Understanding Teacher Decision-Making: Parent-Teacher Alignment in Multidimensional Skill Development

The intervention is an information treatment delivered through a custom website (wiseteaching.net). Treated teachers receive:

Individual student reports showing parent preferences and current ability ratings (0-100) across nine dimensions of human capital grouped into academic, social, and emotional categories
Classroom-level summaries of parent priorities
Pedagogical strategies for improving non-cognitive skills based on feedback from teachers

Teachers access this information through individual login credentials that limit their access to only their classroom's data. The information intervention reduces uncertainty about parent preferences but does not change the weight teachers place on parent preferences or their own preferences.

The intervention is an information treatment delivered through a custom website (wiseteaching.net). Treated teachers receive:

Individual student reports showing parent preferences and current ability ratings (0-100) across nine dimensions of human capital grouped into academic, social, and emotional categories
Classroom-level summaries of parent priorities
Pedagogical strategies for improving non-cognitive skills based on feedback from teachers

Teachers access this information through individual login credentials that limit their access to only their classroom's data. The information intervention reduces uncertainty about parent preferences but does not change the weight teachers place on parent preferences or their own preferences.

2024-04-30

2025-03-01

Teacher belief updating:
Correlation between teacher beliefs about parent preferences and actual parent preferences

Student academic outcomes:
School-provided grades in academic subjects
Student attendance rates

Student non-cognitive outcomes:
Parent ratings (0-100) of child's abilities across nine dimensions of human capital
Teacher assessments of student capabilities in social and emotional domains (report card)

Teaching practices:
Self-reported teacher implementation of suggested pedagogical strategies

Teacher belief updating is measured by the correlation between teachers' perceptions of parent preferences and actual parent preferences. I examine how this correlation changes after the intervention for treatment versus control teachers. Higher correlation indicates more accurate teacher understanding of what parents want for their children.

For student outcomes, I use:
Academic performance comes directly from school-provided grades and attendance records
Non-cognitive skills are measured through parent ratings on a 0-100 scale for nine dimensions of human capital grouped into academic, social, and emotional categories

Parent and teacher outcomes are captured through:
Current ability ratings: Parents rate each student's current abilities across nine skill dimensions on a 0-100 scale
Importance rankings: Both groups rank which skills they consider most important for the student to improve (a subset for teachers)

For teaching practices, I collect self-reported data on:
Implementation of pedagogical strategies (for the treatment group)

This study will be done in conjunction with four schools in four states across India. Schools will administer surveys for teachers and parents of students from 1st to 10th grade. Within each grade, students are grouped into sections, each of which is associated with a class teacher. Each section has approximately 30-40 students. Students will also be surveyed at endline.

Within the largest school in the sample, there are an average of 10 sections in each grade (100 total sections), and each section contains approximately 40 students (~4000 students total). In the smallest school, there are 3-4 sections in each grade, totaling ~1100 students. Across the schools, there are ~8150 student-parent pairs.

Having completed the baseline surveys, I have collected responses from ~3,300 student-parent pairs, representing 43.5% of the full student body.

The experiment randomizes at the classroom (teacher) level following a two-stage design:

1. Random split of grades with either odd or even grades getting information on pedagogical strategies
2. Within each group:
• Grades that get strategy info: 80% receive parent preferences and ratings information
• Grades that do not get strategy info: 20% receive parent preferences and ratings information

This design minimizes the threat of spillovers by assuming teachers within a grade communicate, thus assigning strategy information at the grade level, while parent information is specific to each class. The within-grade randomization proportions were chosen to maximize the power for comparing control and the combined treatment with both strategy and parent information.

Data collection occurs across multiple waves. Baseline surveys took place in April 2024 for the first school, and in September 2024 for the three other schools. The information treatment began in the week after the baseline surveys were completed for teachers. Endline surveys will take place in April 2025 for all schools, with a follow-up wave planned for August 2025. The timing aligns with the Indian academic calendar, with academic years starting in April.

This study will be done in conjunction with four schools in four states across India. Schools will administer surveys for teachers and parents of students from 1st to 10th grade. Within each grade, students are grouped into sections, each of which is associated with a class teacher. Each section has approximately 30-40 students. Students will also be surveyed at endline.

Within the largest school in the sample, there are an average of 10 sections in each grade (100 total sections), and each section contains approximately 40 students (~4000 students total). In the smallest school, there are 3-4 sections in each grade, totaling ~1100 students. Across the schools, there are ~8150 student-parent pairs.

Having completed the baseline surveys, I have collected responses from ~3,300 student-parent pairs, representing 43.5% of the full student body.

The experiment randomizes at the classroom (teacher) level following a two-stage design:

1. Random split of grades with either odd or even grades getting information on pedagogical strategies
2. Within each group:
• Grades that get strategy info: 80% receive parent preferences and ratings information
• Grades that do not get strategy info: 20% receive parent preferences and ratings information

This design minimizes the threat of spillovers by assuming teachers within a grade communicate, thus assigning strategy information at the grade level, while parent information is specific to each class. The within-grade randomization proportions were chosen to maximize the power for comparing control and the combined treatment with both strategy and parent information.

Data collection occurs across multiple waves. Baseline surveys took place in April 2024 for the first school, and in September 2024 for the three other schools. The information treatment began in the week after the baseline surveys were completed for teachers. Endline surveys will take place in April 2025 for all schools, with a follow-up wave planned for August 2025. The timing aligns with the Indian academic calendar, with academic years starting in April.

Randomization done in office by a computer.

Treatment is clustered at the classroom (section) level.

Yes

240 classrooms

3300 students

96 classrooms control, 24 classrooms receive strategy information, 24 classrooms receive parent preferences and ratings, 96 classrooms receive both