Skip to content

Preferences, Selection, and the Structure of Teacher Pay

icpsr_198086

Andrew C. Johnston (2025). American Economic Journal: Applied Economics 2025, 17(3): 310–346

DOI: 10.1257/app.20210763

field value
venue American Economic Journal: Applied Economics 2025, 17(3): 310–346
paper 10.1257/app.20210763
replication data openICPSR 198086
data license CC BY 4.0 (openICPSR project page 198086 (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) 3
graduated / eligible / exported 3 / 5 / 114 (policy claims_manual_regression_cap)
routed / gap / residue (at source) 26 / 48 / 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 🟠 medium “High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
“In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
Table 4, col 1 (Choice (1))		 +0.011 0.009 0.218 Linear Regression reg_71 · sal_va_math_4
n=10398 · d=17
C2 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 1 (Choice (1))		 -0.038 0.098 0.696 Linear Regression reg_71 · deduct_va_math_4
n=10398 · d=17
C3 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 1 (Choice (1))		 +0.083 0.072 0.246 Linear Regression reg_71 · prem_va_math_4
n=10398 · d=17
C4 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 1 (Choice (1))		 +0.018 0.023 0.430 Linear Regression reg_71 · growth_va_math_4
n=10398 · d=17
C5 ⭐ primary 🟠 medium “The only way in which high-performing teachers systematically differ is their preferences for offers including performance pay (Table 4 and Figure 5).” — p. 18
“If teachers entertained two comparable offers, a high-performing teacher (top decile) is 23 percent more likely than a bottom-decile one to select the offer providing an additional $3,000 in merit pay per year.” — p. 20
Table 4, col 1 (Choice (1))		 +0.040 0.015 0.008 Linear Regression reg_71 · reward_va_math_4
n=10398 · d=17
C6 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 1 (Choice (1))		 -0.005 0.041 0.909 Linear Regression reg_71 · retire_va_math_4
n=10398 · d=17
C7 ⭐ primary 🟠 medium “In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
“High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
Table 4, col 1 (Choice (1))		 +0.001 0.002 0.678 Linear Regression reg_71 · replace_va_math_4
n=10398 · d=17
C8 ⭐ primary 🟠 medium “High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
“In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
Table 4, col 2 (Choice (2))		 +0.010 0.016 0.522 Linear Regression reg_72 · sal_va_read_4
n=10398 · d=17
C9 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 2 (Choice (2))		 -0.306 0.163 0.060 Linear Regression reg_72 · deduct_va_read_4
n=10398 · d=17
C10 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 2 (Choice (2))		 +0.135 0.118 0.255 Linear Regression reg_72 · prem_va_read_4
n=10398 · d=17
C11 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 2 (Choice (2))		 +0.047 0.036 0.191 Linear Regression reg_72 · growth_va_read_4
n=10398 · d=17
C12 ⭐ primary 🟠 medium “The only way in which high-performing teachers systematically differ is their preferences for offers including performance pay (Table 4 and Figure 5).” — p. 18
“If teachers entertained two comparable offers, a high-performing teacher (top decile) is 23 percent more likely than a bottom-decile one to select the offer providing an additional $3,000 in merit pay per year.” — p. 20
Table 4, col 2 (Choice (2))		 +0.055 0.025 0.028 Linear Regression reg_72 · reward_va_read_4
n=10398 · d=17
C13 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 2 (Choice (2))		 -0.006 0.067 0.926 Linear Regression reg_72 · retire_va_read_4
n=10398 · d=17
C14 ⭐ primary 🟠 medium “In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
“High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
Table 4, col 2 (Choice (2))		 -0.002 0.003 0.457 Linear Regression reg_72 · replace_va_read_4
n=10398 · d=17
C15 ⭐ primary 🟠 medium “High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
“In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
Table 4, col 3 (Choice (3))		 -0.002 0.001 0.015 Linear Regression reg_73 · sal_danielson_total
n=18838 · d=17
C16 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 3 (Choice (3))		 +0.001 0.010 0.937 Linear Regression reg_73 · deduct_danielson_total
n=18838 · d=17
C17 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 3 (Choice (3))		 -0.002 0.008 0.817 Linear Regression reg_73 · prem_danielson_total
n=18838 · d=17
C18 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 3 (Choice (3))		 -0.001 0.003 0.607 Linear Regression reg_73 · growth_danielson_total
n=18838 · d=17
C19 ⭐ primary 🟠 medium “The only way in which high-performing teachers systematically differ is their preferences for offers including performance pay (Table 4 and Figure 5).” — p. 18
“If teachers entertained two comparable offers, a high-performing teacher (top decile) is 23 percent more likely than a bottom-decile one to select the offer providing an additional $3,000 in merit pay per year.” — p. 20
Table 4, col 3 (Choice (3))		 +0.005 0.002 0.002 Linear Regression reg_73 · reward_danielson_total
n=18838 · d=17
C20 ⭐ primary 🟠 medium In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.
p. 18 · Table 4, col 3 (Choice (3))		 +0.005 0.004 0.276 Linear Regression reg_73 · retire_danielson_total
n=18838 · d=17
C21 ⭐ primary 🟠 medium “In terms of work setting characteristics that policymakers can influence, effective teachers have the same preferences as other teachers with regards to class size, salary, income growth, insurance subsidies, retirement benefits, and supportive principals.” — p. 18
“High-quality teachers do not, for instance, have a stronger preference for more generous pensions, higher salary, or high-performing students.” — p. 18
Table 4, col 3 (Choice (3))		 +0.000 0.000 0.040 Linear Regression reg_73 · replace_danielson_total
n=18838 · d=17

Regression datasets

The 3 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_198086_reg_71 — point ~ va_math_4 + sal + sal_va_math_4 + deduct + deduct_va_math_4 + prem + prem_va_math…

Sensitivity proposal: 🟠 medium maximum across columns (advisory; reviewed manually). Reproduction grade: ✅ match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
va_math_4 Teacher's math value-added score, grade 4 continuous 🟠 medium -12.0 12.0 -0.6589 0.436
sal Job scenario: salary level attribute continuous ⚪ low — — +0.09149 0.00374
sal_va_math_4 Salary attribute x math value-added interaction continuous ⚪ low — — +0.01097 0.0089
deduct Job scenario: insurance premium deduction attribute continuous ⚪ low — — -0.5081 0.0582
deduct_va_math_4 Deduction attribute x math value-added interaction continuous ⚪ low — — -0.03822 0.0977
prem Job scenario: insurance premium level attribute continuous ⚪ low — — -0.05044 0.0241
prem_va_math_4 Premium attribute x math value-added interaction continuous ⚪ low — — +0.08341 0.0718
growth Job scenario: salary growth rate attribute continuous ⚪ low — — +0.1829 0.0122
growth_va_math_4 Salary growth attribute x math value-added interaction continuous ⚪ low — — +0.01829 0.0232
reward Job scenario: merit pay bonus attribute continuous ⚪ low — — +0.04552 0.0059
reward_va_math_4 Reward attribute x math value-added interaction continuous ⚪ low — — +0.04043 0.0152
rating_num Job scenario: number of evaluation rating categories continuous ⚪ low — — -0.07118 0.0247
rating_va_math_4 Rating attribute x math value-added interaction continuous ⚪ low — — -0.02464 0.038
retire_num Job scenario: retirement benefit level attribute continuous ⚪ low — — +0.06089 0.0177
retire_va_math_4 Retirement attribute x math value-added interaction continuous ⚪ low — — -0.004688 0.0412
replace Job scenario: dismissal/replacement risk attribute continuous ⚪ low — — +0.01356 0.000937
replace_va_math_4 Replacement-risk attribute x math value-added interaction continuous ⚪ low — — +0.0007522 0.00181

Response

name description type sensitivity bound lo bound hi estimate s.e.
point Teacher's stated preference points for job offer continuous ⚪ low -100.0 100.0 — —

Public bounds (data-independent) sourced from: point-allocation conjoint response, conventionally sums to 100; value-added measures conventionally reported as standardized (SD) scores.

n = 10,398 samples.

icpsr_198086_reg_72 — point ~ va_read_4 + sal + sal_va_read_4 + deduct + deduct_va_read_4 + prem + prem_va_read…

Sensitivity proposal: 🟠 medium maximum across columns (advisory; reviewed manually). Reproduction grade: ✅ match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
va_read_4 Teacher's reading value-added score, grade 4 continuous 🟠 medium -12.0 12.0 -0.1613 0.778
sal Job scenario: salary level attribute continuous ⚪ low — — +0.09178 0.00372
sal_va_read_4 Salary attribute x reading value-added interaction continuous ⚪ low — — +0.01004 0.0157
deduct Job scenario: insurance premium deduction attribute continuous ⚪ low — — -0.5067 0.0578
deduct_va_read_4 Deduction attribute x reading value-added interaction continuous ⚪ low — — -0.3064 0.163
prem Job scenario: insurance premium level attribute continuous ⚪ low — — -0.04862 0.024
prem_va_read_4 Premium attribute x reading value-added interaction continuous ⚪ low — — +0.1349 0.118
growth Job scenario: salary growth rate attribute continuous ⚪ low — — +0.1834 0.012
growth_va_read_4 Salary growth attribute x reading value-added interaction continuous ⚪ low — — +0.04724 0.0361
reward Job scenario: merit pay bonus attribute continuous ⚪ low — — +0.047 0.00586
reward_va_read_4 Reward attribute x reading value-added interaction continuous ⚪ low — — +0.05457 0.0248
rating_num Job scenario: number of evaluation rating categories continuous ⚪ low — — -0.07187 0.0246
rating_va_read_4 Rating attribute x reading value-added interaction continuous ⚪ low — — -0.02534 0.0586
retire_num Job scenario: retirement benefit level attribute continuous ⚪ low — — +0.06156 0.0175
retire_va_read_4 Retirement attribute x reading value-added interaction continuous ⚪ low — — -0.00625 0.067
replace Job scenario: dismissal/replacement risk attribute continuous ⚪ low — — +0.01362 0.000931
replace_va_read_4 Replacement-risk attribute x reading value-added interaction continuous ⚪ low — — -0.00242 0.00325

Response

name description type sensitivity bound lo bound hi estimate s.e.
point Teacher's stated preference points for job offer continuous ⚪ low -100.0 100.0 — —

Public bounds (data-independent) sourced from: point-allocation conjoint response, conventionally sums to 100; value-added measures conventionally reported as standardized (SD) scores.

n = 10,398 samples.

icpsr_198086_reg_73 — point ~ danielson_total + sal + sal_danielson_total + deduct + deduct_danielson_total + p…

Sensitivity proposal: 🟠 medium maximum across columns (advisory; reviewed manually). Reproduction grade: ✅ match (benchmark-eligible).

Features

name description type sensitivity bound lo bound hi estimate s.e.
danielson_total Teacher's total Danielson evaluation framework score continuous 🟠 medium — — +0.09059 0.0496
sal Job scenario: salary level attribute continuous ⚪ low — — +0.08573 0.00278
sal_danielson_total Salary attribute x Danielson score interaction continuous ⚪ low — — -0.002376 0.00098
deduct Job scenario: insurance premium deduction attribute continuous ⚪ low — — -0.5489 0.0423
deduct_danielson_total Deduction attribute x Danielson score interaction continuous ⚪ low — — +0.0007789 0.00979
prem Job scenario: insurance premium level attribute continuous ⚪ low — — -0.09036 0.0179
prem_danielson_total Premium attribute x Danielson score interaction continuous ⚪ low — — -0.001773 0.00767
growth Job scenario: salary growth rate attribute continuous ⚪ low — — +0.1845 0.00923
growth_danielson_total Salary growth attribute x Danielson score interaction continuous ⚪ low — — -0.001377 0.00268
reward Job scenario: merit pay bonus attribute continuous ⚪ low — — +0.02652 0.00449
reward_danielson_total Reward attribute x Danielson score interaction continuous ⚪ low — — +0.005326 0.00172
rating_num Job scenario: number of evaluation rating categories continuous ⚪ low — — -0.07503 0.0184
rating_danielson_total Evaluation rating attribute x Danielson score interaction continuous ⚪ low — — -0.001084 0.00424
retire_num Job scenario: retirement benefit level attribute continuous ⚪ low — — +0.07628 0.0128
retire_danielson_total Retirement-benefit attribute x Danielson score interaction continuous ⚪ low — — +0.004674 0.00429
replace Job scenario: dismissal/replacement risk attribute continuous ⚪ low — — +0.01425 0.000685
replace_danielson_total Replacement-risk attribute x Danielson score interaction continuous ⚪ low — — +0.0004368 0.000213

Response

name description type sensitivity bound lo bound hi estimate s.e.
point Teacher's stated preference points for job offer continuous ⚪ low -100.0 100.0 — —

Public bounds (data-independent) sourced from: point-allocation conjoint response, conventionally sums to 100.

n = 18,838 samples.