Khollaboration with Khan Academy: Leveraging Computer Assisted Learning, Teachers, and Parents to Facilitate Personalized Learning at Scale

How to cite

Oreopoulos, Philip; Gibbs, Chloe; Price, Joseph. 2021. "Khollaboration with Khan Academy: Leveraging Computer Assisted Learning, Teachers, and Parents to Facilitate Personalized Learning at Scale." AEA RCT Registry. AEARCTR-0007682. https://doi.org/10.1257/rct.7682-2.2

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Philip Oreopoulos* , Chloe Gibbs , Joseph Price

* Primary investigator

Registry ID
AEARCTR-0007682
DOI
https://doi.org/10.1257/rct.7682-2.2
First registered
Last updated
Status
In development
Trial dates
2021-06-21 – 2023-07-31
Countries
Private
JEL codes
I20, I21
Registry record
http://www.socialscienceregistry.org/trials/7682

Abstract

This research investigates the effectiveness of providing teachers with supports for integrating Khan Academy programming with their math instruction. The study will be conducted as a randomized control trial comparing math achievement of students whose teachers have been offered support from a Khan Academy “Khollaborator” to students whose teachers do not receive the Khollaborator support and conduct their math instruction as usual. Teachers in the treatment group will also receive regular progress reports and support for communicating with parents and students. Results will provide information about how to most effectively align Khan Academy with classroom efforts, and how doing so impacts students’ math achievement.

Intervention

The program, called “Khollaboration With Khan Academy” (KWiK), aims to increase student engagement with Khan Academy to provide students with a greater personalized learning experience, allowing students to receive more immediate feedback, progress at their own pace, and focus on mastering course material. To effectively integrate Khan Academy into teachers’ classrooms, treatment group teachers will participate in training prior to the start of the school year and university student coaches (Khollaborators) will regularly communicate with teachers to offer support and guidance. Khollaborators will also provide weekly progress updates to students in treatment group classrooms and their parents.

Experimental design

The research team will employ a randomized control trial to answer the research question: What is the impact of Khollaboration with Khan Academy (KWiK) training for teachers on students’ math achievement? To answer this, researchers will randomize assignment to KWiK at the grade-school level. Students whose teachers are assigned to KWiK will comprise the treatment group, whereas students assigned to teachers not offered training are the control group. Teachers who sign up to participate and are not offered KWiK in the 2021-22 school year will receive the intervention the following year (SY 2022-23).

The primary outcome for both is students’ performance on math standardized assessments. The research team will also use data from Khan Academy to measure usage (in minutes) and performance on the Khan Academy platform (number of level-ups). Secondary outcomes of interest include student attendance in school, and student and teachers’ satisfaction.

Design details

Randomization method
Teachers will be randomized at the school-grade level. Researchers will use Stata to assign each school-grade level unit to a random number uniformly distributed between 0 and 1. This number will be replicable using the associated seed number. Randomization will be stratified at the grade level, meaning that each grade-school unit will be sorted by its random number and ranked in order relative to other grade-school units within the same grade. Grade-school units with a ranking below the 50 percentile will be considered treated, grade-school units with a ranking above the 50th percentile will be considered control.
Randomization unit
Randomization will occur at the school-grade level. Randomization will be stratified at the grade level.
Clusters
Based on current interest survey data, there will be approximately 180 grade-school units.
Observations
Based on current survey data, the study will have approximately 300 teacher participants. Assuming each classroom has approximately 25 students, we expect a study size of 5500 students.
Arms
We plan to randomize school-grade level clusters using a 50-50 ratio. We expect to have approximately 4500 students in treatment group classrooms and 4500 students in control group classrooms.
Primary outcomes
1) Math skill development as measured through Khan Academy progress and performance 2) Math achievement as measured through interim and year-end standardized tests
Secondary outcomes
Secondary outcomes of interest include student attendance in school, disciplinary incidents, on-time grade progression, and student and teachers’ satisfaction.

Cited by

No citation data available for this registration yet.

Keywords