ASA 2020 Fall Data Challenge: Voter Turn-out
Willamette University Statistics Team
1. Goal: Increase Voter Turn-out
This prior election illustrated how voter turnout matters. This year in Pennsylvania, Joe Biden won by only 47,578 votes. By understanding what factors impacted voter turnout in previous years, campaigns can better target people who are actually likely to vote rather than wasting both time and money. As such for our project, we decided to focus our efforts on figuring out what factors best predict that someone will vote.
2. The Data
We used the “IPUMS-ASA U.S. Voting Behaviors” dataset about voting behaviors in the U.S. from the Census Bureau and Bureau of Labor Statistics, provided by the IPUMS organization and curated by ASA.
3. Exploratory Data Analysis
3.1 What percent of people vote?
Interactive map made with plotly
Note that this is weighted with the survey weights provided.
3.2 How do people vote? What methods are most popular?
Bar graphs arranged in shape of United States with geofacet
3.3 What variables are associated with voter turn-out?
3.3.1 Education Level
Hypothesis: People with higher education levels are more likely to vote
3.3.2 Age
Hypothesis: Older people are more likely to vote
3.3.3 Marital Status
Hypothesis: People who are or have been married are more likely to vote
3.3.4 Veteran Status
Hypothesis: Veterans more likely to vote
3.3.5 Labor Force
Hypothesis: People who are employed are more likely to vote
3.3.6 Race
3.4 Why don’t people vote?
4. Analyses
4.1 Regression
Do states with higher rates of mail-in voting have higher rates voter turn-out?
Interactive graphic made with plotly
4.2 Weighted Chi-Squared Tests
4.2.1 Education
## No School Some school but no diploma High school graduate or GED
## Did not vote 0 17324670 44499002
## Voted 0 11995158 61551938
## Some college but no degree Associate degree Bachelors degree
## Did not vote 22710480 10336346 14258097
## Voted 50096529 28466024 67231104
## Masters degree Professional or Doctoral degree
## Did not vote 4034243 1281277
## Voted 29884543 10435339##
## Pearson's Chi-squared test
##
## data: tbl
## X-squared = NaN, df = 7, p-value = NA4.2.2 Race
## White Black More than one race Asian or Pacific Islander
## Did not vote 89352859 13803392 2624594 7073032
## Voted 211146197 32313137 4130562 10128785
## American Indian or Aleut or Eskimo
## Did not vote 1999820
## Voted 2098503## White Black More than one race Asian or Pacific Islander
## Did not vote 0.2973482 0.2993155 0.3885320 0.4111793
## Voted 0.7026518 0.7006845 0.6114680 0.5888207
## American Indian or Aleut or Eskimo
## Did not vote 0.4879606
## Voted 0.5120394##
## Pearson's Chi-squared test
##
## data: tbl2
## X-squared = 1864975, df = 4, p-value < 2.2e-164.2.3 Marital Status
## Did not vote Voted
## 114853697 259817184##
## Chi-squared test for given probabilities
##
## data: tbl4
## X-squared = 56087659, df = 1, p-value < 2.2e-164.2.4 Veteran Status
## No service Yes
## Did not vote 107821916 7031781
## Voted 235033022 24784162##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: tbl5
## X-squared = 1196521, df = 1, p-value < 2.2e-164.2.5 Citizenship
## Born abroad of American parents Born in U.S Born in U.S. outlying
## Did not vote 1064127 100237770 1229508
## Voted 2307519 235363239 1394888
## Naturalized citizen
## Did not vote 12322292
## Voted 20751538##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: tbl5
## X-squared = 1196521, df = 1, p-value < 2.2e-164.2.6 Labor Force
## No, not in the labor force Yes, in the labor force
## Did not vote 42247712 72605984
## Voted 89934150 169883034##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: tbl7
## X-squared = 164185, df = 1, p-value < 2.2e-164.3 Logistic Generalized Linear Models
##
## Call:
## glm(formula = as.factor(Voted) ~ AGE, family = "binomial", data = trim1618)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.9493 -1.2968 0.7173 0.8747 1.1328
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.3325070 0.0161454 -20.59 <2e-16 ***
## AGE 0.0243576 0.0003231 75.39 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 186890 on 152823 degrees of freedom
## Residual deviance: 180929 on 152822 degrees of freedom
## AIC: 180933
##
## Number of Fisher Scoring iterations: 4##
## Call:
## glm(formula = as.factor(Voted) ~ AGE + as.factor(SEX) + Metro +
## RaceSimp + Vet + Citizen + Labor + EduSimp, family = "binomial",
## data = trim1618)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.6036 -1.0220 0.5683 0.8141 2.2380
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) -2.2848579 0.0721985 -31.647
## AGE 0.0316859 0.0003883 81.593
## as.factor(SEX)2 0.1178478 0.0127306 9.257
## MetroCentral city status unknown -0.1309099 0.0197097 -6.642
## MetroNot identifiable -0.1067393 0.0593822 -1.797
## MetroNot in metro area -0.0895915 0.0190150 -4.712
## MetroOutside central city 0.0032961 0.0164771 0.200
## RaceSimpBlack 0.2999303 0.0208877 14.359
## RaceSimpMore than one race -0.0983910 0.0459707 -2.140
## RaceSimpAsian or Pacific Islander -0.5467592 0.0317662 -17.212
## RaceSimpAmerican Indian or Aleut or Eskimo -0.4507596 0.0502083 -8.978
## VetYes 0.0881883 0.0238815 3.693
## CitizenBorn in U.S 0.0986694 0.0642475 1.536
## CitizenBorn in U.S. outlying -0.3652759 0.1009645 -3.618
## CitizenNaturalized citizen -0.2755766 0.0676390 -4.074
## LaborYes, in the labor force 0.3067889 0.0143658 21.356
## EduSimpHigh school graduate or GED 0.7680608 0.0223958 34.295
## EduSimpSome college but no degree 1.4008940 0.0244355 57.330
## EduSimpAssociate degree 1.4976338 0.0275499 54.361
## EduSimpBachelors degree 2.1132725 0.0257172 82.174
## EduSimpMasters degree 2.4268736 0.0339748 71.432
## EduSimpProfessional or Doctoral degree 2.5381708 0.0513136 49.464
## Pr(>|z|)
## (Intercept) < 2e-16 ***
## AGE < 2e-16 ***
## as.factor(SEX)2 < 2e-16 ***
## MetroCentral city status unknown 3.10e-11 ***
## MetroNot identifiable 0.072257 .
## MetroNot in metro area 2.46e-06 ***
## MetroOutside central city 0.841450
## RaceSimpBlack < 2e-16 ***
## RaceSimpMore than one race 0.032331 *
## RaceSimpAsian or Pacific Islander < 2e-16 ***
## RaceSimpAmerican Indian or Aleut or Eskimo < 2e-16 ***
## VetYes 0.000222 ***
## CitizenBorn in U.S 0.124595
## CitizenBorn in U.S. outlying 0.000297 ***
## CitizenNaturalized citizen 4.62e-05 ***
## LaborYes, in the labor force < 2e-16 ***
## EduSimpHigh school graduate or GED < 2e-16 ***
## EduSimpSome college but no degree < 2e-16 ***
## EduSimpAssociate degree < 2e-16 ***
## EduSimpBachelors degree < 2e-16 ***
## EduSimpMasters degree < 2e-16 ***
## EduSimpProfessional or Doctoral degree < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 186494 on 152617 degrees of freedom
## Residual deviance: 163922 on 152596 degrees of freedom
## (206 observations deleted due to missingness)
## AIC: 163966
##
## Number of Fisher Scoring iterations: 44.4 Tree Models
4.4.1 Boosted Trees for Variable Importance
## var rel.inf
## EduSimp EduSimp 51.8737011
## AGE AGE 34.4071612
## RaceSimp RaceSimp 5.9799034
## Citizen Citizen 3.6491491
## Labor Labor 2.3461848
## SEX SEX 1.2491621
## Vet Vet 0.49473824.4.2 Classification Tree
##
## Classification tree:
## tree(formula = as.factor(Voted) ~ AGE + EduSimp, data = trim1618)
## Number of terminal nodes: 3
## Residual mean deviance: 1.137 = 173500 / 152600
## Misclassification error rate: 0.2898 = 44222 / 152618
## node), split, n, deviance, yval, (yprob)
## * denotes terminal node
##
## 1) root 152618 186500 Voted ( 0.3001 0.6999 )
## 2) EduSimp: Some school but no diploma,High school graduate or GED,Some college but no degree 84939 114300 Voted ( 0.3992 0.6008 )
## 4) AGE < 45.5 34903 48310 Did not vote ( 0.5227 0.4773 ) *
## 5) AGE > 45.5 50036 62190 Voted ( 0.3130 0.6870 ) *
## 3) EduSimp: Associate degree,Bachelors degree,Masters degree,Professional or Doctoral degree 67679 62950 Voted ( 0.1758 0.8242 ) *