Childcare, Labor Supply, and Business Development: Experimental Evidence from Uganda

icpsr_197964

Kjetil Bjorvatn, Denise Ferris, Selim Gulesci, Arne Nasgowitz, Vincent Somville, Lore Vandewalle (2025). American Economic Journal: Applied Economics 2025, 17(2): 75–101

DOI: https://doi.org/10.1257/app.20230227

field	value
venue	American Economic Journal: Applied Economics 2025, 17(2): 75–101
paper	https://doi.org/10.1257/app.20230227
replication data	openICPSR 197964
data license	CC BY 4.0 — “License for Data: This work is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) License. Summary of Availability: All data are publicly available.”
regression datasets (in this corpus)	3
graduated / eligible / exported	3 / 21 / 340 (policy `claims_manual_regression_cap`)
routed / gap / residue (at source)	42 / 27 / 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	On the extensive margin, fathers in the childcare arm were 10 pp more likely to be working for a wage, which is a 26 percent increase compared to the control group, where 38 percent of the fathers were in salaried employment. p. 16 · Table 4, col 6 (Labor supply Wage >0 (6))	+0.097	0.040	0.015	Linear Regression	`reg_10__5 · T_2` n=970 · d=17
C2	⭐ primary	🔴 high	“Compared to the control group average of UGX 143,000, this corresponds to a 19 percent increase.” — p. 19 “In particular, households in the childcare arm generated UGX 28,000 more.” — p. 19 Table 5, col 1 (Total household income (1))	+27.838	13.458	0.039	Linear Regression	`reg_2__20 · T_2` n=1411 · d=18
C3	⭐ primary	🟠 medium	We find that the childcare subsidy—alone or when combined with cash—had positive and significant effects of about 0.15 SD on the aggregate score, driven by significant improvements in emergent literacy and motor development. p. 20 · Table 6, col 1 (IDELA score (1))	+0.158	0.057	0.006	Linear Regression	`reg_2__27 · T_2` n=1366 · d=18
C4	⭐ primary	🟠 medium	We find that the childcare subsidy—alone or when combined with cash—had positive and significant effects of about 0.15 SD on the aggregate score, driven by significant improvements in emergent literacy and motor development. p. 20 · Table 6, col 1 (IDELA score (1))	+0.152	0.058	0.009	Linear Regression	`reg_2__27 · T_4` n=1366 · d=18
C5	primary	🔴 high	The cash transfer and the combined treatment also have positive but imprecisely estimated effects on total household income. p. 19 · Table 5, col 1 (Total household income (1))	+4.702	13.000	0.718	Linear Regression	`reg_2__20 · T_3` n=1411 · d=18
C6	primary	🔴 high	The cash transfer and the combined treatment also have positive but imprecisely estimated effects on total household income. p. 19 · Table 5, col 1 (Total household income (1))	+7.827	13.319	0.557	Linear Regression	`reg_2__20 · T_4` n=1411 · d=18
C7	primary	🟠 medium	“Cash transfers had a positive effect on household consumption and no significant impact on child development, and led to an increase in the prevalence of IPV, as reported by the mothers.” — p. 23 “Turning to the cash treatment, the impact on the aggregate score is 0.09 SD and on its components is positive but not statistically significant.” — p. 20 Table 6, col 1 (IDELA score (1))	+0.089	0.057	0.118	Linear Regression	`reg_2__27 · T_3` n=1366 · d=18

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_197964_reg_10__5
icpsr_197964_reg_2__20
icpsr_197964_reg_2__27

icpsr_197964_reg_10__5 — Y1_C_pt_D ~ T_2 + T_3 + T_4 + district_1 + district_2 + district_3 + district_4 + distric…

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	-0.06993	0.0749
`T_2`	childcare-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.09741	0.0398
`T_3`	cash-transfer-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.07239	0.0419
`T_4`	combined childcare+cash treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.04915	0.0408
`district_1`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.05457	0.0651
`district_2`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.02681	0.0644
`district_3`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.1191	0.0685
`district_4`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.05212	0.0603
`district_5`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.1653	0.0599
`district_6`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.1321	0.065
`district_7`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.1355	0.0661
`district_9`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.1284	0.0688
`strat_younger_sibling`	randomization stratum: presence of younger sibling	continuous	⚪ low	0.0	1.0	+0.05243	0.0325
`strat_school`	randomization stratum: preschool/school affiliation	continuous	⚪ low	0.0	1.0	-0.01619	0.0336
`occup_1`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	+0.00426	0.0422
`occup_3`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	+0.0788	0.0435
`strat_rel_tchild`	randomization stratum: respondent's relation to target child	continuous	⚪ low	0.0	1.0	+0.1761	0.0501
`Y1_C_pt_D_bl_r`	baseline value of father wage-employment dummy	continuous	🟠 medium	0.0	1.0	+0.3751	0.0322

Response

name	description	type	sensitivity	bound lo	bound hi	estimate	s.e.
`Y1_C_pt_D`	father wage-employment dummy (extensive margin)	continuous	🟠 medium	0.0	1.0	—	—

Public bounds (data-independent) sourced from: baseline value of underlying binary indicator; binary employment indicator ('D' = dummy); binary indicator; binary location dummy; binary occupation-category dummy; randomization stratification dummy.

n = 970 samples.

icpsr_197964_reg_2__20 — Y5_A_pro_w ~ T_2 + T_3 + T_4 + district_1 + district_2 + district_3 + district_4 + distri…

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	+10.4	19.5
`T_2`	childcare-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+27.84	13.5
`T_3`	cash-transfer-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+4.702	13
`T_4`	combined childcare+cash treatment arm assignment	continuous	⚪ low	0.0	1.0	+7.827	13.3
`district_1`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	-2.687	16.8
`district_2`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+99.3	23.9
`district_3`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+59.78	21
`district_4`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+69.84	20
`district_5`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+53.75	18
`district_6`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+87.29	19.4
`district_7`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+44.13	21.4
`district_9`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	-7.229	15.9
`strat_younger_sibling`	randomization stratum: presence of younger sibling	continuous	⚪ low	0.0	1.0	-0.9013	11.1
`strat_school`	randomization stratum: preschool/school affiliation	continuous	⚪ low	0.0	1.0	+6.033	11.6
`occup_1`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	+12.17	10.6
`occup_2`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	+38.47	15
`strat_rel_tchild`	randomization stratum: respondent's relation to target child	continuous	⚪ low	0.0	1.0	+55.16	13.9
`Y5_A_pro_w_bl_r`	baseline total household income	continuous	🔴 high	—	—	+0.1347	0.0606
`Y5_A_pro_w_bl_miss`	missing-indicator for baseline household income	continuous	⚪ low	0.0	1.0	+30.68	13.1

Response

name	description	type	sensitivity	bound lo	bound hi	estimate	s.e.
`Y5_A_pro_w`	total household income, winsorized	continuous	🔴 high	—	—	—	—

Public bounds (data-independent) sourced from: binary indicator; binary location dummy; binary occupation-category dummy; missingness dummy indicator; randomization stratification dummy.

n = 1,411 samples.

icpsr_197964_reg_2__27 — tot_idela_sd ~ T_2 + T_3 + T_4 + district_1 + district_2 + district_3 + district_4 + dist…

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	-0.5612	0.108
`T_2`	childcare-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.1581	0.057
`T_3`	cash-transfer-only treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.08875	0.0567
`T_4`	combined childcare+cash treatment arm assignment	continuous	⚪ low	0.0	1.0	+0.1521	0.0582
`district_1`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+1.197	0.0933
`district_2`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.8173	0.0921
`district_3`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.5788	0.095
`district_4`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.5881	0.101
`district_5`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.4477	0.0985
`district_6`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.5756	0.0889
`district_7`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+1.011	0.109
`district_9`	district fixed-effect assignment indicator	continuous	🟠 medium	0.0	1.0	+0.4444	0.0911
`strat_younger_sibling`	randomization stratum: presence of younger sibling	continuous	⚪ low	0.0	1.0	-0.1145	0.0463
`strat_school`	randomization stratum: preschool/school affiliation	continuous	⚪ low	0.0	1.0	+0.2827	0.0461
`occup_1`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	-0.08328	0.0559
`occup_3`	household head/father occupation category indicator	continuous	🟠 medium	0.0	1.0	-0.128	0.058
`strat_rel_tchild`	randomization stratum: respondent's relation to target child	continuous	⚪ low	0.0	1.0	-0.0373	0.0637
`tot_idela_sd_bl_r`	baseline child early-learning development score	continuous	🟠 medium	-6.0	6.0	+0.4756	0.0251
`tot_idela_sd_bl_miss`	missing-indicator for baseline IDELA score	continuous	⚪ low	0.0	1.0	-0.1827	0.0837

Response

name	description	type	sensitivity	bound lo	bound hi	estimate	s.e.
`tot_idela_sd`	child early-learning development score, standardized	continuous	🟠 medium	-6.0	6.0	—	—

Public bounds (data-independent) sourced from: binary indicator; binary location dummy; binary occupation-category dummy; missingness dummy indicator; randomization stratification dummy; standardized score (SD units per name/claim); standardized score (SD units).

n = 1,366 samples.