LAK13 Assignment #1 - Analytics: Logic and Structure
Introduction
There is an effort within US public K-12 schools to analyze student performance, and use that analysis to improve schools. The most public and political aspect of this effort has been the use of this data to measure teacher effectiveness. A few school districts have released student and teacher information (Los Angeles, New York City, etc.), and have made them widely available. This data can be used to answer basic questions about student and teacher performance.
Questions
My main question; is this a good idea? Are student scores a good reflection of teacher effectiveness? There are a few other related questions I'd also like to explore:
Data Sources
There is a wide range of raw data, and measures based on this data, available to the public;
The only student and teacher performance data that is in an accessible state is the Colorado School Grades' data from Kaggle. The data from NYC Open Data has always been fairly accessible, but with the wide variety of data types, I'm a bit unsure of their usability. At this time I am unsure if I can get access to the LAUSD data in a usable format.
Next Steps
I would like to continue researching available datasets, and from there identify the ones that seem most usable. Once those have been identified, use R to perform exploratory data analysis, and identify any useful trends. Using the ratings/rankings of different school districts on student and teacher data that has a different measure may also help identify potential flaws in each measure.
There is an effort within US public K-12 schools to analyze student performance, and use that analysis to improve schools. The most public and political aspect of this effort has been the use of this data to measure teacher effectiveness. A few school districts have released student and teacher information (Los Angeles, New York City, etc.), and have made them widely available. This data can be used to answer basic questions about student and teacher performance.
Questions
My main question; is this a good idea? Are student scores a good reflection of teacher effectiveness? There are a few other related questions I'd also like to explore:
- Are the analysis methods between school districts transferable?
- Is the value-added model a 'good' one?
- Should parents use the teacher ratings/rankings to make decisions about where their child goes to school?
- Do these ratings/rankings say anything useful about college and work readiness, unemployment, crime, etc.?
Potential Issues
This issue is a politically charged one, and I'm concerned about getting an unbiased dataset, and that the ratings/rankings contain hidden assumptions that are not based on fact. Having this concern does not mean that I won't use certain datasets, but I will work under a trust and verify policy.
Pulling in datasets from multiple school districts, agencies, and bureaus may cause issues of data comparability. I'm unsure of how to deal with these issues, and would appreciate any suggestions.
There is a wide range of raw data, and measures based on this data, available to the public;
- The L.A. Times released the Los Angeles Teacher Ratings in May 2011. Using data from the Los Angeles Unified School District they calculated the value-added to student scores on a range of measures, by teacher and school.
- In February 2012, the New York City released a smaller set of teacher data, which The Wall Street Journal used to create teacher ratings and posted them to Grading the Teacher. The NYC Open Data site also has a wide range of education data.
- The Colorado Department of Education's SchoolView Data Center contains a variety of data on student and teacher performance. Colorado School Grades also shares this data on their site, and released a dataset on Kaggle for wider distribution and analysis.
- For demographic information The United States Census Bureau's Data Access Tools site contains a wide variety and sources.
- The U.S. Department of Labor, Bureau of Labor Statistics, and the U.S. Department of Justice, Bureau of Justice Statistics both allow access to useful data sets.
The only student and teacher performance data that is in an accessible state is the Colorado School Grades' data from Kaggle. The data from NYC Open Data has always been fairly accessible, but with the wide variety of data types, I'm a bit unsure of their usability. At this time I am unsure if I can get access to the LAUSD data in a usable format.
Next Steps
I would like to continue researching available datasets, and from there identify the ones that seem most usable. Once those have been identified, use R to perform exploratory data analysis, and identify any useful trends. Using the ratings/rankings of different school districts on student and teacher data that has a different measure may also help identify potential flaws in each measure.
Comments
Post a Comment