Author’s note: To see how we analyzed our data and to understand our voter model, please refer to the notes at the end of this article.
Municipal election campaigns are gearing up for November. Because El Paso is a Democratic Party stronghold, the March primaries determined the outcome of the state and national elections in favor of the Democrats. However, the city council races are in contention. Both the city representative incumbents and their challengers are working to identify who will likely cast a vote in November. Once identified, the campaigns will look to get-out-the-vote from the El Paso voters likely to vote in the upcoming elections. But who are the El Paso voters? Campaigns use voter modeling to predict who is likely to vote in November so that they can encourage the voter to cast a ballot in favor of the campaign’s candidate.
To identify the El Paso likely voter, El Paso News analyzed the voting patterns of over half a million El Paso voters over the last 20 years to identify voting patterns and attempt to identify the likely voter in the upcoming elections. Here is what we found.
There are 500,190 El Paso voters in our data set. These are registered voters through June 10, 2022. The County of El Paso Elections Department reported 502,416 voters in El Paso yesterday. That is 2,226 new voters that our data does not include but for our modeling purposes that does not matter. The reason is that our modeling is based on voters’ historical voting patterns in addition to other data that we will explain. The last election was on May 24 where 24,699 of the 497,943 registered voters cast a vote in the May 2022 Primary Runoff Election. New voters who have registered after June 10 have not had an opportunity to cast a vote, thus, these are new voters. The El Paso Democratic Party and some political observers have argued that Beto O’Rourke will drive new voters to the polls in November. As part of our analysis, we will look at a Beto Factor in the upcoming elections. But first it is important to understand what voter modeling is and what campaigns can glean from it.
Voter Modeling
Modeling whether a voter is likely to vote have the inherent biases of those doing the modeling. Modeling are assumptions on why someone votes and the factors that drive voters to the polls. Polls have used the term “likely voter” as someone that is likely to cast a vote but how a voter is identified as a “likely voter” varies between campaigns and pollsters alike. [1] At its most basic, a likely voter is someone that has a history of voting in past elections. But what makes this methodology complicated is that voters vote for different reasons that are difficult to quantify. For example, Texas Democrats are under the belief that Beto O’Rourke’s candidacy is likely to make Texas blue in November. Whether this is a talking point, or an actual belief is unknown, but the Democrats have the inherent bias that they should be in control of Texas’ politics. Thus, any modeling by Democrats must be weighed against the bias of a Democrat renaissance in Texas led by O’Rourke. Another inherent bias is the oft-cited belief that women that are 65 and older is a reliable voter.
It is well documented that voters aged 65 to 74 are the most likely voters nationally, but as readers will learn, does not apply to El Paso voters. Several factors from retirement to fixed incomes are ascribed to this age group suggesting that their voter participation is a factor of the time to follow politics and on issues like taxes and access to benefits that drives them to the polls. The U.S. Census reports that 76% of those between 65 and 74 are likely voters.
Likewise, women voters are more likely to vote is often a cited figure among political strategists. But election figures from the 1980’s to today show that although women slightly outperform men on Election Day, the difference is marginal at best. In 1980, 70% of the votes cast were from women compared to 69.3% of men. In 2020, women were slightly ahead at 74.1% compared to 71.2% of the men. Although the argument can be made that women outperform men in American elections, the historical difference between the two is negligible. Nonetheless, some local political consultants believe that the abortion issue has given the two women incumbents an edge.
In addition to inherent biases and gender, there is also the issue of voter fluidity. The fluid voter is the voter that is eligible to vote but because of school, family or job commitments isn’t available to vote on Election Day. This is most prevalent in the younger voter who leaves the city to attend school and is unable to vote but is registered to vote. Each election cycle, younger voters are touted as a voter to make a difference, but data suggests that is not the case. In the case of El Paso there is also the military factor in the voting demographics. The Army is fluid in that it moves military personnel regularly between jurisdictions.
In the case of polling, much of their “likely voter” identity is based on the unscientific method of asking someone answering in a survey if they intend to vote in the upcoming election. Voters, when faced with answering the question of whether they vote will seldom be honest and answer truthfully that they do not vote. [2]
Modeling a voter does not guarantee a campaign’s victory because an unliked candidate is unlikely to turn a voter that does not like them into a vote regardless of the number of times the candidate touches the voter with messaging. However, voter modeling does allow a campaign to efficiently target voters with messaging that can lead a voter to cast a vote in favor of the candidate. In the case of El Paso, for example, instead of targeting the 63,355 registered voters in District 5, the campaigns of the incumbent Isabel Salcido and her two challengers Richard Genera and Felix Muñoz, could instead focus their efforts on the 40,747 voters that have cast at least one vote throughout their voting history. Salcido won her seat in 2018 with 10,917 votes out of 19,243 ballots cast. Taking that into account, trying to reach 40,000 voters is expensive.
Our modeling suggests that although about 40,000 voters in District 5 have cast at least one vote in their voting history, they are not the likely voters in November. Our research shows that the 40,000 voters, almost all, about 80% of them, have cast votes in only about 10% of the times they are eligible to vote on. It is therefore more practical for the candidates to focus their get-out-the-vote efforts on the 20% of the voters that are likely to vote on Election Day.
But who are those voters?
The El Paso Voter
Now that voter modeling techniques are better understood we can now apply them to the El Paso voter. We took our 20 years of voter data and modeled El Paso’s likely voter. Strictly on voting records, we found that of the 500,190 currently registered voters, only 320,620 have cast at least one ballot since they registered to vote in El Paso. This shows that 64% of currently registered voters in El Paso have cast a vote. It also shows that 179,570, or 36% El Paso voters have never voted in an election.
Based on this figure, any campaign can quickly discard 36% of the voters that will not vote for them.
But that leaves the question of who are the power voters in El Paso? Who are the voters that the candidates want to target? Our data found that there are 288 “super” voters in El Paso. The super voter is one that has never missed an election, are engaged politically and are motivated to vote each election cycle. (See note 2 below for our methodology) They represent less than one percent of El Paso’s voters who have cast at least one ballot.
Most voters who have voted at least once account for about 80% of the voters in our data that have cast a ballot. This suggests that El Paso voters are just not motivated to vote. Only about 50,000 of El Paso voters cast ballots about 20% of the time they are eligible to vote.
The analysis of our data shows that there are about 5,000 El Paso voters who consistently vote over the lifetime of their voter registration. These are the voters that have cast a ballot over 50% of the time they are eligible to vote. Here is where things get interesting when attempting to model the El Paso voter.
Clearly it takes more than 5,000 votes for most El Paso elections. Since this is true, what does this suggest for the 5,000 El Paso voters we identified? Specifically, where do the rest of the voters come from?
Let us dig deeper into our data to see if we can better understand the El Paso voter.
The first question is whether El Paso women out vote their male counterparts. The answer is yes, by over 10%. Over 50% (55.33%) of women voters in El Paso have cast at least one vote compared to 44.67% of the men. In comparison, El Paso women voters outperform their national counterparts.
Another question is where El Paso’s active voters live. Our analysis shows that the majority of El Paso’s voters live in the 79936 zip code. The next largest concentration of El Paso voters lives in the 79912 zip code. Excluding the Fort Bliss and Tornillo zip codes, the least active El Paso voters live in the 79901 zip code. However, readers should be cautious about ascribing too much to the zip code of the voter because our analysis has revealed that several voters use addresses other then their home addresses on their voter registrations. We found several voters using business addresses in the voter rolls.
Another voter demographic often cited by political consultants is the age group of voters. It is common to hear a political consultant argue that voters 65 to 74 are the likely voters. But does this metric hold up to our analysis? We found that voters between the ages of 45 and 64 at 33% is the age group in El Paso that has cast at least one vote. This is followed by the age group between 30 and 44 who, at 25%, is the next age group of El Paso voters. Those between 18 and 29 came in third at 16%. Those over 65 were the last at 14%.
How does this compare to the national voter? According to a study by the U.S. Census Bureau, the 2016 national voter in the 18-29 age group accounted for 15.7% of the voters. Those in the age group of 30-44 were 22.5% of the vote. The largest voting age group nationally in 2016 were those aged between 45-64 at 37.6%. And those aged over 65 represented 24.2% of those who cast ballots in the 2016 elections nationally. [3]
El Pasoans aged between 30 and 44 years old mirrored the national voter at 25% to the national average of 22.5%. The younger El Paso voters, those aged 18 to 29, likewise mirrored their national counterparts at 16.1% to 15.7%. Those in the age group between 45 and 64 were slightly further apart with El Paso voters accounting for 33% of the voters compared to the national average of 37.6%. However, the El Paso voters aged over 65 years old was much lower at 15% then their national counterparts in the 2016 elections who came in at 24.2%. However, there is a discrepancy in the age groups of El Paso voters because the County’s original data has many invalid birthdates in their voter rolls. Thus, as many as about 11% of El Paso voters’ age data is not available.
Nonetheless, those in El Paso who are between 45 and 64 years old are El Paso’s likely voters. They have cast at least one ballot in their voting history. But one ballot years ago does not translate into a vote in November so analysts must further parse the voting records to arrive a targeted list of voters likely to vote in November.
To better target our likely voters, we need to analyze who voted in recent elections. We looked at the voters that cast a vote in the last election and those who cast a ballot in the last three elections in a row. These voters are more likely to cast a vote in November.
Our analysis reveals that there are 24,680 El Paso voters who cast a vote in the last three elections.
These are the voters that will likely be targeted by campaigns for the November elections.
But the November elections include state and national elections which for El Paso voters were settled in the March primaries but nonetheless affect the municipal candidates. To pierce through the election noise, down-ballot candidates in the city council races are analyzing trigger issues to see if they are a factor for the voters and if so, which issue can best be leveraged for their candidate. There are two state-wide issues that may impact the elections.
The Trigger Issues: Abortion And The Beto Factor
The issue of gender-based ballots is important to understand because much of the narrative about turning Texas blue is the belief that voters will be driven to the polls because women will turn out because of the abortion issue. Some political strategists believe that women will surge at the polls because of the abortion issue.
Others argue that the Democrats will also surge at the November polls because of the Beto Factor – the belief that the Beto O’Rourke’s campaign to unseat Greg Abbott will drive Democratic voters to the ballot boxes. Particularly in El Paso, the abortion issue and the Beto factor has several campaigns wondering if the two women city council incumbents have an edge in the election.
However, the underlining trigger issue for the city council races is taxes. How much of a factor it will play on voters voting in November is yet to be fully understood. Two schools of thought exist. One is that there are enough voters focused on the tax issue to select a municipal candidate while the other side argues that most voters will vote on the statewide ballot giving little to no thought on down ballot candidates giving the incumbents the advantage. Much of the difficulty lies in that the municipal elections are in November.
May Municipal Elections
In 2013, El Paso voters approved moving the city elections to November. The last time city council elections were scheduled in May was in 2017. In 2020, El Paso voters voted for city representatives in November. There were significant less voters turning out for the May elections which made predicting which voters were likely to turn out easier when municipal elections were held in November.
El Paso voters will once again be asked to vote to move city elections back to May in May 2023. Opponents to moving city elections back to May argue that the additional cost to the taxpayers for holding a May election and the low voter turnout is the reason to keep the city elections in November.
However, for political consultants, moving city elections to May make predicting who will vote for a city candidate much easier giving an edge to the campaigns who can micro-target voters more specifically. When voters vote in May, their attention is focused on the city’s candidates. When the voter votes in November, several other races play into who the potential voter will be and where their focus lies. Targeting voters for city council races in November is made more difficult because of other factors – like abortion and the Beto O’Rourke’s campaign – distracting the voters from the municipal elections. Elections in November have a much larger turnout making targeting voters more difficult and expensive.
What It Means For City Council District Races
The most important question remains, who are the likely voters that will vote for the city council representatives for districts 1, 5, 6 and 8? Our analysis found the likely voters in these districts to be as follows:
District 1: 47,136 voters
District 5: 40,747 voters
District 6: 31,024 voters
District 8: 27,602 voters
But these voters are voters who have cast at least one ballot in their voting history. For most campaigns targeting them for their vote is expensive and unlikely to yield results for the campaign. A better targeting is to further parse the likely voter into a more manageable list by selecting those that will likely cast a vote in November because they have voted in the last three elections in a row. This is what we found:
District 1 Power Voters: 4,824
District 5 Power Voters: 1,712
District 6 Power Voters: 2,062
District 8: Power Voters 2,550
Clearly these are not enough voters for the campaigns to win with even if they can convince every voter to vote in their favor. However, the analysis demonstrates how difficult it is for the municipal candidates to target their voters in the November election. To address this problem, campaigns will seek to further target their voters through the trigger issues and the use of other data like social media, socio-economic and demographics to mobilize the voters in November.
We have made our raw voter data, voter records and voter modeling publicly available at El Paso Votes for everyone, including campaigns, to use for free. To analyze El Paso voters or create your own voter models just create a free account on the App.
Support Independent Journalism
Choose an amount
Your contribution is appreciated.
DonateVoter Modeling Notes:
- El Paso News analyzed the voting records of 500,190 El Paso voters that are currently registered to vote in November. Our model does not include the 2,226 new voters who have registered since June 10. Our analysis of ballots cast include all voters who cast at least one ballot since November 3, 1998. Our public version of our voter data includes voter data going back to January 24, 2004. Our voter modeling holds true for the public version of the data as well as the additional four years we included in our analysis for this article.
- El Paso News used a proprietary algorithm that analyzed voting patterns over 20 years and then applied additional data points to include social media conversations, petition signatures, board assignments and participation, socio-economic metrics and political engagement through political campaigns and campaign contributions to political candidates.
Footnotes:
- Joshua Blank, “Just What is a ‘Likely Voter’?,” Texas Tribune, March 6, 2012, https://www.texastribune.org/2012/03/06/just-what-likely-voter/.
- Blank, “Just What is a ‘Likely Voter’.”
- Thom file, “Characteristics of Voters in the Presidential Election of 2016,” U.S. Department of Commerce, Economic and Statistics Administration, U.S. Census Bureau, September 2018.
I am curious if your tabulation accounts for only the elections in which a voter is eligible to vote – for instance, runoffs for districts or cities that the voter doesn’t live in. I reviewed my data, as presented, and while it clearly shows I didn’t vote in several elections, I was not eligible for those particular races. Does your analysis account for that?
Thank you for your question. The tabulation EXCLUDES the elections that voters were not eligible to vote in, both by the date their registration becomes effective and by jurisdiction. The screen you are referring to needs a little more cleaning up. There is a better screen intended for the campaigns but available to everyone. Log in to the dashboard and look for “voter tools” at the bottom. Select “look up voters” and the results will display a cleaner score. My App is a work in progress so I still need to clean up some screens and make other functionality available. Sorry for the inconvenience.