Electrifying Education: Evidence from Solar-powered Lights, TVs and Bilingual Videos in Tanzanian Secondary Schools

Last registered on March 01, 2024

View Trial History

Pre-Trial

Trial Information

General Information

Title

Electrifying Education: Evidence from Solar-powered Lights, TVs and Bilingual Videos in Tanzanian Secondary Schools

RCT ID

AEARCTR-0000842

Initial registration date

October 22, 2015

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

October 22, 2015, 2:49 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated

March 01, 2024, 7:18 PM EST

Last updated is the most recent time when changes to the trial's registration were published.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name

Hee Kwon Seo

Affiliation

The University of Chicago

Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status

On going

Start date

2015-09-02

End date

2024-12-31

Keywords

Education, Environment & Energy, Labor

Additional Keywords

JEL code(s)

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Approximately one billion people lack electricity access today, and three billion face the need to use a nonnative language on a daily basis; we study how these two constraints affect learning in public schools. We ask: how does school electrification affect education? How does variation in digital learning content, in particular the language of instruction, affect demand for education and performance? How can we measure the value of these school resources? We pilot-test these questions via tracking the results of a unique experiment conducted with 200 middle schools in northern Tanzania--educating 40,000 students each year--that are in the process of receiving, in a randomized, staggered roll-out fashion, solar-powered lights, TVs and language-varied educational videos over the years 2015 to 2017. We value private welfare changes by examining how students trade off attendance during class hours and remedial hours against their opportunity costs of time. We are additionally interested in complementing this analysis by using natural predictors of grid placement to identify the impact of energy expansion over the past decade on school performance (as well as on other measures of economic productivity as they relate to growth).

External Link(s)

Registration Citation

Citation

Seo, Hee Kwon. 2024. "Electrifying Education: Evidence from Solar-powered Lights, TVs and Bilingual Videos in Tanzanian Secondary Schools." AEA RCT Registry. March 01. https://doi.org/10.1257/rct.842-2.0

Former Citation

Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Experimental Details

Interventions

Intervention(s)

We install solar-powered lights, TVs and language-varied educational videos over the years 2015 to 2017.

Intervention Start Date

2015-09-17

Intervention End Date

2016-11-30

Primary Outcomes

Primary Outcomes (end points)

O-level test scores and pass rates of 2016 and 2017 graduating cohorts

Primary Outcomes (explanation)

O-level test scores: normalized GPA (grade point average) as given by the national examination records
O-level pass rates: indicators for passing (getting at least Division IV)

Secondary Outcomes

Secondary Outcomes (end points)

school-reported "hours of video-viewing per average week"; school-reported percentage of students "staying late to use electricity"

Secondary Outcomes (explanation)

Experimental Design

Schools are randomized into the following treatment groups.
1. “Solar x TV x No Video (G1)” receiving solar lights and TVs (“facilities”) only;
2. “Solar x TV x English Videos (G2)” receiving solar facilities and English videos;
3. “Solar x TV x Bilingual Videos (G3)” receiving solar facilities and bilingual videos;
4. “No Solar x No TV x English Videos (G4)” receiving English videos only;
5. “No Solar x No TV x Bilingual Videos (G5)” receiving bilingual videos only;
6. “Control (G6)” receiving neither solar facilities nor video.

G1 receives the solar systems in 09/2015. G2 receives the solar systems and English learning materials in 09/2015. G3 receives the solar systems and bilingual learning materials in 09/2015. G4 receives the English learning materials only in 09/2015, but receives the solar systems at the end of 2017 (after the endline). The G5 receives bilingual learning materials in 2015, and receives the solar systems at the end of 2017. The G6 receives only the solar system at the end of 2017.

This design allows us to compare G1-G5 with G6.

This design also allows for a "pooled" comparison of schools with solar (G1-G3) vs. schools without solar (G4-G6), whose result we report to summarize.

This design finally allows us to look at results for 2016 (year 1) and 2017 (year 2) both separately and pooled together. These results will be reported in the final paper to be produced including the 2016 and 2017 outcome data.

Experimental Design Details

Not available

Randomization Method

School-randomization is done by Stata. Schools are stratified at the region-level and then at 6-school previous year national examination GPA-average bands. Code is reproducible.

Randomization Unit

Randomization for the energy / video intervention is done at the school level.

Was the treatment clustered?

Yes

Experiment Characteristics

Sample size: planned number of clusters

164 schools.

The initial number of schools received from the government containing schools without electricity in the project-assigned districts was 208. This list was used to randomize the schools; however, subsequent visit to deliver intervention materials let to a reduction in sample size due to "no electricity" eligibility restriction. Out of 208 schools, 34 schools were already connected to the national grid or had solar facilities installed in all classrooms. Further 10 schools were relatively newly established schools, whose students were too young to register for the 2016 junior-secondary exit exams. These schools were excluded from the program, as they should never have been included in the sample in the first place. No intervention was delivered and no data was collected from these 44 schools. Thus the sample number of clusters is 164 schools.

Sample size: planned number of observations

Approximately 20,000 students. 10,171 students in the graduating cohort of 2016. The number of students in the graduating cohort of 2017 is pending verification.

Sample size (or number of clusters) by treatment arms

1. “Solar x TV x No Video (G1)” receiving solar lights and TVs (“facilities”) only (35 schools randomized, 26 schools verified to be eligible upon visit and registered into program);
2. “Solar x TV x English Videos (G2)” receiving solar facilities and English videos (35 schools randomized, 29 schools registered);
3. “Solar x TV x Bilingual Videos (G3)” receiving solar facilities and bilingual videos (36 schools randomized, 31 schools registered);
4. “No Solar x No TV x English Videos (G4)” receiving English videos only (35 schools randomized, 29 schools registered);
5. “No Solar x No TV x Bilingual Videos (G5)” receiving bilingual videos only (33 schools randomized, 24 schools registered);
6. “Control (G6)” receiving neither solar facilities nor video (34 schools randomized, 25 schools registered).

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Power calculations can be seen in section 3.5.1 of "IGC_FInal-report_cover2.pdf" targeting 2016 outcome data. In the final paper, which will include 2017 data pending receipt, results will be reported analyzing results for 2016 (year 1) and 2017 (year 2) both separately and pooled together.

Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

IRB