Voting: A Data Story

Please note: This story is best viewed on a desktop browser.

Introduction

In the United States, voting is a core component of how we think of “democracy.” At its most basic level, democracy is giving people a say in the decisions that affect them. Voting is just one way people have a say in decisions that affect them.

This data story provides information to organizations and municipalities registering and engaging voters in the Hartford region. The questions we explore include:

• Where is there currently a high number of non-registered, potential voters in the Hartford region? 

• Where were the highest turnout and lowest turnout in the 2016 election in the Hartford region? 

• Where are young voters ages 18-29 more active and less active in the Hartford region?

Each of these questions offers insight into where there needs to be the most focus on voter registration and education in the Hartford region.

For this story, we started by analyzing the Connecticut Secretary of State Voter Registration data. This dataset includes registered voters as of July 2020, and the Citizen Voting Age Population data (2018 5-year U.S. Census American Community Survey estimates). We also used the official Voter Turnout Data from the Connecticut Secretary of State and the Connecticut Department of Motor Vehicle (DMV) data. This dataset shows the number of applicants for DMV products that did not register to vote. (See the Appendix for more information on these datasets, including their limitations.)

Find additional resources to help guide your organizational planning about on-the-ground voter registration and engagement activities with our Voter Maps.

This project is brought to you with the generous support of the Hartford Foundation for Public Giving.

Not-Yet-Registered Voters

Where is there currently a high number of non-registered, potential voters in the Hartford region?

• By looking at the raw numbers of estimated eligible, not-yet-registered voters, the following municipalities have the highest estimated numbers: Hartford (14,126), Enfield (9,283), and Manchester (6,079).

• Some of Connecticut’s smaller and more rural towns have a higher percentage of eligible, not-yet-registered voters than the urban areas, including Somers (31%), Enfield (27%), and Andover (24%). Some of this is likely accounted for by correctional facilities in some of these towns.

• Several towns have an estimated 0% of eligible not-yet-registered voters, which serves as a reminder to use these numbers to suggest trends rather than for reporting on absolute numbers.

• With the DMV data, we find a total of 13,000 voters eligible in the Hartford region that have not yet registered to vote. Hartford has the most number of eligible not-yet-registered voters through the DMV at 3,014, followed by Manchester (1,114) and East Hartford (1,055).

According to CVAP, an estimated 14,000 eligible not-yet-registered voters live in Hartford but 3,000 according to the DMV data. Why do we see a higher number of non-registered voters in CVAP than in the DMV data? 

• Not everyone obtains a driver’s license or photo ID. According to a 2015 analysis of American Elections Study (2012) data by Project Vote, approximately 7% of Americans did not have photo identification at that time.

• Some Connecticut drivers have not needed to renew their licenses since the beginning of this program. 

• The CVAP estimate only accounts for voting-age citizens. It cannot account for other factors, such as criminal justice system involvement, that would make a person ineligible to vote.

This information suggests that the scale of the work of registering voters may be larger than would be expected in some of Connecticut’s more rural towns.

Explore these data further with the Not-Yet-Registered cross-tab in the Appendix.

Voter Participation in the 2016 Presidential Election

Where were the highest turnout and lowest turnout in the 2016 election? Where should there be the most focus on voter education?

• 2016 voter turnout in the Hartford region was highest in Somers (95%), Andover, and Bolton (both 88%), and lowest in Hartford (63%) and East Hartford (67%).

• Towns with the highest percentage of currently registered residents who voted in Connecticut in the 2016 election include Andover (81%), Bolton (78%), and Wethersfield (76%). Hartford has the lowest percentage of current residents who voted in Connecticut in 2016 (45%), followed by East Hartford (57%), and East Windsor (59%), respectively.

While towns in the Greater Hartford region aim for 100% voter turnout in 2020, the 2016 voter turnout data provided by the Connecticut Office of Secretary of the State can be used to set benchmarks for the 2020 election. 

We can look at voter turnout in two ways with the datasets we are using. 

• First, we can look at the official number of registered voters who voted in the 2016 presidential election - the official voter turnout rate.

• Second, we can look at the percentage of currently registered residents that voted in Connecticut in the 2016 presidential election to get a sense of participation of currently registered residents in this upcoming election. 

In the 2016 election, Somers had the highest voter turnout among registered voters in the region at 95%, followed by Andover and Bolton (both at 88%). Andover has the highest percentage of currently registered residents who also voted in Connecticut the 2016 election at 81%, followed by Bolton (78%) and Wethersfield (76%).

Towns with the lowest voter turnout in 2016 also have the lowest percentage of currently registered residents who voted in Connecticut in the 2016 election. These include Hartford (63% 2016 turnout, 45% current residents), East Hartford (67% 2016 turnout, 57% current residents), and East Windsor (72% 2016 turnout, 59% current residents), respectively.

Note: There is natural churn in the registration of voters -- aging in, moving out, and passing away--resulting in the number of currently registered residents being lower than official turnout. 

The gap between the official 2016 turnout and the participation in the 2016 election of currently registered residents who lived in Connecticut at the time gives a sense of the amount of work required for voter registration and education activities. A town with a much lower percentage of current residents who voted in Connecticut suggests more residents need information about where to vote, voting requirements, and information about local races.

Areas of high opportunity to increase it for 2020: Let’s get to 100%

• Hartford→ 18% gap in voter participation and of currently registered residents voted in Connecticut in 2016

• Somers → 20%  gap in voter participation and of currently registered residents voted in Connecticut in 2016

• Vernon → 15%  gap in voter participation and of currently registered residents voted in Connecticut in 2016

Explore these data further with the 2016 Turnout Compared to Currently Registered Residents cross-tab in the Appendix.

Young Voter Registration and Participation

Where are young voters ages 18-29 more active and less active in the Hartford region?

• In the Greater Hartford region, individuals aged 18-29 experience the lowest voter registration rates. Hartford, East Hartford, Vernon, Enfield, and Manchester have the highest number of eligible, not-yet-registered young voters.

• The Hartford region’s municipalities with the highest percentage of eligible not-yet-registered voters aged 18-29 are Hartford, Vernon, and Enfield.

As organizations and town officials are working to ensure that each eligible constituent is registered to vote, many are also working to encourage registered voters to vote. We can look at previous voting behavior of those currently registered to get a sense of the scale of work required to get every registered voter out to the polls on November 3. 

According to the Census Bureau, Individuals between the ages of 18 and 29 have historically experienced lower voting registration rates than other age brackets (explore the raw data). The same trend shows up when analyzing 2020 registration rates in the Greater Hartford region. 

Municipalities with the highest percentage of eligible not-yet-registered voters aged 18-29 are Hartford (49%), Vernon (45%), and Enfield (45%), while municipalities with the highest number of eligible not-yet-registered voters aged 18-29 are Hartford (14,509), Manchester (4,246), and East Hartford (3,452). 

See how your town is doing below. Click your town name in the table to see how many of your town’s currently registered voters aged 18-29 voted in 2016, and how many potentially eligible young people have yet to register.

Explore these data further with the Registration and Voting Ages 18-29 cross-tab in the Appendix.

Conclusion

While this data analysis cannot give exact knowledge of how voters will act in 2020, it provides insight into where to focus efforts to increase voter participation in your town. This analysis shows that there is still much work needed to reach our goal of 100% eligible voter registration and 100% registered voter participation in Connecticut.

To explore more of the voting data go to the Get Out the Vote Data Project. You can also download and print or share this data story as a pdf. Be sure to sign up for our CTData newsletters to receive data news and updates, or follow us on Facebook, Twitter, or LinkedIn. 

This project is brought to you with the generous support of the Hartford Foundation for Public Giving.

Appendix

Background 

This project aims to provide practical information for voter registration and education efforts to organizations and municipalities in the Hartford region. The work was made possible with the generous support of the Hartford Foundation for Public Giving.

Data Limitations

As we note throughout this data story, each of the datasets we use comes with their particular limitations. Here is a more detailed description of each dataset and its limitations.

All data should be used to identify general areas to target voter registration and education efforts rather than for reporting on actual numbers. It can help identify areas for increased registration and turnout but does not necessarily provide exact counts.

Secretary of State Voter Registration Data. This dataset, created by CTData, is a point-in-time dataset of currently registered voters on June 30, 2020. This dataset allows us to look at ages of currently registered voters and what previous Connecticut elections they voted in - but not the town in which they voted. The Hartford Courant recently analyzed voter registration data and found that there has been a significant increase in voter registration between June 15 to Sept 15. This dataset includes only some of these registrations.

When using this dataset to look at past voter behavior, there are a few critical points to keep in mind: 

• When categorizing the dataset by age groupings, the youngest age group (in our exploration, ages 18-29) will always have lower past voting participation because only some of this age group was eligible in the 2016 presidential election.

• The number and percentage of currently registered residents who voted in the 2016 election will always be lower than the official 2016 presidential election turnout because of state out-migration (people who voted in 2016 but have since moved out-of-state) and because of people who have died since voting in 2016.

Citizen Voting Age Population (CVAP). This dataset, a U.S. Census Bureau project, contains estimates of citizens of voting age and has margins of error. This estimate is calculated using the 5-year American Community Survey estimates. For this analysis, we use the 2014-2018 data set.

Census population estimates count people in jails and prisons in the town where they are incarcerated at the time of the count. Some of these individuals are eligible to vote, but most are not. The Census Bureau cannot accurately estimate the number of ineligible voters in any town, so these individuals are included in the eligible population. 

These estimates also result in some towns having an estimated 100% of eligible voters registered. However, the DMV file shows this not to be the case. 

See more technical documentation here. 

Connecticut Department of Motor Vehicles (DMV) Eligible, Not-Yet-Registered Voters. In 2016 the DMV began voter registration of applicants for new and renewed licenses or photo identification unless applicants opted out. The DMV shares data they have collected since the beginning of this program with the Secretary of State’s Office to identify eligible, not-yet-registered voters. By comparing the number of non-applicants with the voter registration file, the Secretary of State’s Office identifies the voters that are not in the Voter File, meaning they are not registered. This number is significantly different from the estimate of eligible, not-yet-registered voters using the CVAP data set.

There are several explanations for why the number of eligible, not-yet-registered voters (CVAP) is different when comparing these two datasets. First, not everyone obtains a driver’s license or photo ID. According to a 2015 analysis of American Elections Study (2012) data by Project Vote, approximately 7% of Americans did not have photo identification at that time. Additionally, some Connecticut drivers have not needed to renew their licenses since the beginning of this voter registration program. Lastly, the CVAP estimate only accounts for citizens of voting age and cannot adjust the estimates based on criminal justice system involvement.

2016 Election Results. These data, available through the Secretary of State’s Elections Management System data portal, show the number of registered residents who voted in person as a percentage of the number of registered residents in the town to determine the voter turnout rate. The number of absentee ballots mailed out and cast are not counted in the turnout calculation.

Town Population Estimates. The dataset used to identify the population ages 18-29 in each town was created using the American Community Survey 5-year population estimate (2014-2018) and includes both citizens and non-citizens. This is different than the Citizen of Voting Age dataset, which does not disaggregate by age. We analyzed the data to create the unique age groups we used for this exploration.

Data Tables