Skip to content

An Experiment in Candidate Selection

icpsr_124501

Katherine Casey, Abou Bakarr Kamara, Niccoló F. Meriggi (2021). American Economic Review

DOI: 10.1257/aer.20200125

field value
venue American Economic Review
paper 10.1257/aer.20200125
replication data openICPSR 124501
data license CC BY 4.0 (openICPSR project page 124501 (maintainer-read 2026-07-20)) — “This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.”
regression datasets (in this corpus) 5
graduated / eligible / exported 5 / 33 / 118 (policy claims_manual_regression_cap)
routed / gap / residue (at source) 65 / 104 / 0

Claims & tasks

Every claim the paper makes in its main body about a parameter, joined to the benchmark task (dataset · coefficient) that captures it. The effect, s.e. and p are the captured regression's own reported values (the paper's Stata figures, carrying the paper's SE method); the row shows the paper's verbatim quote (page · exhibit), so you can confirm the regression matches what the paper says. The prominence column is the within-study inclusion signal (⭐ headline → primary → secondary). The sensitivity column is the privacy level of the most-sensitive variable in the regression (🔴 high / 🟠 medium / ⚪ low).

# prominence sensitivity claim effect s.e. p task instance (dataset · coef)
C1 ⭐ primary 🔴 high “This rate increases by 23 percentage points with treatment, a large and highly statistically significant effect, which is driven by party responsiveness in safe and weak seats.” — p. 1577
“The new selection mechanism has a large positive impact on representation: the estimated treatment effect on selecting the voters’ first choice aspirant is 23 percentage points, standard error 10.6 (panel A of Table 3).” — p. 1592
Table 3, col 1 ((1))
+22.956 10.608 0.034 Linear Regression reg_1 · treat
n=91 · d=23
C2 primary 🔴 high “The coefficient on development record in Table 4 is a precisely estimated 0.04 for both voters and party leaders (standard error 0.01 and 0.02, respectively), which implies that a one standard deviation unit increase in public goods provision is associated with a 4 percentage point gain in an aspirant’s vote share in the V2 and P1 surveys.” — p. 1596
“Overall, the headline results that come through this triangulation approach are first, that both voters and party officials value aspirants who have a demonstrated record of providing public goods in the constituency.” — p. 1602
“Development record stands out as the one index that registers positively and significantly for both voters and party officials across all of these specifications.” — p. 1596
Table 4, col 1 (Aspirant's share in voter polls (1))
+0.043 0.013 0.001 Linear Regression reg_14__1 · devz
n=385 · d=31
C3 primary 🔴 high Estimates for the conscientiousness indicator suggest that returning the money is associated with a precisely estimated increase of 5 (9) percentage points in vote share among voters (party officials).
p. 1596 · Table 4, col 1 (Aspirant's share in voter polls (1))
+0.051 0.025 0.042 Linear Regression reg_14__1 · return
n=385 · d=31
C4 primary 🔴 high “The coefficient on development record in Table 4 is a precisely estimated 0.04 for both voters and party leaders (standard error 0.01 and 0.02, respectively), which implies that a one standard deviation unit increase in public goods provision is associated with a 4 percentage point gain in an aspirant’s vote share in the V2 and P1 surveys.” — p. 1596
“Overall, the headline results that come through this triangulation approach are first, that both voters and party officials value aspirants who have a demonstrated record of providing public goods in the constituency.” — p. 1602
Table 4, col 2 (Aspirant's share in party official survey (2))
+0.042 0.020 0.044 Linear Regression reg_14__2 · devz
n=367 · d=30
C5 primary 🔴 high Estimates for the conscientiousness indicator suggest that returning the money is associated with a precisely estimated increase of 5 (9) percentage points in vote share among voters (party officials).
p. 1596 · Table 4, col 2 (Aspirant's share in party official survey (2))
+0.089 0.036 0.016 Linear Regression reg_14__2 · return
n=367 · d=30
C6 primary 🔴 high Online Appendix Table A3 shows that stronghold voters prefer aspirants with better economic development records, and interestingly, further share party officials' preference for those with more professional qualifications.
p. 1605 · Table A.3, col 4 (Stronghold Races Only: Aspirant's share in V2 (4))
+0.085 0.030 0.007 Linear Regression reg_16__1 · pqz
n=235 · d=20
C7 primary 🔴 high Online Appendix Table A3 shows that stronghold voters prefer aspirants with better economic development records, and interestingly, further share party officials' preference for those with more professional qualifications.
p. 1605 · Table A.3, col 5 (Stronghold Races Only: Aspirant's share in P1 (5))
+0.101 0.059 0.092 Linear Regression reg_16__2 · pqz
n=226 · d=19
C8 secondary 🔴 high Wealth and public service motivation also at least marginally predict voter preferences, and while not precise predictors for party officials, the estimated coefficients are positive and comparable in magnitude for both groups.
p. 1596 · Table 4, col 1 (Aspirant's share in voter polls (1))
+0.034 0.015 0.029 Linear Regression reg_14__1 · weaz
n=385 · d=31
C9 secondary 🔴 high Wealth and public service motivation also at least marginally predict voter preferences, and while not precise predictors for party officials, the estimated coefficients are positive and comparable in magnitude for both groups.
p. 1596 · Table 4, col 1 (Aspirant's share in voter polls (1))
+0.038 0.021 0.070 Linear Regression reg_14__1 · psmz
n=385 · d=31
C10 secondary 🔴 high Development record stands out as the one index that registers positively and significantly for both voters and party officials across all of these specifications.
p. 1596 · Table A.3, col 4 (Stronghold Races Only: Aspirant's share in V2 (4))
+0.063 0.016 2.7e-04 Linear Regression reg_16__1 · devz
n=235 · d=20

Regression datasets

The 5 regression dataset(s) graduated into the benchmark corpus from this study's replication package — each backs one or more claims above. Expand each for its features, response, (sound, data-independent) public bounds, and reproduction grade.

icpsr_124501_reg_1 — v2first_won ~ treat + 2.strata + 3.strata + 4.strata + 5.strata + 6.strata + 7.strata + 8…

Sensitivity proposal: 🔴 high maximum across columns (advisory; reviewed manually). Reproduction grade:match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
_cons intercept continuous 1 1 +21.86 18.8
treat randomized primary-election treatment assignment indicator continuous ⚪ low 0.0 1.0 +22.96 10.6
2.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 29.4
3.strata randomization stratification cell/block identifier categorical ⚪ low +4.167 27
4.strata randomization stratification cell/block identifier categorical ⚪ low +22.17 30
5.strata randomization stratification cell/block identifier categorical ⚪ low +33.33 29.9
6.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 26.2
7.strata randomization stratification cell/block identifier categorical ⚪ low -8.333 29.1
8.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
9.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
10.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
11.strata randomization stratification cell/block identifier categorical ⚪ low +66.67 21.2
12.strata randomization stratification cell/block identifier categorical ⚪ low +66.67 20.1
13.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 40.5
14.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 35.4
15.strata randomization stratification cell/block identifier categorical ⚪ low -8.333 29.1
16.strata randomization stratification cell/block identifier categorical ⚪ low +50 28.1
17.strata randomization stratification cell/block identifier categorical ⚪ low -8.333 35.2
18.strata randomization stratification cell/block identifier categorical ⚪ low -33.33 21.2
19.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
20.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
21.strata randomization stratification cell/block identifier categorical ⚪ low -33.33 21.2
22.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37
23.strata randomization stratification cell/block identifier categorical ⚪ low +16.67 37

Response

name description type sensitivity bound lo bound hi estimate s.e.
v2first_won respondent's V2 top-choice aspirant ultimately won continuous 🔴 high

Public bounds (data-independent) sourced from: binary indicator.

n = 91 samples.

icpsr_124501_reg_14__1 — v2_vs ~ pqz + weaz + devz + psmz + loyz + iqz + locz + cez + return + 2.strata + 3.strata…

Sensitivity proposal: 🔴 high maximum across columns (advisory; reviewed manually). Reproduction grade:match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
_cons intercept continuous 1 1 +0.09827 0.0243
pqz standardized professional-qualifications index of aspirant continuous 🟠 medium -6.0 6.0 +0.0458 0.028
weaz standardized wealth index of aspirant continuous 🔴 high -6.0 6.0 +0.03365 0.0152
devz standardized development-record index of aspirant continuous ⚪ low -6.0 6.0 +0.04306 0.0127
psmz standardized public-service-motivation index of aspirant continuous 🟠 medium -6.0 6.0 +0.03787 0.0206
loyz standardized party-loyalty index of aspirant continuous 🔴 high -6.0 6.0 +0.008518 0.0311
iqz standardized cognitive-ability (IQ) score of aspirant continuous 🟠 medium -6.0 6.0 +0.006651 0.0253
locz standardized local-ties/localness index of aspirant continuous 🟠 medium -6.0 6.0 +0.01686 0.0304
cez candidate's social/political connections index (standardized) continuous 🟠 medium -6.0 6.0 -0.004624 0.0209
return aspirant returned campaign money (conscientiousness indicator) continuous ⚪ low 0.0 1.0 +0.05117 0.0248
2.strata randomization stratification cell/block identifier categorical ⚪ low +0.1034 0.0269
3.strata randomization stratification cell/block identifier categorical ⚪ low +0.05106 0.0338
4.strata randomization stratification cell/block identifier categorical ⚪ low +0.03598 0.0344
5.strata randomization stratification cell/block identifier categorical ⚪ low +0.279 0.0733
6.strata randomization stratification cell/block identifier categorical ⚪ low +0.2561 0.0479
7.strata randomization stratification cell/block identifier categorical ⚪ low +0.3326 0.126
8.strata randomization stratification cell/block identifier categorical ⚪ low +0.3791 0.0251
9.strata randomization stratification cell/block identifier categorical ⚪ low +0.2159 0.0203
10.strata randomization stratification cell/block identifier categorical ⚪ low +0.3395 0.174
11.strata randomization stratification cell/block identifier categorical ⚪ low +0.1369 0.118
12.strata randomization stratification cell/block identifier categorical ⚪ low +0.3534 0.0592
13.strata randomization stratification cell/block identifier categorical ⚪ low +0.1673 0.0598
14.strata randomization stratification cell/block identifier categorical ⚪ low +0.01655 0.0273
15.strata randomization stratification cell/block identifier categorical ⚪ low +0.01857 0.0305
16.strata randomization stratification cell/block identifier categorical ⚪ low +0.06426 0.0584
17.strata randomization stratification cell/block identifier categorical ⚪ low +0.01953 0.0472
18.strata randomization stratification cell/block identifier categorical ⚪ low +0.02579 0.0354
19.strata randomization stratification cell/block identifier categorical ⚪ low +0.0235 0.0277
20.strata randomization stratification cell/block identifier categorical ⚪ low +0.3872 0.0721
21.strata randomization stratification cell/block identifier categorical ⚪ low +0.08145 0.0397
22.strata randomization stratification cell/block identifier categorical ⚪ low +0.05989 0.0582
23.strata randomization stratification cell/block identifier categorical ⚪ low +0.08967 0.0276

Response

name description type sensitivity bound lo bound hi estimate s.e.
v2_vs wave-2 survey vote share for aspirant continuous 🔴 high 0.0 1.0

Public bounds (data-independent) sourced from: binary indicator; standardized z-score index; vote share is a proportion.

n = 385 samples.

icpsr_124501_reg_14__2 — p1_vs ~ pqz + weaz + devz + psmz + loyz + iqz + locz + cez + return + 2.strata + 3.strata…

Sensitivity proposal: 🔴 high maximum across columns (advisory; reviewed manually). Reproduction grade:match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
_cons intercept continuous 1 1 +0.06875 0.0392
pqz standardized professional-qualifications index of aspirant continuous 🟠 medium -6.0 6.0 +0.1089 0.0413
weaz standardized wealth index of aspirant continuous 🔴 high -6.0 6.0 +0.02873 0.0273
devz standardized development-record index of aspirant continuous ⚪ low -6.0 6.0 +0.04192 0.0205
psmz standardized public-service-motivation index of aspirant continuous 🟠 medium -6.0 6.0 +0.03178 0.0318
loyz standardized party-loyalty index of aspirant continuous 🔴 high -6.0 6.0 +0.02337 0.0597
iqz standardized cognitive-ability (IQ) score of aspirant continuous 🟠 medium -6.0 6.0 +0.05784 0.0381
locz standardized local-ties/localness index of aspirant continuous 🟠 medium -6.0 6.0 +0.04062 0.0505
cez candidate's social/political connections index (standardized) continuous 🟠 medium -6.0 6.0 -0.00551 0.0379
return aspirant returned campaign money (conscientiousness indicator) continuous ⚪ low 0.0 1.0 +0.08859 0.0359
2.strata randomization stratification cell/block identifier categorical ⚪ low +0.1153 0.0384
3.strata randomization stratification cell/block identifier categorical ⚪ low +0.05312 0.049
4.strata randomization stratification cell/block identifier categorical ⚪ low +0.04093 0.0427
5.strata randomization stratification cell/block identifier categorical ⚪ low +0.3471 0.0925
6.strata randomization stratification cell/block identifier categorical ⚪ low +0.2975 0.0588
7.strata randomization stratification cell/block identifier categorical ⚪ low +0.6368 0.175
8.strata randomization stratification cell/block identifier categorical ⚪ low +0.3813 0.0452
9.strata randomization stratification cell/block identifier categorical ⚪ low +0.2253 0.0333
10.strata randomization stratification cell/block identifier categorical ⚪ low +0.4441 0.177
11.strata randomization stratification cell/block identifier categorical ⚪ low +0.184 0.144
12.strata randomization stratification cell/block identifier categorical ⚪ low +0.397 0.0827
13.strata randomization stratification cell/block identifier categorical ⚪ low +0.1777 0.0603
14.strata randomization stratification cell/block identifier categorical ⚪ low +0.05466 0.042
15.strata randomization stratification cell/block identifier categorical ⚪ low +0.03833 0.052
16.strata randomization stratification cell/block identifier categorical ⚪ low +0.05615 0.0551
17.strata randomization stratification cell/block identifier categorical ⚪ low +0.06935 0.0545
18.strata randomization stratification cell/block identifier categorical ⚪ low +0.07394 0.0704
19.strata randomization stratification cell/block identifier categorical ⚪ low +0.04313 0.0445
21.strata randomization stratification cell/block identifier categorical ⚪ low +0.1091 0.0475
22.strata randomization stratification cell/block identifier categorical ⚪ low +0.01909 0.0538
23.strata randomization stratification cell/block identifier categorical ⚪ low +0.06245 0.0446

Response

name description type sensitivity bound lo bound hi estimate s.e.
p1_vs party official's predicted vote share for aspirant continuous 🔴 high 0.0 1.0

Public bounds (data-independent) sourced from: binary indicator; standardized z-score index; vote share is a proportion.

n = 367 samples.

icpsr_124501_reg_16__1 — v2_vs ~ pqz + weaz + devz + psmz + loyz + iqz + locz + cez + return + 2.strata + 3.strata…

Sensitivity proposal: 🔴 high maximum across columns (advisory; reviewed manually). Reproduction grade:match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
_cons intercept continuous 1 1 +0.1097 0.0244
pqz standardized professional-qualifications index of aspirant continuous 🟠 medium -6.0 6.0 +0.08455 0.03
weaz standardized wealth index of aspirant continuous 🔴 high -6.0 6.0 +0.02222 0.0152
devz standardized development-record index of aspirant continuous ⚪ low -6.0 6.0 +0.06293 0.016
psmz standardized public-service-motivation index of aspirant continuous 🟠 medium -6.0 6.0 +0.01207 0.0187
loyz standardized party-loyalty index of aspirant continuous 🔴 high -6.0 6.0 -0.05167 0.0309
iqz standardized cognitive-ability (IQ) score of aspirant continuous 🟠 medium -6.0 6.0 +0.0388 0.0304
locz standardized local-ties/localness index of aspirant continuous 🟠 medium -6.0 6.0 -0.02495 0.0331
cez candidate's social/political connections index (standardized) continuous 🟠 medium -6.0 6.0 +0.008096 0.0223
return aspirant returned campaign money (conscientiousness indicator) continuous ⚪ low 0.0 1.0 +0.02203 0.0261
2.strata randomization stratification cell/block identifier categorical ⚪ low +0.08925 0.035
3.strata randomization stratification cell/block identifier categorical ⚪ low +0.06392 0.0359
4.strata randomization stratification cell/block identifier categorical ⚪ low +0.05319 0.0343
16.strata randomization stratification cell/block identifier categorical ⚪ low +0.06676 0.0463
17.strata randomization stratification cell/block identifier categorical ⚪ low +0.02271 0.0518
18.strata randomization stratification cell/block identifier categorical ⚪ low +0.02761 0.0372
19.strata randomization stratification cell/block identifier categorical ⚪ low +0.02962 0.0264
20.strata randomization stratification cell/block identifier categorical ⚪ low +0.3853 0.088
21.strata randomization stratification cell/block identifier categorical ⚪ low +0.08697 0.0385
22.strata randomization stratification cell/block identifier categorical ⚪ low +0.08656 0.0553
23.strata randomization stratification cell/block identifier categorical ⚪ low +0.0896 0.0238

Response

name description type sensitivity bound lo bound hi estimate s.e.
v2_vs wave-2 survey vote share for aspirant continuous 🔴 high 0.0 1.0

Public bounds (data-independent) sourced from: binary indicator; standardized z-score index; vote share is a proportion.

n = 235 samples.

icpsr_124501_reg_16__2 — p1_vs ~ pqz + weaz + devz + psmz + loyz + iqz + locz + cez + return + 2.strata + 3.strata…

Sensitivity proposal: 🔴 high maximum across columns (advisory; reviewed manually). Reproduction grade:match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
_cons intercept continuous 1 1 +0.07716 0.0374
pqz standardized professional-qualifications index of aspirant continuous 🟠 medium -6.0 6.0 +0.1012 0.0588
weaz standardized wealth index of aspirant continuous 🔴 high -6.0 6.0 +0.03335 0.0304
devz standardized development-record index of aspirant continuous ⚪ low -6.0 6.0 +0.05585 0.0277
psmz standardized public-service-motivation index of aspirant continuous 🟠 medium -6.0 6.0 +0.01826 0.0384
loyz standardized party-loyalty index of aspirant continuous 🔴 high -6.0 6.0 -0.05815 0.0627
iqz standardized cognitive-ability (IQ) score of aspirant continuous 🟠 medium -6.0 6.0 +0.08158 0.0445
locz standardized local-ties/localness index of aspirant continuous 🟠 medium -6.0 6.0 +0.02211 0.0626
cez candidate's social/political connections index (standardized) continuous 🟠 medium -6.0 6.0 +0.005411 0.0432
return aspirant returned campaign money (conscientiousness indicator) continuous ⚪ low 0.0 1.0 +0.07726 0.0413
2.strata randomization stratification cell/block identifier categorical ⚪ low +0.1007 0.04
3.strata randomization stratification cell/block identifier categorical ⚪ low +0.06621 0.0508
4.strata randomization stratification cell/block identifier categorical ⚪ low +0.05478 0.039
16.strata randomization stratification cell/block identifier categorical ⚪ low +0.0594 0.0518
17.strata randomization stratification cell/block identifier categorical ⚪ low +0.06826 0.0582
18.strata randomization stratification cell/block identifier categorical ⚪ low +0.04929 0.0663
19.strata randomization stratification cell/block identifier categorical ⚪ low +0.05442 0.0437
21.strata randomization stratification cell/block identifier categorical ⚪ low +0.09465 0.0485
22.strata randomization stratification cell/block identifier categorical ⚪ low +0.01939 0.0595
23.strata randomization stratification cell/block identifier categorical ⚪ low +0.05956 0.0475

Response

name description type sensitivity bound lo bound hi estimate s.e.
p1_vs party official's predicted vote share for aspirant continuous 🔴 high 0.0 1.0

Public bounds (data-independent) sourced from: binary indicator; standardized z-score index; vote share is a proportion.

n = 226 samples.