Why

What

Who

In many neighborhoods in LA, fewer than 30% of residents have a high school diploma. Music-based mentoring can build confidence and improve academic achievement.

Music-based mentoring that includes 5 hours of instruction from a professional music teacher, at home practice, performances, and family engagement.

Of children living in low-income communities in Los Angeles County, 1,000 children who participate kindergarten through 5th grade, an additional 1,000 children who participate 6th grade through high school and 1,000 children who participate from kindergarten through 12th grade.

For a $34.4 million investment over 13 years covering 3,000 kindergarten thru 12th graders ($1,500 per child annually)

We use a variety of data sources to tailor our analysis to the location and to make our estimates as accurate as possible. However, sometimes data is unavailable, either because it has not been collected, or because it is not shared publicly. We encourage government agencies and other organizations to share their de-identified data online and researchers to prioritize the following gaps in the data.

Educational Data for At-Risk Population: Data from the California Department of Education is available for certain populations, including children on free and reduced lunch. However, approximately 2/3 of children in Los Angeles are eligible for this program, which is likely greater than the at-risk population Harmony Project is designed to target. Because of this, for baseline data on suspensions, truancies, and graduation we used the bottom quartile of schools when collecting data.

However, it would have been more ideal to have better information on the the population that enroll in Harmony Project and the academic outcomes for the communities they come from.

Substance Use: Substance use measures can be inconsistent. For substance use among elementary school-aged children, we use information from a study that tracked whether children in a cohort had tried various substances. For high school substance use, we used a question from a delinquency study in Los Angeles that asked participants about their drug use. We were unable to make any sort of adjustments to get data to approximate the at-risk population we model the program for.

School Attendance: Baseline data for school attendance was difficult to find. We ultimately decided to use a measure that looked at percent of children missing three or more days in the past month, however ideally we would have had an average number of days missed per year.