Predicting H.S. Dropouts With Pervasive Databases
rhadamanthus writes "As seen on the Houston Chronicle: 'With a new computer database available at every campus this fall, teachers can keep a virtual eye on every student and identify those at risk of leaving. For the first time, educators can look up a student's attendance, discipline, immigration status, grades, and test scores at one source and use that information to predict dropouts. ... "All students will know someone is watching them, tracking them, and is interested in their success," school board member Laurie Bricker said at a press conference today.' Hooray for surveillance in the HISD."
You know...
If all the databases with personal information were merged some really really interesting things could be derived asbout you. Think about it. What if your bank thought that you were going to die in 15 years and wouldn't give you a home loan?
Some documents are declared top secret because they combine information available to the puplic in creative ways.
Help I'm a rock.
The intention of predictive models is to find underperformers and work harder to engage them before it's too late.
The reality of predictive models is that it's a self-fulfilling prophecy. If any of the counselors, teachers, receptionists, principals, or aides approach an underperformer with a speech about how they need to buck up before they drop out, all that many kids will hear is, "they know I'm a failure, so why try?"
For a small minority of kids, this gets even worse. We have discussed the profiling it takes to predict violence. This sounds a lot like the same arguments raised which lead to flame-out sentiments like "they know I'm violent, so I've got nothing to lose."
[
Somebody sure is watching and tracking individual students, but they're definitely not interested in the student's success -- collecting all this data together and using it to generate mass "trends" will likely end up in having various kids who are doing well being sat down and had a talking-to by the school's guidance counsellors about not dropping out, merely because they don't fit the trend.
Back in my senior year of high school, we had some sort of tracking system that was based primarily on attendance. It flagged me as a student that was going to fail out, never mind my 3.9 GPA and my acceptance to Stanford based upon the entrance exams (untimately did not go to Stanford because I could not afford the $25k/year). I had a meeting with our vice principal telling me I was in serious trouble with my attendance. What a joke.
Visit Jonesblog and say hello.
I always stayed up until a few hours before I had to go to school last semester. I nearly failed French.
At the same time, I was working on an extremely educational (to me at least) programming project and some web sites.
Would my school's system see my drop in French test scores as a sign of impending doom? Would it correlate that with the departure of Jane Doe, who dropped out due to a pregnancy and accuse me of being the father?
Had I been sidetracked, I never would have had this site of mine on this slasdot article. I wouldn't have gotten a local computer store to invest time/money in my first commercial program.
You can't reduce anything as complex as a human being to mere comprehendable numbers. Anyways, this new system sounds like it'll be great fun to mess with.
(On another note, it's hilarious how schools are scared to put a picture of a student on the school's website without a notorized rights waver, yet they jump at the chance to make a national database of students.)
You can't judge a book by the way it wears its hair.
"These are very tangible and reasonable criteria they're using to make these determinations."
As a part time university teacher I found out when one student was giving me trouble that he had been giving others trouble as well. I was not told of the trouble because that information would have biased my perseption of and treatment of that student. That is an important principle that may be violated here if "teacher" were to get ahold of that data.
Schools administrations would use the data for those things that were most important to them. This may include avoiding lawsuits, eliminating trouble makers. With limited budgets and overcrouded classrooms the insentives would be to diminish classroom size and be able to apply budget to where it would be most effective. You know the current political envirionment is one to privatize or business-tize all activity.
Now with that information would be very useful to at tracking teachers. Lets see, at teacher that has mass defections, well lets get rid of them. Or classes that have certain individuals attending, show scores dropping (trouble makers and cultural disruptors). The data mining capabilities are endless. But of course the adminsitrators would never think of using the data in these ways.
I am reminded of a story where a friend had a meeting with their boss. The boss offered them a project. They said they would like it. It was given to someone else. The reason was the boss said that even though she said yes, her body language said no. The same danger is here with the interpretation or "profiling" of individuals from scan data.
Kill them all and let God sort it out.
A quick HOWTO in turning a democracy into a plutocratic fascist state:
The outcry was initially the collection of the data. We were told not to worry, it is for private industry's use and, besides, we don't have a constitutional right to privacy in business.
Now the outcry is the use of data mining and aggregation to take the data thus collected and use it to profile our private lives in minute detail, effectively creating a defacto, if hybrid, police-surveillence state. And you dismiss it as "they're not creating any data that's not already there", as though that somehow negates the consiquences of such behavior.
The initial public outcry against the collection of private data on private individuals was right then, and it would be right today were it not for the deafening silence of those who have recognize a battle long since lost.
The public outcry against the sale and exchange of data between private corporations (and government agencies) was right then, and it is right today, even if the number of voices has declined over the years.
And the outcry against aggregating and mining this data to microanalyze our individual lives is justified, appropriate, and dismissed at our own peril. This isn't the start of a slippery slope we're talking about here, this is another in a long series of runs down it we're skiing
The Future of Human Evolution: Autonomy