|
Field
Trial Status
|
Before
in_development
|
After
completed
|
|
Field
Abstract
|
Before
We conduct an adaptive experiment to select one out of six treatment arms of biweekly phone calls to parents that deliver reading exercises for first graders. The experiment is carried out in two waves - start to midterm, and midterm to endterm - during term 3 of the first grade in a sample of 108 private primary schools in Kenya. The method uses Bayesian econometrics to estimate average treatment effects and exploration sampling (Kasy and Sautmann, 2021) to assign parents to treatment arms in each wave. The objective is to identify the treatment arm with the highest average "correct words per minute" (cwpm) score.
|
After
The application is a test of six formats for interactive phone calls to parents of first-graders in 108 primary schools in Kenya that encourage reading at home through short exercises. Each experimental wave carries out biweekly calls during one half of term 3 of the Kenyan school year and uses the exploration sampling algorithm in wave 2 of the experiment to identify the call with the highest rate of call engagement, that is, the listener starts the reading exercises.
|
|
Field
Last Published
|
Before
June 12, 2021 06:49 AM
|
After
April 09, 2022 11:21 AM
|
|
Field
Experimental Design (Public)
|
Before
We carry out an adaptive experiment with 6 treatment arms and a control group to identify the treatment arm that in expectation leads to the greatest improvements in reading fluency, through engaging parents in reading regularly with their children at home. In the first wave of the experiment, student IDs are randomly assigned to each arm and the control; in the second wave, we calculate the probability that each treatment arm is optimal using Bayesian estimation and then use Exploration Sampling (Kasy and Sautmann, 2021) to assign the wave-2 sample to these arms.
|
After
We carry out an adaptive experiment with 6 treatment arms and a control group to identify the treatment arm that in expectation leads to the greatest parental engagement and therefore improvements in reading fluency, through engaging parents in reading regularly with their children at home. In the first wave of the experiment, student IDs are randomly assigned to each arm and the control; in the second wave, we calculate the probability that each treatment arm is optimal using Bayesian estimation and then use Exploration Sampling (Kasy and Sautmann, 2021) to assign the wave-2 sample to these arms.
|
|
Field
Additional Keyword(s)
|
Before
|
After
adaptive sampling, methodology
|
|
Field
Public analysis plan
|
Before
No
|
After
Yes
|
|
Field
Public locations
|
Before
No
|
After
Yes
|
|
Field
Building on Existing Work
|
Before
|
After
No
|