Corrigendum to “The Complicated Partisan Effects of State Election Laws”

• Published date: 01 March 2018
• Date: 01 March 2018
• DOI: 10.1177/1065912917735547
• Subject Matter: Corrigenda

https://doi.org/10.1177/1065912917735547

Political Research Quarterly

2018, Vol. 71(1) 242

© 2017 University of Utah

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1065912917735547

journals.sagepub.com/home/prq

Corrigenda

735547PRQXXX10.1177/1065912917735547Political Research Quarterly

correction2017

Corrigendum to “The Complicated

Partisan Effects of State Election Laws”

Burden, Barry C., David T. Canon, Kenneth R. Mayer, and Donald P. Moynihan. 2017. “The Complicated Partisan

Effects of State Election Laws.” Political Research Quarterly 70 (3): 564–76. (Original DOI: 10.1177/1065912917704513)

In an article exploring the difficulties of estimating causal effects with clustered data, MacKinnon and Webb (2017)

extended some of the analysis reported in our recent article. In doing so, they uncovered two errors in Table 7.

Specficially, the early voting variable was miscoded in the 2008 to 2012 models and an incorrect clustering variable was

applied. A corrected version of the table is reported below. None of the other analyses in our article were affected.

References

MacKinnon, James G., and Matthew D. Webb. 2017. “Pitfalls When Estimating Treatment Effects Using Clustered Data.” The

Political Methodologist 24(3): 20–31.

Table 7 (corrected). Difference-in-Difference Models of the County Democratic Vote Share.

2004–2008 2008–2012

Δ Early voting −0.12 (1.49) −0.34 (1.53) 0.49 (.086) −0.25 (0.71)

Δ Ex-felons barred 0.17 (0.47) −1.69*** (0.58) 0.77 (1.07) 1.17*** (0.41)

Δ ID requirement 0.67 (1.98) 1.94 (1.95) −1.89 (1.47) −0.56 (1.40)

% African American — 0.03 (0.02) — 0.09*** (0.01)

% Hispanic — 0.09*** (0.02) — 0.06*** (0.01)

Median income (in 1,000s) — 0.05* (0.03) — 0.03 (0.01)

% college graduates — 0.04 (0.03) — −0.014 (0.015)

% sixty-five or older — −0.26*** (0.06) — 0.06** (0.04)

Population (in 100,000s) — −0.024*** (0.005) — 0.009 (0.007)

Population density — −0.00005*** (0.00001) — 0.00003*** (0.00001)

Constant 4.74*** (0.47) 3.31 (2.16) −1.64*** (0.34) −5.78*** (1.34)

R2.002 .26 .03 .35

n3,112 3,112 3,112 3,111

Cell entries are linear regression coefficients with robust standard errors clustered by county in parentheses. Dependent variable and key

independent variables are differenced. The same states held gubernatorial elections in 2004, 2008, and 2012, so this variable is not included.

“EDR,” “Early voting + EDR,” and the thirty-day closing date did not change in any state between elections, so the differences do not vary and

cannot be included in the models. EDR = Election Day registration.

*p < .10. **p < .05. ***p < .01 (two-tailed test).