Job evaluations

[Balta1701]

Well, that depends on the grade level, but I'd say it's anywhere from 25 for elementary to 80-100 for middle school and HS.

I'd figure a couple move every year, sometimes kids get sick, then you throw out a full year because of disruptions due to extreme weather events like has happened in Tennessee this year, etc.

[StrangeSox]

Trends can be measured in miniuta. It is done 24/7 in trading.

 

The very nature of the data is different and the trends are measured for different reasons.

 

For teachers, you've got annual or maybe biannual evaluations at best. That's a very small set of data until you get to long time periods (because each child is a sub-set of data, so you really don't have 40 unique points in one set), and any "trend" could be overwhelmed by a single outlier.

[southsider2k5]

With an enormous amount of data compared to what you're trying to compile with here.

 

There are millions of kids in this country. As much as I hate the federalization of education, it isn't going away.

[southsider2k5]

The very nature of the data is different and the trends are measured for different reasons.

 

For teachers, you've got annual or maybe biannual evaluations at best. That's a very small set of data until you get to long time periods (because each child is a sub-set of data, so you really don't have 40 unique points in one set), and any "trend" could be overwhelmed by a single outlier.

 

 

There are millions of kids in this country. As much as I hate the federalization of education, it isn't going away.

 

[StrangeSox]

There are millions of kids in this country. As much as I hate the federalization of education, it isn't going away.

 

What's the sampling frequency for a teacher?

[StrangeSox]

Trends can be measured in miniuta. It is done 24/7 in trading.

 

What is the volume of data? How long would it take to compile a similar volume of data for a single teacher? Remember, you're evaluating individual effectiveness, not overall educational standards effectiveness.

[southsider2k5]

What's the sampling frequency for a teacher?

 

You don't need it but once a year. The previous year sets the baseline for each invididual student when they come into the classroom. Being able to set their baseline against every other piece of data that schools have been tracking for decades, such as race, gender, poverty rates, previous test scores, previous grades, etc, you can easily find similar students, and also similar situation teachers and measure them.

[StrangeSox]

You don't need it but once a year. The previous year sets the baseline for each invididual student when they come into the classroom. Being able to set their baseline against every other piece of data that schools have been tracking for decades, such as race, gender, poverty rates, previous test scores, previous grades, etc, you can easily find similar students, and also similar situation teachers and measure them.

 

But this still doesn't address Balta's contention--that you won't see the signal from your personal performance until 2-3 years later, which means you're being evaluated on the [grade level-2] teacher in your school's performance vs the national comparable [grade level-2] teacher performance.

 

You can't measure a signal two years before it appears.

[StrangeSox]

^This is assuming that Balta's claim is correct and that the signal can't be detected in a single school year.

[southsider2k5]

^This is assuming that Balta's claim is correct and that the signal can't be detected in a single school year.

 

I can't imagine that teaching can be done with a completely undetectable knowledge time bomb that doesn't explode for years. I don't buy it.

[StrangeSox]

I can't imagine that teaching can be done with a completely undetectable knowledge time bomb that doesn't explode for years. I don't buy it.

 

There's undoubtedly going to be some signal lag. If you are a fantastic teacher who got a whole group of dead-average kids and got them all really motivated to learn by the end of the year, well, you've been a fantastic teacher, but test-based results aren't going to show the full impact. They may be much better students now, but they haven't had the time to fully exploit what you've given them. Next year's teacher is going to get a group of kids who improved over the last year, but are now eager to get to the top. You won't see that reflected in the evaluation of your abilities, but the next teacher will.

 

I'd like for Balta to provide some references here, but I see no reason to dismiss the claim out-of-hand.

 

 

edit: effective teaching isn't just a measure of cramming knowledge into someone's head, which is really all a test can give you, but also motivating them to continue to learn, expand, explore and experiment. That impact will not show up instantaneously but will benefit them for the rest of their lives, more than getting a few more kids marginally better at algebra.

Edited April 29, 2011 by StrangeSox

[StrangeSox]

I also have no idea how you'd ever effectively measure that impact in an objective manner

[southsider2k5]

There's undoubtedly going to be some signal lag. If you are a fantastic teacher who got a whole group of dead-average kids and got them all really motivated to learn by the end of the year, well, you've been a fantastic teacher, but test-based results aren't going to show the full impact. They may be much better students now, but they haven't had the time to fully exploit what you've given them. Next year's teacher is going to get a group of kids who improved over the last year, but are now eager to get to the top. You won't see that reflected in the evaluation of your abilities, but the next teacher will.

 

I'd like for Balta to provide some references here, but I see no reason to dismiss the claim out-of-hand.

 

Then you see an emergence in the very next year, not years down the road. At the sametime, there had to have been some sort of improvement over the kids who flat out didn't care at all. The reality of teaching doesn't happen with convincing kids to learn finally on the last day. There is a drop dead date in the middle, at some unspecific point, where when they aren't getting it, they quit caring about it.

[southsider2k5]

I also have no idea how you'd ever effectively measure that impact in an objective manner

 

That you really can't. Which is why I hate the federal and state mandates/standards of education, and the evaluations of it.

[StrangeSox]

Then you see an emergence in the very next year, not years down the road. At the sametime, there had to have been some sort of improvement over the kids who flat out didn't care at all. The reality of teaching doesn't happen with convincing kids to learn finally on the last day. There is a drop dead date in the middle, at some unspecific point, where when they aren't getting it, they quit caring about it.

 

Well that was a heavily framed hypothetical, reality will require a ton of signal filtering.

[StrangeSox]

That you really can't. Which is why I hate the federal and state mandates/standards of education, and the evaluations of it.

 

The problem there, to me, is you could get really, really, really bad local standards of education. Thinking of Texas recently and how they wanted to completely rewrite history from a heavily conservative point of view and redefine science to more or less include stuff like astronomy, I would be strongly opposed to city or county-level standards and no higher oversight. Granted, that's a State-level board making terrible decisions, but we've seen plenty of similar stuff like Dover, PA.

[Texsox]

A couple issues that were not address, children are not randomly assigned. Comparing two unequal groups of students is difficult. I teach US history 1492-1865, they will not have that particular subject again. What is the second snapshot? Econ? World Geography?