School data statistics education
January 21, 2011 5:56 PM   Subscribe

You're a teacher, a researcher interested in education, a school employee, a parent with kids "in the system". If you had access to all the data you possibly needed within the realm outlined above, what kind of questions would you want to answer?

I'm looking for research areas that would be possible given access to grades, performance, demographics over multiple years. Preferrably something more exciting than ranking schools or districts or states against each other.
posted by ttyn to Education (16 answers total) 7 users marked this as a favorite
 
Best answer: NELS is a huge longitudinal data set on all sorts of education related variables.
posted by k8t at 6:12 PM on January 21, 2011


link: http://nces.ed.gov/surveys/nels88/
posted by k8t at 6:13 PM on January 21, 2011


Best answer: I would better confront the question if I knew what it was you wanted information for and to what with it.

I have always believed that we Americans need to study how things are done in other countries and then see how we can use what works in ours...for education: why is Finland so successful in its educational system?
posted by Postroad at 6:20 PM on January 21, 2011


Best answer: Effects of early childhood programs, across a variety of demographics, through high school. (Also whether it has any effect on truancy/dropouts.)

Also be interesting to see the effects of robust elementary school art, music, P.E., etc., curricula on later achievement.
posted by Eyebrows McGee at 6:21 PM on January 21, 2011


Best answer: I would like to see someone expand analysis of the recent Center for American Progress report on "educational productivity" to include, as much as possible, an adjustment for transportation costs in remote or sparsely populated areas, and an examination of spending on infrastructure/physical plant considerations inherent in the U.S.'s aged school facilities. Ulrich Boser, the report's principal author, acknowledged "shortcomings" with CAP's study; I'd like to see something that can address them.

I would also like to see analysis and comparison of single-school districts in rural areas with enrollment of less than 250 students and examine their post-graduation endeavors.

There are many uses of various longitudinal reports: predictive studies to develop at-risk student early-warning systems, for example.

I would like to see analyses that will help people who jump on the U.S. in country-comparison studies like PISA or TIMSS to remember that the U.S. is a huge, geographically, linguistically, culturally, and economically diverse nation, unlike Finland.

I assume you are familiar with the Data Quality Campaign and NCES?
posted by jgirl at 6:26 PM on January 21, 2011


Best answer: I'd love information on how/whether length of school day and volume of homework affect educational outcomes. And whether computer use in K-6 is correlated with much of anything long-term beyond how much money the district is spending.
posted by Andrhia at 6:41 PM on January 21, 2011 [1 favorite]


Response by poster: These are awesome. Keep em coming!
posted by ttyn at 7:11 PM on January 21, 2011


Response by poster: Note: countries outside of the US are out of scope, unfortunately.
posted by ttyn at 7:12 PM on January 21, 2011


Best answer: I would want to find out if a particular teaching style was more effective than another. Are certain teachers within a grade at a particular school producing kids year in and year out that outperform their cohorts in their district on some standardized or consistent over time test? Is one teacher in the 3rd grade getting better results? Why? One year it could be the kids, but if it is a consistent pattern you have to attribute the teacher. If you could find a way to measure a teachers impact on learning versus a norm, you could try to replicate their methods, reward them more monetarily and use them to help in professional development with their peers. If I had a trove of national or even multiple school data I would do the same study in each school and then try to compare the successful teachers to see if there are common traits and methods. I think the same data could also point to weaknesses in the district curriculum map.

Second, I would look at the correlation between real estate prices and some measure of success (standardized test scores? ugh I hate them but...) in a district. I think that people are willing to pay for a good education. Sure, we claim it is a public school, but each district has its own taxing authority in my state of NY. If you told me what a 4 bedroom 3500 square foot house was selling for in that district, I would bet that I could guess what their scores were in some test. Ok, so what do you do with that? I am not sure yet, but if you can determine if throwing money at education is as important as teaching technique or the teacher her/himself, you might be getting somewhere.
posted by AugustWest at 8:26 PM on January 21, 2011


Best answer: The effectiveness of differentiated instruction in classrooms with students who have severe behavioral problems.
posted by HotPatatta at 8:27 PM on January 21, 2011


Best answer: OK...not sure if you can 'prove' any of this, but teaching to the test has become the norm in some states, sometimes to the exclusion of all else.

In addition to the current state test, schools are sometimes selected at random to pilot the next revision of the state or district test. The school or district could also be involved with a national education non-profit and have additional testing placed upon students as well. With so much time devoted to testing, content has been reduced to the most minimal of information. Some students don't have basic knowledge of any history (important dates, etc...) or geography (where their city is in relation to the state capital, etc...)

I believe the fact that students are exposed to so many multiple-choice tests, their ability on higher order thinking skills has been diminished. Sure, they can score well on a test that they probably just guessed the answer on by bubbling in a letter. But, can they crawl higher up on Bloom's Taxonomy to synthesize something new on their own.

'Race to the Top' and No Child Left Behind' have created test-taking students that have difficulty with other subjects that don't use the multiple-choice model. What if we pared down the testing and returned to teaching content, creating tests based on the content, not benchmarks or essential questions.

Now we have administrators (locally, in my area) that find loopholes in the reporting of testing data that skew their schools' results into being better than they actually are. Nationally some principals and school superintendents have been fired for doing just that...changing the results so the district looks better and can receive more money for 'improving' test scores.

Good luck with finding topics... I probably vented more than I did assisting you, sorry.
posted by bach at 11:58 PM on January 21, 2011


Best answer: For native speakers of English: Does studying foreign languages earlier correlate to higher grades in English later?
posted by mdonley at 3:54 AM on January 22, 2011


Best answer: @August West, Teach for America is doing the first study you suggest and releasing some data as they go. (One confounding issue with such a study is that there are typically several ways to be an excellent teacher.)

The second study you suggest is released by major metro newspapers most years when the new school testing data comes out. But we've known for decades that a) impoverished students need considerably more money spent on them to come close to the same outcomes as wealthier students; and b) even in the same school, impoverished students will typically underperform their wealthier counterparts because of the problems they face at home. Home price is an excellent proxy for parental education (matters a LOT in student achievement), housing stability (matters quite a bit), low lead poisoning levels (matters a lot in my personal district where lead poisoning is among the highest in the nation, but probably not a lot in an average district), high levels of early intervention for special needs students, decent nutrition (matters a metric assload), etc. It's not the money itself that matters. It's the home environment that comes with having a certain amount of money, including the educated parents.

@bach, there's a recent study, and now I can't remember where I saw it, showing how students who have more content knowledge (social studies, geography, history, great literature, etc.) actually do better even on the bubble-in tests. And how reading scores have gone down since instead of reading actual literature (including kid lit; they don't have to be doing Oliver Twist in grade school), kids now read "teaching texts" that are stripped down, boring, and intended to teach a particular reading skill. You might need more specific curricular data than the OP is going to be able to get, though; you'd have to see whether a particular district's reading curriculum includes "great lit" or whether it's all "teaching texts"; you won't know that from just looking at class names. You'd also have to know whether a school is doing "reading across the curriculum" (where they read in ALL their classes ... some curricula try to REMOVE as much reading as possible from non-reading classes to focus on JUST the skills for that class), which you'd basically have to investigate case-by-case. Etc.
posted by Eyebrows McGee at 6:52 AM on January 22, 2011


Best answer: I think it is critical to first agree on what is the goal of a school or schools in general. This is not as simple as give them a good education or a well rounded one or have them meet minimal standardized test scores.

I can tell you that New Trier High School on the North Shore of Chicago has a different metric by which they measure than the district in southern Illinois that EB is a member of the school board.

Devise a metric for determining if a school is successful or has met its goals. For example, the public school my child attend seeks as its mission to create critical thinkers who can think deeply and use knowledge acquired rather than simply acquire the knowledge. Is there some sort of metric that can tell me if the school is successful? Perhaps it is the grades a student gets in their first year of college or what job they have directly out of high school.

I would love to see if there was a way to measure how expectations effect performance. Do the higher expectations and pressures that come from high performing districts help their students reach higher scores than a school where the goal is simply to get a diploma. Does expectation and environment effect results as much as intelligence?

Is there some sort of data that is predictive of future success? Can you determine that if a child gets some sort of score on standardized tests in the first and second grade that they are likely to have x scores on their SAT or ACT? If yes, then you can determine if they reached their potential. If they exceed the predicted score is it because of the curriculum, teacher, parents help or something else? Can you determine if a school helps all of its students reach their capabilities?

Is there any difference in performance between students who qualify for a free lunch and those who do not in the same school? If you can determine that scores or success is somehow based significantly on environment, then maybe it would be better for the students if the school hired a social worker rather than another math teacher.

I am sure there are studies out there that come down on both sides, but I would still like to look at class size versus performance. Is there a tipping point in class size over which all the students suffer and are smaller classes really materially beneficial?

If I had national data, I would look at performance by grade versus school configuration. Do 5th graders that are housed in elementary schools do better, worse or the same as 5th graders in a middle school? I would also look at school start times versus performance.

The number one thing I would want to know is if there is a correlation between teacher and standardized test score. If you could track student performance for teachers that have taught in more than one district and in districts that are different in terms of socio-economic levels and see how scores compare.

Putting aside all the random ideas I have here and in my previous post, I think schools need to teach or facilitate a different kind of learning than that on which such things as Race to the Top is measured. Teaching content and drilling for tests that measure your memory is a waste of time for a 21st century worker. All the world's knowledge is on a smartphone and googlable in a few seconds. Or come ask a question on AskMeFi.

We need to teach our children how to apply and analyze data and information. The very task your friend is doing with this data. Here is a data set. WHat can we learn from it? That is what kids need to learn. Be critical thinkers. That will require different structures in schools. Sitting in 5 rows 5 deep makes no sense. Kids should learn to work in teams collaboratively. Teachers need to relearn or adapt their teaching styles. Use and develop Wikis as a class. Kids need to help each other, challenge each other and question each other. They need to become active participants in their own education not just passive learners.

posted by AugustWest at 10:22 PM on January 22, 2011


Response by poster: @AugustWest, those are fantastic and exactly what I'm looking for. Thanks!
posted by ttyn at 10:57 PM on January 22, 2011


Best answer: I want to know if there are deleterious effects to _not_ doing enrichment or differentiated instruciton for what my town calls "high-end learners" (no, I am not kidding about that), somethign that my lcoal elementary school principal accepts as a sad but normal state of affairs.

You can look at system-wide effects or what happens to individual students or classrooms or cohorts, I just want to know the scale of the cost for leaving these kids to drift while concentrating so heavily on teaching-to-the-test and kids with IEPs.
posted by wenestvedt at 11:23 AM on January 24, 2011


« Older Back pain and swollen lump at base of spine   |   Bioidentical HRT M.D. in Austin, TX Newer »
This thread is closed to new comments.