kitchen table math, the sequel: 24/7

Sunday, October 24, 2010


Sorry to be AWOL...I am CHAINED TO MY DESK!

(Teaching 2 courses I've never taught before, both of them composition & literature, which means lots of paper reading and lots of reading-reading.)

I'm doing no math at all. None. Not even the SAT Question of the Day.

I'm spending so much time on English literature, fables, folk tales, and short stories that I think I'll just start posting some of the work I'm doing there for the time being.

For starters, here's a fantastic site for English students:

Getting an A on an English paper


VickyS said...

Me too, Catherine! I've been teaching high school Biology and Economics this year--in addition to the day job! It is incredible how much work it is--too much to keep up here, unfortunately. Maybe I'll be back online next summer to tell you guys all about it.

LynnG said...

Biology and Economics? In addition to a day job? Wow!
Can't wait to hear about it.

Glen said...

Okay, during the lull, maybe I can ask a question. Does anyone know of statistics comparing math performance in East Asia with math performance among East Asian-American kids? It's too easy to dismiss a Japanese vs. "average American" comparison, but since Japanese, Korean, Chinese, Taiwanese, Hong Kongese, Singaporean math results cluster so tightly, we ought to be able to compare the cluster to East Asian Americans and learn more than if we compared them to "average Americans".

LynnG said...

I haven't seen any data specifically like that, Glen, but it probably exists somewhere. I HAVE seen data that shows East Asian-American kids doing far better than their non-East Asian-American peers, but I don't think that is what you are asking about.

You might poke around on the National Center for Education Statistics website. I've found a lot of useful stats there in the past.

Glen said...

Thanks, Lynn

Debbie Stier said...

Thanks for the site on getting an A on an English paper. Just sent to my son (via Facebook by the way, because that's the only place I know I can reach him, for SURE).

VickyS said...

Glen, most of the Chinese kids I know (and I know quite a few, having been involved for many years in our local Chinese-American community) go to "Saturday school" where they study two subjects: Chinese and math. They also do a lot of Kumon. If this holds for other regions of the country, then this population "after schools" a lot more than the the average American and I'm not sure a comparison between the Asian-Americans and the Asians would say much about the US math system per se.

Glen said...

Vicky, that's a good point. Around here (Silicon Valley), the Chinese shuttle buses pick the kids up every day from the schools in our neighborhood and take them to their afterschool Chinese/math programs. That makes them even more similar to the kids in Asia (going to their English/math programs), so if the Asian kids here still couldn't keep up with their cousins in Asia, it would strongly suggest that something important (not genes, not class, not family structure, not afterschooling...something else) was lacking.

Of course it could also be that they would do just as well as their cousins. Or maybe the results would depend on what was included in the test. I don't know.

I also wonder about diminishing returns effects. It could be that one high-quality hour of math (or maybe somewhat more of mixed high-quality and junk-food math) a day pushes most kids forward as much as they can be pushed in one day. It could be that a good afterschool program could make the school math program quality almost irrelevant--the kid's buffers are full for that day. If so (and I don't know), then Asian kids with a good afterschool program here could match the results of their cousins in Asia, not because the school program here was just as good but because, under those circumstances, school math would hardly matter. (It WOULD matter for less-fortunate kids without such programs, of course, but the school would see such results as evidence that the school program was fine, and the problem was with certain kids.)

Anonymous said...

"I also wonder about diminishing returns effects. It could be that one high-quality hour of math (or maybe somewhat more of mixed high-quality and junk-food math) a day pushes most kids forward as much as they can be pushed in one day."

There almost *HAS* to be some sort of diminishing returns effect at some point ... I get tired of thinking eventually.

What isn't obvious one way or the other is how close the various kids are to that point (and, pretty obviously, it isn't a clean cut line, but rather a matter of the first hour is 100% useful, the 2nd hour is only XX% as useful as the first, etc.).

One high-quality hour of math for a 4th grader strikes me as a LOT of thinking. Having said that, Kumon isn't quite what I have in mind for "high quality." Kumon is great for drilling speed, but you have diminishing returns here, too ... once you are at 80-90% of whatever your limit is, the additional drill isn't going to speed you up as much as the earlier drill time did.

My idea of "high quality" would be more in the line of some highly skilled mathematics teacher working with the kids on age/grade appropriate math olympiad type problems, rather than spending another ten hours moving the accuracy/automaticity of an algorithm from 92% to 93% of some maximum.

But ... if the tests care a lot about speed/automaticity on basic/core algorithms then I suspect that one could keep getting value out of additional hours.

What do the tests that we would use to compare the populations care about?

-Mark Roulo

Glen said...

What do the tests that we would use to compare the populations care about?

Someone else will have to answer that, but as for what we SHOULD care about, I think we should use adaptive testing (tests that choose subsequent questions based on your answers to previous questions) to explore a much larger space of mathematical capabilities, so we can find out more about who knows what.

I've complained that our Calif. std tests are so limited that a Stanford physics professor wouldn't be able to outscore a careful third grader, because careless arithmetic errors count against you, but no knowledge beyond simple arithmetic can count FOR you. The tests are blind to abilities, such as solving tough word problems, that kids in Singapore spend years developing.

Adaptive testing could tell us things such as that Group A can't reliably do single-digit multiplication in their heads (and that this ability turns out to correlate surprisingly strongly with ability to handle fractions), or Group B is amazingly good at contest-level word problems, or Group C includes a surprisingly large cadre of kids who can do math far beyond their classroom level (so, must be getting it outside of school, which the questionnaire shows they are), etc.

Crackpot theories can explain low-quality observations as well as good theories can. Until we have higher-resolution testing, I'm afraid we'll be stuck with crackpot theories driving educational practice.