Correction for Bernhard et al., Men and women candidates are similarly persistent after losing elections

Political Sciences Correction for “Men and women candidates are similarly persistent after losing elections,” by Rachel Bernhard and Justin de Benedictis-Kessner, which published June 21, 2021; 10.1073/pnas.2026726118 (Proc. Natl. Acad. Sci. U.S.A. 118, e2026726118).

The authors note that: “We have identified two coding errors. First, due to a typo, we unintentionally omitted California local elections from 2001. Second, for a small number of simultaneous elections for multiple seats in the same jurisdiction in California, we grouped candidates from different races into the same race. These errors led to reported sample sizes and effect sizes that were slightly inaccurate. However, they do not affect the main results of the paper.”

In the abstract: “Using a regression discontinuity design and data that feature 212,805 candidates across 22,473 jurisdictions….” should instead read “Using a regression discontinuity design and data that feature 213,966 candidates across 22,519 jurisdictions…”

On page 1, left column, third full paragraph, line 10: “drawing on state and local election data that feature 212,805 candidates across 22,473 jurisdictions” should instead read “drawing on state and local election data that feature 213,966 candidates across 22,519 jurisdictions.”

On page 2, left column, sixth full paragraph, line 1: “We use data on 141,116 state legislative elections nationwide, 23,466 local elections from California cities and counties, and 9,132 mayoral elections across the country” should instead read “We use data on 141,116 state legislative elections nationwide, 24,470 local elections from California cities and counties, and 9,130 mayoral elections across the country.”

On page 2, right column, first full paragraph, line 6: “Together, these sources of data encompass 212,805 candidates across 22,473 jurisdictions between 1950 and 2018” should instead read “Together, these sources of data encompass 213,966 candidates across 22,519 jurisdictions between 1950 and 2018.”

On page 3, right column, second full paragraph, line 17: “The size of this effect varies between 14 percentage points (in California local races)” should instead read “The size of this effect varies between 13 percentage points (in California local races).”

On page 3, right column, fourth full paragraph, line 5: “In California county, city, and school district elections, men and women are both 13 percentage points less likely to run again if they barely lose rather than barely win.” should instead read “In California county, city, and school district elections, men candidates are 13 percentage points less likely to run again if they barely lose rather than barely win and women candidates are 12 percentage points less likely to do so.”

In Footnote ††: “The difference in California local races of 0.1 percentage points is similarly insignificant (p = 0.974)” should instead appear as “The difference in California local races of 1.8 percentage points is similarly insignificant (p = 0.628).”

The authors also note that, due to the coding errors, Fig. 1B and Fig. 2 appeared incorrectly. The corrected figures and their legends appear below.

Fig. 1.

(A–C) Regression discontinuity effects, by jurisdiction.

Fig. 2.

Regression discontinuity effects, by gender.

Additionally, several minor changes have been made to the supporting information.

The online version has been corrected and the SI Appendix has been updated.