The Problem with the Ubiquitous School Rating System in the US
Using test scores as a measure of school quality is widely criticized because test scores are 70-80% attributed to parent income, not school quality. As a result, using test score data as a major factor in school ratings creates bias towards schools where parents have high incomes. And this system unfairly underrates schools where parents have diverse or low incomes.
So, why is it that parent incomes determine test score outcomes?
Consider the extremes: a wealthy family vs a family experiencing poverty. The experience of the child from 0-5 years old in these two homes is wildly different when it comes to education. The child in the wealthy family has an abundance of books, educational games, flashcards, gets read to every night, hears more words, attends an expensive preschool that actually has a curriculum, maybe one parent stays at home, and most likely both parents are highly educated. By the time Kindergarten hits, the child is hard wired for learning and tends to do very well on the standardized tests that start in 3rd grade.
And in addition to a 5 year head start, the child in the wealthy family has additional resources like paid tutors, test prep, etc.
The child in the family that is experiencing poverty probably has less educated parents, is more likely to be in a single parent household, and education likely takes a back seat to the need for food, water and shelter.
The point of this comparison is that the achievement gap between these two kids is not the school’s fault.
When straight test scores are used to compare schools, the schools where parents have lower incomes get lower scores, and vice versa.
We don’t think the size of parents wallets should be a factor in determining school quality.
In addition, the ratings don’t consider children with disabilities (CWD). Schools that have higher concentrations of these students get lower scores because these children tend to score lower on the test.
SchoolSparrow’s Rating System
Our system evaluates student performance in the context of the school’s population of economically disadvantaged students and students with disabilities. Our algorithm is a multi-variable regression with input variables that describe the demographics of school, including the % of kids that are considered economically disadvantaged, and the % of kids that have disabilities.
We also publish a diversity rating that describes a school’s racial diversity. Research has shown that kids in diverse environments (if it’s done right) get significant cognitive benefits such as better teamwork and problem solving, more empathy and acceptance of others, and possibly even higher IQ’s. These benefits cannot be replicated at a homogenous school.
In this way we elevate thousands of high quality schools that are currently completely obscured from view. We call these schools Out-Performers. We’ve learned through interviews with staff at these Out-Performing Schools, that a common theme is a strong principal that has developed a safe and inclusive culture with high morale and low teacher turnover.
Students thrive in these environments.