers. Both keypress and voice responses could be processed manually and distributedly assuming no easily accessible automated system. E.g., something on rp2012.org to listen to responses and identify the valid/ron paul supporters.
 
You can't integrate a robocall into Google Voice. (dont think)

I've seen discussions of developers trying to do this, but it is far from complete, and isn't necessary either. Voicent + CallCentric does the job if you have the money and hardware.

Let me make some constructive suggestions / ask some questions though.....

1. Folks utilizing the phone bank and not being paid, thus the cost for IDng a supporter is $0. Granted the big factor is that the amount of calls being made is probably not high enough to mobilize more than 1000 supporters (at THIS point, I assume the rate will steadily increase over time) So my question would be... at a cost of $4 per supporter as compared to $0 a supporter why is the robopoll a better option than 100% manual phonebanking?

Besides speed, I don't have any more reasons to prefer robopolling over phonebanking than the three I gave in the OP.

2. I believe the campaign is targeting its registered support list. This is not necessarily a list of just R's it will certainly include I's that support Ron as well. However, i'm not 100% sure on this point.

That's a good point. Steve Bierfeldt asked me to call him tomorrow, so I'll see what he says about this. But unless the Ron Paul supporters are so hardcore that they read this forum, they won't know the robopoll was on behalf of Ron Paul, and therefore there won't be any blowback from the overlap.

3. Do you plan to use the robopoll in conjunction with a manual GOTV operation? Stated Differently, will you use the robopoll to ID support and then utilize manual callers for the GOTV operation?

This depends on whether the official campaign accepts my robopoll results. If they don't, then I will simply give them to Jun Dam, who I'm sure will have his phonebanking volunteers make GOTV calls. But I have no control over what anyone does with the results.

4. At a cost of $4 per supporter why should i support this venture over the campaigns?

If you're not maxed out to the official campaign yet, then donate to it first.

If you have the time to phonebank, then volunteer with the official campaign.

If you are maxed out, the official campaign isn't giving you any phonebanking work yet, and your time is worth less than the Federal minimum wage (which was $4.25/hr when I moved to Australia in 2003), then go ahead and phonebank for rp2012.org.

Otherwise, donate to these robopolls.
 
A CPU/bandwidth issue can be addressed with multiple users/computers. Both keypress and voice responses could be processed manually and distributedly assuming no easily accessible automated system. E.g., something on rp2012.org to listen to responses and identify the valid/ron paul supporters.

If you're volunteering to develop this, then by all means, go right ahead.
 
what about robo-polling two questions for correlation purposes. for example, you ask callers to pick who they are going to vote for THEN ask what issue is going to determine your vote or what issue is most important to them...or something like that. along the way you find paul voters but you also get a sense of what issues are important to voters of other candidates. thoughts? i'd be curious to know why people are supporting Cain...then Ron could start emphasizing different issues when he talks in an effort to 'steal' voters..everyone is doing it TO Ron..OR you could mail out the corresponding slim-jim to that person...I wouldn't mind helping develop that...it would be rather simple..
 
Last edited:
Nice work. When selecting future sample sizes, will exclusionary criteria include that the number has not been dialed before?
 
what about robo-polling two questions for correlation purposes. for example, you ask callers to pick who they are going to vote for THEN ask what issue is going to determine your vote or what issue is most important to them...or something like that. along the way you find paul voters but you also get a sense of what issues are important to voters of other candidates. thoughts? i'd be curious to know why people are supporting Cain...then Ron could start emphasizing different issues when he talks in an effort to 'steal' voters..everyone is doing it TO Ron..OR you could mail out the corresponding slim-jim to that person...I wouldn't mind helping develop that...it would be rather simple..

This sounds like Fritz Wenzel's job. It would also more than double the cost of the poll. We're just trying to find existing Ron Paul supporters for the Ames straw poll.
 
That 18% who stayed on through the survey and selected Ron are exactly the kind of people that would be likely to caucus Ames. Awesome work!
 
Back
Top