This study analyzed the impact of Indiana's voter photo ID law on voter turnout. It found that voter turnout decreased between 2002 (before the law) and 2014 (after the law) at both the registered voter and voting age population levels. A multivariate regression analysis also indicated that the voter ID law requirement in 2014 had a statistically significant negative relationship with turnout compared to 2002, while factors like higher unemployment, percentages of younger/older voters, and rental housing were also associated with lower turnout. However, the author notes that more research is needed to fully understand the variables affecting turnout.
1. The Indiana Voter ID Law
A County-Level Analysis of Impact on
Voting Age Population Voter Turnout
Stephanie Brinkerhoff-Riley
May 4, 2016
2. Before the Voter ID Law
• Voters must be registered to vote
• To vote in person, a voter was required to
state their name and sign a voter log
• Voter identification was verified if necessary
by comparing signatures on the voter log
and voter registration card
• Laws were passed at federal level to increase
voter protections and registration
• VRA (1965); NVRA (1993)
• States expanded early voting, same day
registration, online and absentee voting
• Voter turnout in the Presidential election
year of 1996 was 63% and 2000 was 56%
• Voter turnout in the non-Presidential year of
1998 was 44% and 2002 was 39%
• Voter turnout trending down despite efforts
4. The Indiana Voter Photo ID Law
• First of its kind in the country
• Took effect for 2006 elections
• Requires unexpired government-issued photo ID to vote in person
• ID must largely conform to the name listed in the voter registration log
• Stated purpose is to reduce in person voter fraud
• Democrats sued and US Supreme Court affirmed the law with the
willingness to hear additional issues if a discriminatory application could be
shown (2008)
• 33 states have followed with similar laws
• Research on impact on voter turnout, specific groups of voters and voter
fraud has been contradictory
• Disagreement exists on the appropriate data and regression model
• Impact is small (1-2%) which complicates chosen methodology
• 2002 and 2014 allows for a before and after comparison: no statewide
federal races or ballot initiatives
5. Conflict Theory: social systems generate conflict as a pervasive feature of
society as opposing interests compete for control of society (government)
Research Question: Do voter photo identification requirements depress voter
turnout?
Type of Research Question: Difference
Hypothesis: Voter photo identification requirements depress voter turnout
IV= photo identification requirement in place starting in 2006
DV= voter turnout, calculated as a percentage of the voting age population
(2002; 2014)
Projected Causal Relationship: negative
• Voting age population (VAP) data from US Census Bureau (2002;2014)
• Registered Voters (RV) data from Indiana Election Division
• Voter turnout data for each of the 92 counties from Indiana Election Division
• County-Level Data (92) from US Census Bureau; IU Kelley School of
Business
Independent t-Test: Difference in Mean turnout of RV and VAP (2002;2014)
Multivariate Linear Regression: DV= Turnout VAP (2002;2014)
Independent Dummy Variable (2014=1; 2002=0)
IV= Percent Poverty; Minorities; Registered Voters; College Degrees; No HS
Diplomas; Unemployment; Rental Housing; VAP 18-24; VAP 65+
6. Results: Independent Sample t-Tests
Registered Voter Turnout Voting Age Population Turnout
Mean SD Min Max Mean SD Min Max
2002
.43 .064 .24 .56 .39 .072 .23 .60
2014
.35 .063 .24 .49 .32 .066 .21 .50
t=8.280 p=.000 t=6.674 p=.000
7. Results: Multivariate Linear Regression
Model Adjusted R Square 0.593
Turnout of VAP (DV) Signif. Std. Error B Beta
2014 (compared to 2002) (IV) (Dummy) .000 .017 -.096 -.629
Percent of VAP 18-24 (IV) .908 .161 -.019 -.010
Percent of VAP 65+ (IV) .004 .232 .678 .252
Percent of VAP Registered (IV) .000 .051 .393 .403
Percent of VAP 25+ No HS Diploma (IV) .557 .140 -.082 -.051
Percent of VAP College Degree (IV) .182 .207 -.277 -.136
Percent of Population in Poverty (IV) .016 .214 .522 .263
Percent of Population Minority (IV) .544 .114 -.070 -.048
Unemployment Rate (IV) .000 .005 -.020 -.275
Percent of Housing as Rental Units (IV) .000 .121 -.461 -.326
8. Implications:
• Clearly a decrease in voter turnout in Indiana (RV and VAP)
• Voter Photo ID Law may explain some of the variation in turnout
• More research must be done to find and understand the variables
missing from the regression model
• Voter turnout was trending down before the law was passed
Questions?
Editor's Notes
My project was to look at the impact in Indiana on voter turnout from the Indiana Voter ID Law
Series of events which focused the country on archaic election systems and the potential for election fraud
Bush v Gore
HAVA
Lake County
Carter-Baker Report
Appearance of voter fraud can be just as damaging or more so to the election process and people’s confidence in the system
First of its kind
Requirements
S Ct affirmed
33 states followed suit
Research on impact contradictory
The voting age population, those 18 years old and above, for each county is calculated and reported by the U.S. Census Bureau as of July 1st of each election year. The Census Bureau's Population Estimates Program utilizes current data on births, deaths, and migration to calculate population change since the most recent decennial census and produces annual estimates regarding changes in population and demographics.
County level control variable data was also collected by the U.S. Census Bureau through the statistical data utility for the State of Indiana, which was developed and is maintained by Indiana Business Research Center at Indiana University’s Kelley School of Business
The t-tests demonstrate with a more than 99% confidence level the difference between 2002 and 2014 in both the turnout of the voting age population and the turnout of registered voters is statistically significant.
The difference in means for turnout for the voting age population and registered voters is also substantively significant.
VAP drop was 7%.
RV drop was 8%.
The model explains 59.3% of the variation in voter turnout between 2002 and 2014.
There is a confidence level of more than 99% that the unemployment rate, the amount of housing which are rental units, the number of people registered to vote, the size of the voting age population which are 65 years or older and the 2005 voter photo identification law had a statistically significant impact on voter turnout.
There is a confidence level of more than 98% that the size of the population living in poverty had a statistically significant impact on voter turnout.
B Values impact on Voter Turnout
Beta for Dummy Variable
Beta for Registration and Rental Housing
Explanation for Poverty and Rental Housing
2014: 1% of the voter turnout was 13,884 and .1% was 1,388
.678 = 9,413