About a month ago, I conducted a poll for the upcoming RPV State Convention. Many people helped promote the poll. Sadly, I have been quite swamped recently, and haven’t been able to sit down and compile a thorough analysis of all the data. In light of that fact, I am going to open source the data. Below is a summary of information about the poll, and it’s strengths and weaknesses, and then the datasets. Feel free to use and analyze the data. If you post analysis of it, please link back to this bog. Also there were some decent costs associated with this, so if you appreciate the data or other articles I have written, a donation would be greatly appreciated.
2013 Republican Convention Poll
Date Range: 3/7-3/22
845 Self-Identified Likely Voters
Format and Goals:
The following poll was conducted over the course of two weeks to gain information about voter preferences in the upcoming 2013 RPV State Convention held in Richmond, VA on May 18th to decide the Republican nominee for Lieutenant Governor and Attorney General. The poll was conducted using surveymonkey.com. It was posted on williedeutsch.com, reposted on bearingdrift.com and promoted via social media.
Since the Republican nomination will be determined on multiple ballots, this poll was created to see who is likely to pick up support as the balloting progresses, and who different candidates supporters are likely to support.
The poll begins with a screen question asking people if they will be a delegate at the upcoming state convention. It then asks the following questions.
- Rank the Lieutenant Governor candidates in the order of who you currently support.
- On a scale of 1-5, with 1 meaning there is no way you will vote for anyone else, and 5 meaning you are actively looking for someone else, how committed are you to your number one choice for LG?
- On a scale of 1-5, with 1 meaning there is no way you will vote for anyone else, and 5 meaning you are actively looking for someone else, how committed are you to your second choice for LG?
- Who do you support for Attorney General?
- What is your gender?
- Which category below includes your age?
- What unit are you from?
All candidates were listed in alphabetical order by last name on the poll.
While the poll is better at blocking spammers than most polls, it is not perfect. Someone could easily find a way to vote from different IP Addresses. (Full data including the IP address of every voter will be released if people would like to examine the data for anomalies.)
Another weakness is the geographical spread for this poll. While the respondents come from a wide number of units, there is nothing scientific about the spread. In a weighted vote convention, having delegates from the right units is more important than the sheer number of votes you have.
One criticism of this poll is that the candidates listed at the bottom of the poll will have a larger number of 6th and 7th place votes because people would be lazy, and simply leave the candidates listed at the bottom. While the last candidate on the list has a large number of 6th and 7th place votes, the voting patterns for the second to last candidate don’t correlate. As a result, this seems to be a very weak criticism.
The strength of this poll is the ability to get a feel for who candidates support on future ballots. This poll also has two checks against spamming. 1.) There is an IP Blocker which blocks multiple people from voting from the same IP address. 2.) The number of questions, and depth of the poll also acts to weed out poll spammers. Comprised of 845 people, the sample size is significant for a convention that will be made of almost ten thousand delegates. While this poll is not perfect, it provides a better look at the field than any other poll currently available.
Data Sets are Below: