This gets to a critique that I have had about the ways in which D65 does data analysis. I know Tom mentioned he is a Bayesian. When they were doing the analysis for the new school the first thing they should have done was a probabilistic regression looking at the impact of busing on an outcome measure. You control for class (free/reduced…
This gets to a critique that I have had about the ways in which D65 does data analysis. I know Tom mentioned he is a Bayesian. When they were doing the analysis for the new school the first thing they should have done was a probabilistic regression looking at the impact of busing on an outcome measure. You control for class (free/reduced lunch eligibility) and race (which they require every student to report during registration) and see what the results look like.
There are plenty of low-income kids and kids of color that walk to school outside of the Fifth Ward so you have good counter examples.
So in terms of the survey, Tom it would be cool to foia any anonymized data sets that show this stuff. I am sure there are all sorts of privacy issues, so it may be difficult.
My sense is that they don't conduct any inferential analysis (which is part of the problem). The Ed.Ds that administrators like Horton and Turner have do not require any statistical training. Horton's dissertation was based on interviews of like 10 people. Turner's program at UIC doesn't even require a dissertation!
If there were disaggregated public anonymized data on performance, it would be very interesting to use for a proper statistical analysis.
This gets to a critique that I have had about the ways in which D65 does data analysis. I know Tom mentioned he is a Bayesian. When they were doing the analysis for the new school the first thing they should have done was a probabilistic regression looking at the impact of busing on an outcome measure. You control for class (free/reduced lunch eligibility) and race (which they require every student to report during registration) and see what the results look like.
There are plenty of low-income kids and kids of color that walk to school outside of the Fifth Ward so you have good counter examples.
So in terms of the survey, Tom it would be cool to foia any anonymized data sets that show this stuff. I am sure there are all sorts of privacy issues, so it may be difficult.
My sense is that they don't conduct any inferential analysis (which is part of the problem). The Ed.Ds that administrators like Horton and Turner have do not require any statistical training. Horton's dissertation was based on interviews of like 10 people. Turner's program at UIC doesn't even require a dissertation!
If there were disaggregated public anonymized data on performance, it would be very interesting to use for a proper statistical analysis.