Skip to content

Incentivizing School Attendance in the Presence of Parent-Child Information Frictions

icpsr_154261

Damien de Walque, Christine Valente (2023). American Economic Journal: Economic Policy 2023, 15(3): 256–285

DOI: https://doi.org/10.1257/pol.20210202

field value
venue American Economic Journal: Economic Policy 2023, 15(3): 256–285
paper https://doi.org/10.1257/pol.20210202
replication data openICPSR 154261
data license CC BY 4.0 — “This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.”
regression datasets (in this corpus) 4
graduated / eligible / exported 4 / 15 / 142 (policy claims_manual_regression_cap)
routed / gap / residue (at source) 25 / 0 / 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 “In the control group, absences reported by parents in a household survey are not significantly correlated with actual absences observed during unannounced attendance checks.” — p. 2
“While in control schools, parental self-reported knowledge of their daughter’s school absences has no predictive power on the probability that their daughter was absent at a random attendance check, in treatment schools the coefficient associated with parent-reported absences is significant and more than doubles.” — p. 28
“In the control group, however, the estimated increase is positive but small at only 0.009, and it is statistically insignificant, indicating that the quality of attendance monitoring by parents is low.” — p. 24
Table 4, col 1 (Outcome: Absent at attendance check between October 10 and November 3, 2016; Experimental arm: Control (1))		 +0.009 0.007 0.243 Linear Regression reg_20 · misseddays
n=473 · d=11
C2 ⭐ primary 🟠 medium “Compared to a control group mean of 0.65, the information only treatment increased attendance by 4.5 percentage points (6.9 percent), the parent cash treatment increased attendance by 6 percentage points (9.2 percent), and the child incentive treatment increased attendance by 8.3 percentage points (12.8 percent).” — p. 17
“In our experiment, where the value of the transfer is equivalent to 7 percent of the national GDP per capita—a nonnegligible sum for poor households—the estimated effect of the information treatment on attendance is as large as 75 percent of the effect of a CCT providing the same information.” — p. 17
“Incentivizing girls directly is nearly twice as effective as simply providing information, and in our baseline specification (column 1 of Table 2), this difference is statistically significant at the 10 percent level.” — p. 17
Table 2, col 1 (Share present at attendance check (1))		 +0.045 0.023 0.047 Linear Regression reg_8 · g_info
n=173 · d=13
C3 ⭐ primary 🟠 medium “Compared to a control group mean of 0.65, the information only treatment increased attendance by 4.5 percentage points (6.9 percent), the parent cash treatment increased attendance by 6 percentage points (9.2 percent), and the child incentive treatment increased attendance by 8.3 percentage points (12.8 percent).” — p. 17
“In our experiment, where the value of the transfer is equivalent to 7 percent of the national GDP per capita—a nonnegligible sum for poor households—the estimated effect of the information treatment on attendance is as large as 75 percent of the effect of a CCT providing the same information.” — p. 17
“To summarize, we find evidence that providing high-frequency information to parents about their daughter’s school attendance increases school attendance even in the absence of any transfer and that this effect is not statistically distinguishable from that of a CCT to parents also providing the same information.” — p. 21
Table 2, col 1 (Share present at attendance check (1))		 +0.060 0.022 0.008 Linear Regression reg_8 · g_parents
n=173 · d=13
C4 ⭐ primary 🟠 medium “Compared to a control group mean of 0.65, the information only treatment increased attendance by 4.5 percentage points (6.9 percent), the parent cash treatment increased attendance by 6 percentage points (9.2 percent), and the child incentive treatment increased attendance by 8.3 percentage points (12.8 percent).” — p. 17
“Incentivizing girls directly is nearly twice as effective as simply providing information, and in our baseline specification (column 1 of Table 2), this difference is statistically significant at the 10 percent level.” — p. 17
“Incentivizing girls with vouchers allowing them to buy a choice of goods is at least as effective as incentivizing parents with the cash equivalent of these vouchers.” — p. 21
Table 2, col 1 (Share present at attendance check (1))		 +0.083 0.022 2.6e-04 Linear Regression reg_8 · g_girls
n=173 · d=13
C5 primary 🟠 medium “While in control schools, parental self-reported knowledge of their daughter’s school absences has no predictive power on the probability that their daughter was absent at a random attendance check, in treatment schools the coefficient associated with parent-reported absences is significant and more than doubles.” — p. 28
“In particular, simply introducing the attendance report card without financial incentives more than doubles this probability, which reaches 46 percent of the coefficient expected in the case of perfect monitoring (0.045).” — p. 24
Table 4, col 2 (Outcome: Absent at attendance check between October 10 and November 3, 2016; Experimental arm: Info (2))		 +0.021 0.006 0.001 Linear Regression reg_21 · misseddays
n=406 · d=10
C6 secondary 🔴 high “In column 4 of Table 2, we present estimates of the effect of our treatments on attendance obtained when controlling for the baseline characteristics for which there was at least one statistically significant difference between experimental arms and confirm that results are virtually unchanged.” — p. 18
“The observed increase of 4.88 (8.41) percentage points in daily attendance in the information (child incentive) arm translates into 9 (16) more days of instruction, or an 8 percent (13 percent) increase, to be compared with 8–9 percent increases in average test scores relative to the control mean.” — p. 19
Table 2, col 4 (Share present at attendance check (4))		 +0.049 0.024 0.043 Linear Regression reg_13 · g_info
n=173 · d=25
C7 secondary 🔴 high In column 4 of Table 2, we present estimates of the effect of our treatments on attendance obtained when controlling for the baseline characteristics for which there was at least one statistically significant difference between experimental arms and confirm that results are virtually unchanged.
p. 18 · Table 2, col 4 (Share present at attendance check (4))		 +0.059 0.024 0.014 Linear Regression reg_13 · g_parents
n=173 · d=25
C8 secondary 🔴 high “In column 4 of Table 2, we present estimates of the effect of our treatments on attendance obtained when controlling for the baseline characteristics for which there was at least one statistically significant difference between experimental arms and confirm that results are virtually unchanged.” — p. 18
“The observed increase of 4.88 (8.41) percentage points in daily attendance in the information (child incentive) arm translates into 9 (16) more days of instruction, or an 8 percent (13 percent) increase, to be compared with 8–9 percent increases in average test scores relative to the control mean.” — p. 19
Table 2, col 4 (Share present at attendance check (4))		 +0.084 0.024 5.5e-04 Linear Regression reg_13 · g_girls
n=173 · d=25

Regression datasets

The 4 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_154261_reg_13 — mean_presentg ~ g_info + g_parents + g_girls + _Idistrict_2 + _Idistrict_3 + _Idistrict_4…

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 +0.5745 0.111
g_info treatment arm dummy, information intervention continuous ⚪ low 0.0 1.0 +0.0488 0.0239
g_parents treatment arm dummy, parent cash incentive continuous ⚪ low 0.0 1.0 +0.05883 0.0236
g_girls treatment arm dummy, girls' cash incentive continuous ⚪ low 0.0 1.0 +0.08405 0.0238
_Idistrict_2 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.05607 0.0857
_Idistrict_3 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.08055 0.0858
_Idistrict_4 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.08111 0.0954
_Idistrict_5 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.03477 0.104
_Idistrict_6 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.01028 0.0508
_Idistrict_7 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.06009 0.101
_Idistrict_8 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.1084 0.0895
_Idistrict_9 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.006212 0.0805
_Idistrict_10 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.07373 0.0934
_Idistrict_11 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.07482 0.072
b_misseddays baseline parent-reported days absent, October continuous ⚪ low — — -0.01156 0.0104
b_highmonqual baseline dummy high monitor/teacher qualification continuous 🟠 medium 0.0 1.0 -0.04335 0.0834
b_lportuguese baseline dummy speaks Portuguese continuous ⚪ low 0.0 1.0 +0.09635 0.0806
b_lndau baseline dummy speaks Ndau language continuous ⚪ low 0.0 1.0 -0.009464 0.0526
b_lshona baseline dummy speaks Shona language continuous ⚪ low 0.0 1.0 +0.06168 0.0781
b_lchibarue baseline dummy speaks Chibarue language continuous ⚪ low 0.0 1.0 -0.01369 0.0981
b_lotherlang baseline dummy speaks other language continuous ⚪ low 0.0 1.0 +0.02431 0.0658
b_rcato baseline dummy religion Catholic continuous 🔴 high 0.0 1.0 +0.0839 0.0846
b_rchrist baseline dummy religion Christian (other) continuous 🔴 high 0.0 1.0 +0.09039 0.0792
b_rzioni baseline dummy religion Zionist church continuous 🔴 high 0.0 1.0 +0.08497 0.0711
b_ratheist baseline dummy religion atheist/no religion continuous 🔴 high 0.0 1.0 +0.1898 0.0812
b_rother baseline dummy religion other continuous 🔴 high 0.0 1.0 +0.1603 0.0856

Response

name description type sensitivity bound lo bound hi estimate s.e.
mean_presentg mean daily school attendance rate continuous ⚪ low 0.0 1.0 — —

Public bounds (data-independent) sourced from: binary indicator; binary indicator (xi expansion); share/rate variable.

n = 173 samples.

icpsr_154261_reg_20 — absent ~ misseddays + _Idistrict_2 + _Idistrict_3 + _Idistrict_4 + _Idistrict_5 + _Idistr…

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.
_cons intercept continuous — 1 1 +0.2044 0.0693
misseddays number of school days missed continuous ⚪ low — — +0.008684 0.00732
_Idistrict_2 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.07096 0.082
_Idistrict_3 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1488 0.101
_Idistrict_4 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1261 0.0684
_Idistrict_5 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.08261 0.0664
_Idistrict_6 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.05653 0.069
_Idistrict_7 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.08438 0.0923
_Idistrict_8 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.376 0.0949
_Idistrict_9 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1728 0.0846
_Idistrict_10 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.1078 0.0695
_Idistrict_11 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.09786 0.0912

Response

name description type sensitivity bound lo bound hi estimate s.e.
absent indicator student absent during spot check continuous ⚪ low 0.0 1.0 — —

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

n = 473 samples.

icpsr_154261_reg_21 — absent ~ misseddays + _Idistrict_2 + _Idistrict_3 + _Idistrict_4 + _Idistrict_6 + _Idistr…

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.
_cons intercept continuous — 1 1 +0.3206 0.0259
misseddays number of school days missed continuous ⚪ low — — +0.0207 0.00575
_Idistrict_2 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.09599 0.107
_Idistrict_3 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +8.47e-05 0.0364
_Idistrict_4 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.002535 0.03
_Idistrict_6 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.3461 0.0259
_Idistrict_7 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.1138 0.0609
_Idistrict_8 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.01223 0.0801
_Idistrict_9 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.06128 0.0473
_Idistrict_10 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.2516 0.0259
_Idistrict_11 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.02752 0.08

Response

name description type sensitivity bound lo bound hi estimate s.e.
absent indicator student absent during spot check continuous ⚪ low 0.0 1.0 — —

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

n = 406 samples.

icpsr_154261_reg_8 — mean_presentg ~ g_info + g_parents + g_girls + _Idistrict_2 + _Idistrict_3 + _Idistrict_4…

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.
_cons intercept continuous — 1 1 +0.6071 0.0277
g_info treatment arm dummy, information intervention continuous ⚪ low 0.0 1.0 +0.04504 0.0225
g_parents treatment arm dummy, parent cash incentive continuous ⚪ low 0.0 1.0 +0.05992 0.0222
g_girls treatment arm dummy, girls' cash incentive continuous ⚪ low 0.0 1.0 +0.08285 0.0222
_Idistrict_2 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.05773 0.0351
_Idistrict_3 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.0191 0.0404
_Idistrict_4 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1069 0.0485
_Idistrict_5 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.05749 0.0522
_Idistrict_6 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.03518 0.0485
_Idistrict_7 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1436 0.0328
_Idistrict_8 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 -0.09181 0.0307
_Idistrict_9 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.04728 0.0293
_Idistrict_10 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.1116 0.0521
_Idistrict_11 district identifier for attendance sample (indicator) continuous 🟠 medium 0.0 1.0 +0.08712 0.0328

Response

name description type sensitivity bound lo bound hi estimate s.e.
mean_presentg mean daily school attendance rate continuous ⚪ low 0.0 1.0 — —

Public bounds (data-independent) sourced from: binary indicator; binary indicator (xi expansion); share/rate variable.

n = 173 samples.