Follow by Email

Friday, October 10, 2014

Homework-performance correlation

Good article available by link in today's Washington Post by Alfie Kohn on the studies relating to the impact of homework on learning/performance in kids.  The article came out in 2012, but for some reason there's a link in today's paper, and only as I was writing this did I realize that it was a 2012 article.  I might have blogged about it back then for all I can remember.

Anyway, the upshot is that by now it's agreed that homework does absolutely no good in all of elementary school, but now the dispute is whether there is any benefit of homework in high school.  And a 2012 study by Adam Maltese et al was trying to say that there was some benefit.  Here's a cite and link to that study:

Adam V. Maltese, Robert H. Tai, and Xitao Fan, “When Is Homework Worth the Time?  Evaluating the Association Between Homework and Achievement in High School Science and Math,” The High School Journal, October/November 2012: 52-72.  Abstract at

Now that I think back, I didn't have any homework in elementary school, and I did fine.

Here's the abstract of the study:

"Even with the history of debate over the merits of homework, there are significant gaps in the research record regarding its benefit to students. The focus of this study is on the association between time spent on homework and academic performance in science and math by assessing survey and transcript data from two nationally representative samples of high school students collected in 1990 and 2002. Using multiple linear regressions and controlling for students’ background, motivation, and prior achievement, we investigated how much variance in science and math course grades and achievement test scores could be explained by time spent on homework in those classes. The results indicate that there is no consistent significant relationship between time spent on homework and grades, but a consistently positive significant relationship between homework and performance on standardized exams."

The article reaffirms my belief that people need to have licenses before they do studies.  What a waste of time and money for no useful results at all! 

I would have to read more closely to understand how they controlled for background, motivation, and prior achievement. 

As Kohn describes it, the studied focused on math and science homework, and it relied on two large preexisting datasets -- from the National Education Longitudinal Study [NELS] and the Education Longitudinal Study [ELS] -- in which thousands of students are the single question of how much time they spent on homework per day.  From that, the authors apparently are able to use statistical tests to determine whether there's a correlation between that number and student grades or standardized test performance.

Kohn point out an immediate problem that the authors seem to sweep under the rug.  One of the studies came back with "37 minutes" and the other came back with "60 minutes", even though (apparently) the same question was asked of the same basic dataset.  I haven't read far enough to understand how the authors did or didn't deal with this.

But the bottom line of even this study is that it found no correlation between homework time and grades, and it found an extremely tenuous connection between homework time and a tiny increase in standardized test scores.  Somehow, the authors still seemed to think that the study supported assigning homework (just doing it smarter), but Kohn disagrees.

Anyway, I find the whole debate to be very off-base.  It's a bunch of people who are not scientists trying to do science.  Real scientists are pretty bad at this kind of science themselves -- just witness all the debates about clinical studies in the pharmaceutical area, not to mention all the cases of falsified data, selection bias, etc. etc.  But the bottom line is that this is just a stupid way to approach the question of whether homework benefits kids or not. 

One of the major problems that I don't think this study can possibly answer is the simple fact that "slower" kids need more time to figure things out.  If everyone is really "doing" the homework, the students who do MORE of it might well be the slower kids.  Obviously, a lot of slow kids give up and do less.   But that's my basic point -- some kids are naturally faster than others at picking things up, and doing stuff like homework.  That shouldn't be controversial -- a large percentage of kids are diagnosed with ADHD.  An untreated kid with ADHD might take several times as long to do a simple homework assignment as another kid.  Either that or he or she won't do it.  Also, in my life, I've encountered some very impressive people who were clearly slower than me, but took more time to figure things out, with the result that they were just as successful (or more) in school than me. 

And we all know kids who were extraordinarily quick -- kids who could get high every day and still pull in straight A's.

So none of this high level statistical crap is going to get us anywhere with this problem.

The only way to really do it would be to get in the minds of the kids themselves.  I.e. follow them in school, figure out how they learn, etc. etc.  The best way to do that is for those of us who are reasonably self aware to think back on our own experiences, and then report back.  We are the ones who are best able to figure out how to control for different variables.

I myself was reasonably smart, but probably had a kind of ADD that caused me to get distracted on homework such that when the going was tough, it took me a long time to work things out.  The same ADD probably caused me to miss key points in class, that I had to make up for at home.  And as a result I often got help from my older brother on math, physics, and chemistry.  As I think about it, I probably would have been much better served if the school had just had a mandatory study hall of some sort, where students did their "homework" under supervision/tutelage of more advanced students.  In other words, there would be a set time for doing the homework, and the kids would work during that period and at the end of it would be done.  And then the "advanced" students could report back to the teachers on what concepts kids were having trouble with. 

Anyway, I'm getting a bit distracted from my topic right now.  What I'm saying is that for me, homework was not an efficient use of time.  But I do remember that there were many times when I didn't understand what was going on in class, and it only became clear through doing the homework, often with the help of my brother.  So at least in my case, the homework was an integral part of the classes -- if I hadn't done it -- or had done less of  it, I would not have succeeded in those classes.  In fact, I would have failed them.  But I'm sure there were other kids who understood what the teacher was saying all along, and could do the homework in just a few minutes.

And of course my nature was that I was competitive enough that I wanted other kids to think I was smart, so I didn't want to show up in class not knowing anything.  And if someone had asked me how much time I spent on homework, I probably would have lied and underestimated, with the idea that if I gave a real number, the questioner would think I was slow.

From the discussion above, you can see that there are just too many variables going on for the question "how much time do you spend on homework" to do any good in trying to figure out a correlation.  Still, I think it's a question worth exploring, but really, as I just said, it ought to be through a really careful and informed "study" of a set of truly self-aware individuals.  Yes, I realize that that's the opposite of a "scientific" study.  But liberal-arts academics have long ago proved themselves absolutely incapable of making any good use of scientific studies.  

Thursday, October 9, 2014

I'm feeling nonplussed

I now know why I never knew for sure what "nonplussed" meant.  It  now has two diametrically opposed meanings.  Worse, it might be impossible to tell which meaning was intended from the context.  

Today's WashingtonPost had an article about a hawk that had collided with a drone.  Here's the end of the article, a quote from the drone operator:

“Equipment is fine. The reason it dropped was entirely because my first reaction was to reduce throttle to reduce any risk to the hawk. It fell straight down,” Schmidt said in an e-mail. “The hawk seemed completely non-plussed; he flew off without any signs of damage.”

Here's what Google gives for the definition:



1.    1.
(of a person) surprised and confused so much that they are unsure how to react.
"he would be completely nonplussed and embarrassed at the idea"

2.    2.
(of a person) not disconcerted; unperturbed.

So was he completely unperturbed, or completely surprised and confused?  I'll never know - the video stops on impact.

Lesson:  Stay away from this word.

Friday, September 19, 2014

Washington Post in Need of Editors

At 10:18 yesterday morning, on PostEverything, Elizabeth Hawksworth, a Toronto-based writer, posted an article making the claims that "It’s estimated that 1.5 students out of every 100 will commit suicide at some point during their college career," and that "Suicide rates among college students have increased by 200 percent since the 1950s."   

For the casual reader, both of the cited statistics might have been alarming. Thankfully, they are both wrong.  As of this morning (i.e. the next day) at 7:10 a.m., there has been no correction of either. 

Don't get me wrong -- Ms. Hawksworth has an important message to convey, based on her own experience.  Articles like this are worth reading.  But they should be accurate.

The first statistic -- 1.5% out of 100 will commit suicide -- if true, would make college more dangerous than a military combat tour of duty (ok, I don't have statistics on that, but I'm pretty sure fewer than one in a hundred have died in recent wars).  If you click on the link and scroll around a bit, you find that the 1.5% relates to suicide attempts, which is obviously going to be a much higher number.  The link doesn't give a statistic for how many commit suicide "at some point during their college career," but it does say that there are 1100 suicides at colleges per year, which is 7.5% out of 100,000 students (per year).  We'd have to have additional information to convert the "per year" figure to "at some point during their college career," but in any event, it's clear that the actual suicide rate is much lower than 1.5%.  Still, 1100 is a lot of kids, and that's something we should all worry about.

The author links to a secondary source -- for these statistics, and cites additional secondary sources.  So it's hard to tell if the statistics they report are even accurate.

Some of the commenters noticed this problem, but those comments were not at the top of the comment list, so casual readers would have missed them.   

As for the second statistic -- that suicide rates among college students have increased by 200 percent since the 1950s, that sounds very believable to me, but that's not what the linked source says either.  The linked source (something with the "American Association of Suicidology" logo) says:  

"8. Youth (ages 15-24) suicide rates increased more than 200% from the 1950’s to the late 1970’s. From the late 1970’s to the mid 1990’s, suicide rates for youth remained stable and, since then, have slightly decreased."

So that's different on two counts -- first, it's not talking about college students, it's just talking about "youth", and second, the actual statistic makes it clear that the 200% increase occurred 35 years ago (between the 1950s and late 1970s) and that the rate has "slightly decreased" since then.  

But to add to the confusion, this statistic is inconsistent with the statistic presented in her other source, which says that suicide rates in the 15-24 age group have tripled since the 1950s.

It would be very interesting to have some honest statistics about college students here, e.g. how suicide rates among college students differ from non-college students in the same age group -- I had heard that the prevalence of depression among college students has escalated, simply because today there are medications available to control depression.  But I don't know which way that cuts -- if there are more depressed college students, maybe that should mean more suicides, but if the medication is working, then maybe not.  But we're not getting helpful statistics here on PostEverything. 

Suicide is serious and tragic, and public awareness of the problem is important.  But let's keep the statistics straight.

So what exactly is "PostEverything"?  If they are simply in the business of posting everything, then perhaps I shouldn't complain.  But the "Post" appears to stand for Washington Post, and the "everything" just seems to be a reference to the scope of possible subject matter.  According to the  kickoff description by Adam B. Kushner, PostEverything "is an attempt to expand the conversation outward," to topics like"

"Should we worry about the robot takeover of U.S. jobs? Ask an economist. What are some of the dumbest things people think about American foreign policy? Ask a political scientist. How do football teams draft prospects, what does it feel like to confess to atheism in a deeply religious place, is Russia really seeding Crimea with more Russian citizens, and how did university sexual-assault policies get to be so daft? Ask the people who are in a position to answer."

I still think they have a responsibility to read what is posted, and get the author to make corrections as necessary.

Friday, September 12, 2014

Tony Robbins Firewalk Scam

I have to admit that even though I don't like or really believe in Tony Robbins, I honestly believed the idea that doing the fire walk -- walking over hot coals without getting hurt -- was an issue of mind of matter.

Maybe that's because I'm a gullible idiot.

Today I just happened upon this video.  This reports on the July 2012 San Jose firewalk incident in which several people suffered burns and may or may not have been hospitalized.  The initial reports (including from this video) were that there were screams of agony and that some people suffered third degree burns.  Robbins' defenders quickly jumped in and claimed that nobody was hospitalized and nobody suffered third degree burns, and that it's not unusual for some participants -- maybe 1% -- to experience some blistering.

I'm not here to judge that dispute -- I'm perfectly willing to believe that nobody was seriously hurt and that the initial news reports were exaggerated.

The video showed two Robbins acolytes talking about how the firewalk was a demonstration of mind over matter.  A contemporaneous Fox video featured a woman recalling that Tony Robbins had spent about an hour going over how to do the firewalk -- walk at a normal pace, don't look down, and keep repeating "cool feet" or something like that to yourself.

Other Tony Robbins' apologists managed to turn this thing into one more big advertisement for Tony Robbins.  Here's a paean to his greatness from Marianne Schnall, in the HuffingtonPost.  Marianne did the fire walk on a previous occasion and adores Tony Robbins, so she knew immediately that the news reports had to be wrong.  Apparently Arriana Huffington is an acolyte and firewalk-believer as well; Schnall quotes her as saying in an email: "It was a powerful experience of the inner strength we have to create the lives we want, not the lives we settle for -- an inner strength greater than we often give ourselves credit for. And my tiny blisters were a reminder of that!"

But what I learned from the ABCnews video is that the whole firewalking thing is just a sham.  Or maybe a scam.  There were clips of several people saying that it had been debunked, including someone from "Mythbusters" as well as Steven Salerno, author of "Sham," who says that coal is not a good conductor of heat -- if you get the temperature right and are walking fast enough, you won't feel a thing.  I guess it's like moving your finger through a lit candle flame -- you can do it without feeling a thing, but if you slow down too much, or hold your finger above the flame, you'll get burned.  If you've never tried that, go ahead -- it's cheaper than a Tony Robbins seminar.

So I'm not blaming Tony for causing people to get burned.  I'm blaming him for causing people to believe that they were NOT getting burned because of some "mind-over-matter" principle.

Don't get me wrong -- I absolutely believe in "mind over matter" in any number of contexts (that's the placebo effect).  Just not in firewalking.

So that's one more reason I don't like Tony Robbins.

Tuesday, September 9, 2014

Charles Severance and Leo Tolstoy

In an odd coincidence, Google is today celebrating Leo Tolstoy's 186th birthday, with a nifty Google Doodle celebrating some of his better works:

And today, the newspapers have also announced that former political candidate Charles Severance has been indicted in three murders of prominent citizens of Alexandria, Virginia.

I do not wish to make light of these murders, which were brutal and senseless in every sense of those words.  The Washington Post article itself does not clearly describe a motive, but one of the comments does, so I'll reproduce that comment in full, without vouching for it in any way:

9/8/2014 10:57 PM EDT
Severance had a long-running beef with Alexandria law enforcement and courts related to his child custody case. Inter alia, Severance has a history of mental illness, which warped his perceptions of everything that was happening. Dunning's husband was the Sheriff; Lodato's father was the Judge. I think there might have been a separate beef with Kirby. (WashPost and other local media ran a few stories detailing these connections shortly after the Lodato murder. You should be able to research it.)

If that's correct, and if everything else in the article is true, there's a pretty good circumstantial case against Severance.  Will be interesting to see how this plays out.

Anyway, I just felt that somebody had to point out the coincidence.  No offense intended.

Saturday, September 6, 2014

The cruelty of owning a pet snake

If you've ever been inside a non-name-brand pet store, you might have wandered by a cage full of little rodents.  Mice or rats, mostly babies.  Crawling on top of each other.  Barely room to breathe.  And perhaps a few displays away there are also mice and rats, for sale as pets, in more comfortable surroundings.  What's up with the overcrowded rodents?  They are feeder animals.  They will be bought by snake owners, and fed to the snake.  They live their lives out in filth and discomfort, and end it between the jaws of a small-brained reptile.

I don't know why people own snakes as pets.  I have to think that most snake owners are trying to compensate for something.  There is no affection in a snake.  There is no shared understanding in the gaze of a snake.  They are reptiles.  They do no know us, they do not understand us, they do not think like us.  The common ancestor of humans, mice,and rats lived approximately 65 million years ago (sourcesourcesource).   And it looked like this:

(Photo Credit:  Carl Buell)

That animal (Protungulatum donnae) was thus your great-great-great-etc.-for-a-while grandmother, it was also a great-great-great-etc.-for-an-even-longer-while grandmother to today's mice and rats. Technically, therefore, we, mice, and rats are cousins, several times removed.

The inquiry into the common ancestor of snakes and humans is basically ungooglable, mainly because nobody in their right mind cares.  In fact, we know that the common ancestor lived more than 315 million years ago, because by that time, there were already synapsids (an umbrella group that includes our ancestor mammals but not reptiles) running around.  So the difference is at least 250 million years worth of generations.

Our great-great- . . . - grandmother on our "snake" side might have looked like this Crassigyrinus (credit:  Dmitri Bogdanov)

Or possibly like this cacops (credit:  Dmitri Bogdanov):

You can look a rat in the eye and empathize with one another.  Some people think that a rat has about the same intelligence as a dog (of course, some people don't).  And rats can do some pretty cool tricks.  Although mice are not as intelligent as rats (at least in the way we humans tend to measure intelligence), you can look into their eyes and reach an understanding, of sorts. Can't do that with a snake.  Snakes are scaly and alien to us.

And there's no snake in this picture.

So if you're thinking about getting a snake, just don't.  There will always be people in this world who think nothing of feeding mammals to reptiles.  But if you haven't thought about it in these terms, and it disturbs you, then you and I are alike.  Just get a pet rat instead.  Maybe even choose from the feeder rats.  Sometimes these are actually just unwanted former pet rats.

You won't regret it.  Rats are much cooler than snakes.

Note at the beginning I said "non-name-brand".  That's right.  PetCo and PetSmart are a bit more humane -- they don't sell live mice and rats as feeders.  But they will sell you pre-killed mice and rats, I'm told.  So that's not really all that much better, is it?

Wednesday, September 3, 2014

Over-enforcement of traffic laws ruins lives

I just find it an annoyance -- every few years I inadvertently go a bit too fast and I get pulled over.  Maybe once every 50,000 miles.  And recently, I've gotten two photo-enforcement tickets.  The result is an annoyance -- I have to pay some money -- I think the worst has been $150 or so, and my insurance rates go up, even though I've never had an accident and never filed a claim.  So it's annoying and it's unfair and it's inefficient.  If I'm speeding, it's typically very temporary and I'm not going abnormally faster than anyone else on the road.  There's no danger.  But somebody has to make money, and it's the every-more-prevalent urge to make money without actually MAKING SOMETHING or somehow making the world a better place.

I acknowledge that this is a tricky issue.  I actually think on the whole, the idea of having cops patrolling the highways isn't such a bad thing -- they can be there to help when there's an accident, and they can stop the truly dangerous drivers.  And yes, it's also important that the speed limits be enforced, at least to some extent (if you're going more than 15 mph over the speed limit, you're risking getting caught).  So even though I'm annoyed by what I see as over-enforcement, in the end I have to admit it's my own fault -- I just need to be more careful 100% of the time.  If the alternative is some kind of automatic enforcement -- like photo enforcement -- such that everybody gets a ticket when they go above a certain speed, I tend to prefer extra "chances" I get with human police.

But today's WashingtonPost article by Rodney Balko reminded me that once again, what for me might be an annoyance is for many people a devastating and life crushing defeat.   The article is extremely long and full of anecdotes, but the upshot is this:  poor people get pulled over a lot more than rich people,  I'm talking about what defense lawyers in Missouri call “poverty violations” -- driving with a suspended license, expired plates, expired registration, and a failure to provide proof of insurance -- all violations that tend to occur if you're poor and have to make hard choices about how to spend your money.  Anyway, the poor tend to rack up these violations, and then they can't afford the fines, and eventually warrants issue, and the poor are thrown in jail.  All because of a few traffic violations.

I don't have a solution.  I think some of these "poverty violations" are problems -- cars should have current registrations, people should have insurance etc.  But there's something wrong when people get thrown in jail over this stuff.