Beliefs, Information and the Education Plans of Middle School Children in the Dominican Republic

There are two parts to this study: a "Policy Pilot" in which videos are shown in 7th and 8th grade classrooms in a given school, and an "individual treatment" sample in which short informative videos are shown and in-depth surveys are conducted individually with students and their parents.

Policy Pilot

The policy pilot consists of two sets of four videos that were shown to 7th and 8th grade students in spring 2015. The four videos, each approximately 15 minutes long, tell a story of several characters in 8th grade considering their future educational choices. Each video contains an “infographic” describing in more detail the potential benefits to obtaining a higher level of education. Across the two sets of videos, the infographics differ by their information content. The “informative” infographics contain detailed statistics on income levels and distributions by level of education, potential non-monetary benefits to education, and scholarships for higher education. The “persuasive” infographics contain qualitative information on income differences by level of education, scholarships, and non-monetary benefits, but do not include detailed statistics.

In addition to the videos, posters were distributed to be displayed in the classrooms. Along the same lines as the infographics, persuasive and informative posters were designed to either emphasize the benefits of schooling qualitatively, or provide average earnings information for different schooling levels. The informative poster was further split into two types, one that contained earnings information for men, and another that contained earnings information for women.

Individual Treatments

The individual treatments were designed to provide more detailed tests the effects of providing earnings information and distributions at different levels of education. Students or parents were randomly selected to be provided with the informative infographic that describes earnings averages and distributions at each level of education. Earnings of either men or women were provided based on the gender of the student.

2015-04-14

2016-05-15

Primary:
Enrollment

Secondary
Test scores and test participation
Survey-based secondary outcomes

The primary outcome variable for the 2015 interventions will be enrollment in the 2015-2016 school year. Enrollment will be determined through "matriculas" data obtained from the Ministry of Education in the years following the study.

The key secondary outcome variable will be the student’s scores on the national Pruebas Nacionales exam taken in 8th grade. We will also measure impacts on participation in the Pruebas Nacionales exam. Based on the school and individual surveys, we plan on analyzing treatment effects on a number of intermediate outcomes. These outcomes include beliefs of average earnings and the distribution of earnings, expected self earnings and the distribution of self earnings, educational aspirations and plans, and time use. The exact set of secondary outcomes to be analyzed and the form of the analysis will be informed by the primary impact analysis

2015 Design

Policy Pilot

Six hundred schools were randomly selected from the universe of public upper primary schools in the Dominican Republic. These 600 schools were then randomly assigned to one of three groups:
T1: Persuasive videos (200 schools)
T2: Informative videos (200 schools)
C: No videos (200 schools)

Randomization was conducted by grouping schools into rural/urban status and region (north, southeast, and southwest). Within each of these 6 groups, one school was randomly selected. Based on the predicted dropout rate of that school, two additional schools with the closest predicted dropout within the group were selected. Randomization into T1, T2, and C groups was conducted within these triplets.
The type of informative poster (either men’s or women’s earnings information) was randomly assigned across the 200 schools in T2.

Individual Treatments
250 schools were randomly selected to be included in the individual treatments:
T1: 50 schools
T2: 50 schools
C: 50 schools
Other public schools: 50 schools
Other private schools: 50 schools

Within each selected school, up to 28 students were selected to be included in the individual treatments. Among these students, the first 12 appearing on the roster were selected to be visited at home to have surveys and/or information provided to parents. These 12 were randomly assigned as follows:

P1: 4 students: information for parent
P2: 4 students: information for student
P3: 4 students: surveys only

For the remaining 16 students, surveys were conducted with the students only. Of these, half were assigned to receive information:

S1: 8 students: information for student
S2: 8 students: surveys only

Policy pilot schools and classrooms: Randomization using Matlab
Individual treatments: Randomization using Stata based on school roster number

Policy pilot: School
Individual treatments: Individual student

Yes

Policy Pilot: 600 schools
Individual treatments: 7000 households (student, parent, or both)

43,000 students; 3,000 parents

Policy Pilot:

T1: Persuasive videos (200 schools)
T2: Informative videos (200 schools)
C: No videos (200 schools)

Individual Treatments:
P1: 1000 students
P2: 1000 students
P3: 1000 students

S1: 2000 students
S2: 2000 students