What's Up With Schools?
Cynthia Clifford
Strategic Energy Management Data Analyst at CLEAResult -- Creative Problem Solver | Data-Driven Insights | Client-Centric Solutions Specialist
"Education is not preparation for life; education is life itself." John Dewey
An Analysis of Massachusetts Public Schools---Who's Doing Well and Who is Not?
I'm a high school math teacher transitioning into data analytics. I've spent years thinking about what works in schools and what doesn't. I grew up right outside of Boston and graduated from Brookline High School. When I had the opportunity to learn Tableau by analyzing public school data from Massachusetts, I jumped at the chance.
I used pre-covid 19 data and imagined I had been hired by the Secretary of Education of the Commonwealth of Massachusetts to analyze educational data and create a report and dashboard to illuminate the results of the analysis. The full dashboard can be seen on Tableau Public.
The Secretary of Education wants a big picture of school performance and wants to increase access to post-secondary education. He wants to know what factors affect college admission and attendance rates. Diversity, equity, and inclusion are key areas of interest.
Key Insights
The Data
The?dataset?comes from Kaggle and is a compilation of data from the Massachusetts Department of Education 2017 reports. There are 1,861 rows and 272 columns of data in the file. There are many metrics in the file.
The Analysis
Which Schools Struggle the Most?
The Secretary of Education is interested in knowing which schools struggle the most. He wants to know which schools have the lowest graduation rates. To answer this, I created a bar graph showing graduation percentage and filtered out null values so only high schools remained, since the original data included all public schools in the state. Then I chose to display the bottom 20 high schools by graduation rates.
The high school which initially appeared to have the lowest graduation rate, a rate of 0%, was Curtis-Tufts High School in Medford. Upon further investigation, this high school is a therapeutic high school that transitions students to other schools for graduation. Thus, this school was eliminated by filtering out not only null values but values of zero.
Many of the schools with low graduation rates are alternative schools of various sorts, and providing the Secretary of Education with this information was deemed to be useful. In particular, I determined that, in schools with low graduation percentages, between 71% and 100% of the students are high-needs students.
A comprehensive data dictionary is found in the "About the Data" section of the Massachusetts Department of Education website. High-needs students are defined as students who are either economically disadvantaged, English language learners, or with disabilities (meaning they have an Individualized Education Plan).
Further investigation is needed to determine whether the state is adequately meeting the needs of its neediest learners, as a significant high-needs population is strongly correlated with low graduation rates.
College Attendance
The Secretary of Education is wondering how college attendance could be increased. He's considering investing in building more schools in order to lower the average class size to hopefully increase the % attending college.
I investigated the existing relationship between class size and college attendance.
I created a scatter plot and initially, it appeared that college attendance and class size were, counterintuitively, positively correlated.?On closer inspection, however, there are outliers on both ends of the spectrum.?A number of schools with low-class sizes also had low college attendance rates but, as in the graduation rates data, many of these schools are special schools designed for interventions for the neediest students and are not representative of the bulk of schools where class sizes ranged from around 12 to 19 and college attendance ranged from around 70% to 90%. It does not appear that class size was the overriding factor in college attendance in most schools.
领英推荐
To investigate further, I added a filter for percent of economically disadvantaged students (schools with a high percentage of economically disadvantaged students are shown in orange). There is a definite relationship here, with lower college attendance occurring with more economically disadvantaged students.
If the goal is to improve college acceptance rates, there are some outliers whose programs could be studied to see what lessons could be learned from their success.?For example, City on a Hill Charter Public School Circuit Street has 56.3% economically disadvantaged students and a 91.1% college attendance rate. It has an average class size of 12. New Mission High School, with an average class size of 16.5, has 53.1% economically disadvantaged students and a college attendance rate of 89.5%. And Match Charter Public School has an average class size of 26.5, 56.4% economically disadvantaged students, and a college attendance rate of 90.7%. Since these outliers differ significantly in class size, additional factors could be explored to determine which factors are most strongly correlated with program success.
I created additional scatter plots to see if other factors were more strongly associated with college attendance.?I looked at how college attendance rates were related to the percentage of economically disadvantaged students and added the filter of the percentage of non-native English speakers (higher percentages of non-native English speakers are shown in orange).
I found a fairly strong relationship here, but only for schools with a high percentage of economically disadvantaged students. In schools without a high percentage of economically disadvantaged students, there was no discernible relationship.
The graph below shows this clearly. There is not a significant relationship between schools with high percentages of English language learners and college attendance rates except when non-native English speakers are also economically disadvantaged. Then there is a strong relationship.
I don't recommend that the state build more schools to reduce class sizes.?I recommend that schools that have demonstrated extraordinary success with college attendance rates for economically disadvantaged students be examined in detail to determine which factors are leading to success.?These factors could then be replicated.
4th Grade Math
The Secretary of Education believes that 4th-grade math is key to a student's future and would like to focus on improving the state's MCAS 4th Grade Math Passing rate.
He wants to know which districts are above the desired threshold of 50% passing and wants to collect the names of these districts with plans to invite some of these teachers to train the rest of the state’s teachers on how they can improve math scores.?
I filtered out schools that didn’t have 4th-graders or 4th-grade MCAS scores and organized the schools by their math MCAS pass percentages.
If the Secretary of Education is correct, based on their MCAS Math Scores, about 75% of Massachusetts school districts are adequately preparing students for success. The highest performing districts are Hingham and Winchester, two districts in which over 90% of the students have passing MCAS Math scores. Teachers could be recruited from these districts to train other teachers in the state. (Note: In my hometown, 76.63% of 4th graders passed the MCAS Math Test. Acceptable but leaving room for improvement.)
As a final step, I made a dashboard and included a few KPIs.
Insights and Recommendations
There is a strong negative correlation between the percentage of economically disadvantaged students and the rate of college attendance. There are, however, some noteworthy outliers who seem able to prepare large proportions of their economically disadvantaged students for higher education. I recommend these outlier schools be looked at in-depth and studied to determine what factors, qualitative and quantitative, lead to their success.
High-needs populations tend to graduate with lower percentages than the rest of the state, and the problem is exacerbated if the population also has significant numbers of English language learners. Massachusetts is doing a good job of preparing the majority of its students but further investigation is needed to determine whether the state is adequately meeting the needs of its neediest learners.
75% of school districts have over a 50% pass rate on the 4th grade MCAS Math test. Two districts have over 90% of their students passing the test. Teachers from those districts, Hingham and Winchester, should be enlisted to train other teachers in the state.
Thank you so much for reading my project. Suggestions are welcome. If you have any questions, feel free to comment below or reach out to me via email at [email protected] or connect with me on?LinkedIn where you can see my other projects.
I'm currently looking for opportunities as a data analyst. If you know of any opportunities, please reach out. Thank you!
Data Scientist at EMeRG | Data Science | Statistics | Machine Learning | Deep Learning | Artificial Intelligence | LLM | Big Data | Python | Research | SQL | PowerBI | Web Scraping |
1 年Excellent work! By the way, inspired by your insights!
Data Analyst | Excel | Tableau | SQL | Educator
2 年Excellent write up and analysis Cynthia.
Editor of 'The AI Way' a weekly email newsletter focussed on Education and AI. | Pioneering AI in Education & Self-Learning | Explore AI's Frontier with My Weekly Newsletter |1340+ Subscribers & Growing
2 年You did a great analyst.
Data Analyst | ETL | Excel | SQL | Tableau
2 年Loved it! You did a very thorough analysis.