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:
  • 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.

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.

Comments

Popular posts from this blog

"Why A.I. Isn’t Going to Make Art" by Ted Chiang

Culinary Math and Visual Mnemonics

Planning for the Fall: Flipping my class and shadow grading