Why Math Word Problems Fail
noodle.com
noodle.com
What they want is to do well in school. How they do well is to pass their tests. How they pass their tests is by learning the answers to the questions. Making indirect questions that "appeal to real-life problems" is obviously worse, because the main real-life problem students have is passing their tests.
Don't get me wrong, I'd love for this to not be the case. If students could be assessed on their ability to actually understand and use the material to generate new insights and solve interesting problems, I think mathematics education would be a lot better. But as long as we're assessing students on equation-memorising, let's not pretend otherwise in the way we teach them.
1. From the text of the word problem, remember what section of the textbook it belongs to, and the standard form of equation to be solved in that section.
2. Match the parameters of the problem to the parameters of the equation.
3. Solve the equation.
For instance, "find the maximum" means creating an expression that has the form of a parabola (the only function they learned with an extremum, up to that point in the course), and solve for the vertex of the parabola.
I've always imagined a particular kind of "interactive word problem"—one you might be able to present in an interview, or on a computerized test—where the word problem itself contains no numbers that fit any sort of formula or pattern. Instead, the student is expected to notice that the real-world situation could be solved by analyzing it with the help of a certain algorithm—if they just had certain pieces of data. From this, they are then expected to ask for those pieces of data. The pieces they ask for might not all be available—in which case they're expected to come up with another method. (Also, data can be available that's entirely useless for any method of solving the problem.)
Imagine an episode of CSI: you see the blood spatter on the floor, the bullet holes in the walls. A real "word problem" would just describe those, then leave it to the student to come to the conclusion that they could use some formula they know to triangulate a source trajectory if they could get positionings for those spatter-marks and holes. Really, the asking for that data—the analysis step—is worth 90% of the mark; successfully applying the formula is just a bonus.
Honest question that sounds mean: isn't elementary algebra required for graduation from high school and admission to university? My university didn't offer elementary algebra, and I thought students without elementary algebra skills needed to at least start out in the community college system. In my public school system, the advanced math kids took elementary algebra in 8th grade, and the less mathematically inclined took it in 9th grade.
Where I taught, they gave the incoming freshmen a math test, and the kids were put into Calculus, Algebra, or a "remedial" math course. Algebra was the lowest level course that could be offered by an accredited college, and the "remedial" kids eventually had to pass Algebra.
This was a large state university, so the kids came in with a wide range of skills, from a state where there is a lot of variation in how math is taught in high school. They were bright, affluent kids. Many of them had taken algebra and even calculus, had gotten good grades, but had not really learned those subjects well enough to skip them in college.
I had pretty much the same thoughts as you express, but as the semester progressed, I developed the opinion that these kids had been screwed by the system and still deserved a chance. They were not going to be engineers. Most of my students were planning to major in psychology or business.
Those that do well are the students that understand the game, and play along. They are not learning math, but do well "against" math word problems.
The question as educators is simply, "is this the definition of good?"
Instead, I get word problems like "how long is this task going to take, if we can give you this much resource?" and "how long is it going to take to drive from A to B at this time, given this knowledge of traffic conditions?" and "what's the best tradeoff of time spent travelling against airfare, to travel from here to there, given these limiting conditions?" and so on and so on.
Word problems don't exist? Nonsense.
There is a clear line between problems described with words, and these "word problems". The topic was about math, but physics is worse. My high school physics teacher had a hobby of writing to text book publishers about how "silly" some problems were. They're designed to be difficult and relevant to specific material, but in the process reality goes out the window.
I remember being given problems that were plain wrong, e.g. Q: How many miles would a person walking 3miles/h walk in 2 hours? A: 6 miles. I always asserted that the answer is 6, not 6 miles, since the unit is given in the question. Small errors like this in teaching add up to math that looks like a collection of arbitrary commandments that must be memorized. It's understandable people want to move away from this. But this solution is to fix all aspects of teaching so that everything is completely correct, not to pretend that math allows multiple correct answers.
We should be teaching that you can express a well formed question in words, to which there should be a unique correct answer. Word problems shouldn't be an added layer of obfuscation, but rather a map from the semantics of math to the semantics of the real world, to test that the student understands this mapping. It's a bit like talking to a backend programmer and saying "if a user did X, then they did Y, what would your function return, and is that correct"? It forces the programmer to consider not the internal logic of their own code, but how that ultimately maps to what the user sees.
This seems to be what you're encouraging in your final paragraph. This brings up the question: why is the common perception of common core so different from what common core is designed to do?
But the common perception isn't built off an average or large-scale aggregate, it's built on what people see and hear. A single comically bad implementation can be seen by millions of people thanks to Reddit, Facebook, and the like. A single good implementation will be seen by the students, maybe their parents, and maybe a few more teachers. Most examples people see are comically bad, so the common perception is that common core is comically bad.
How can you claim that I would support requiring a full sentence answer, when I already said "I always asserted that the answer is 6, not 6 miles, since the unit is given in the question."?
I understand the common core requires full sentence answers (among other things) and I think this is wrong. I will repeat my reason. Math is not a different domain to ordinary reasoning, and word problems are not different to ordinary questions. So if "How many miles..." can be answered with "6" as a question in the English language (which it always can) then, "6" is a valid answer in the context of a math question. Requiring more, or accepting less (e.g. by accepting "6" as an answer to "How far..."), is obscuring the nature of math.
Note this isn't an argument against showing work, but it is an argument that, in general, any valid reasoning should be accepted as showing work. The particular method should not be prescribed. But that is a less important point.
To attempt a summary -- the way math is taught in America is broken, and word problems are horribly misused. Because neither the students nor the teachers really understand math, they attribute their distaste for these useless word problems to the inherent nature of word problems, or to them not being "real-world" enough. Word problems are not the problem; Russian word problems accomplish their goal quite well and are not held in such distaste by Russian students. And no, it's not because those ones are more "real-world". Basically, none of this is the real problem, these sorts of attempts to address it will not help, and attempts to address it by teachers and curriculum-writers who don't actually understand math can actually make the problem worse.
But really, I recommend reading the whole thing.
Two planes, which are 2400 miles apart, fly toward each other. Their speeds differ by 60 miles per hour. They pass each other after 5 hours. Find their speeds.
Show that a Matrix A is of rank 1 if and only if A can be written as A = xy^t for some column vectors x and y.
I'm more excited by the second problem. I'm going to use concepts I mostly understand to prove ("if and only if") a concept. In so doing, I'll understand all those concepts better.
In the first problem, it looks too much like I'm trying to tell the teacher something he/she wants to hear: the speeds. It's hard for the reader to see how he/she learns something about arithmetic.
When doing the math that mathematicians do, what you learn is often explicit: "prove that for all..." It's more motivating.
Of course I imagine the makers of Common Core have more expertise and experience in teaching students; I have no plan to reform math education.
I have sensors for certain kinds of information. I would like to be able to process it in order to convert it into other types of information.
That isn't to say the second type of problem isn't important: to be able to say "solution X is the same as solution Y, even though their forms are very different" is powerful when you're trying to analyze and evaluate solutions to a problem. It's just a less common problem to solve.
This is quite the challenging problem in linear algebra, considering that it's not true!
[[0]] = [0]*[0]^t
And indeed, I would say that I've learned math much more successfully when I've had a real world math problem. The trouble is, real world problems tend to be horribly complex.
You're in an awkward position where you don't want to give artificial problems, but you also don't want to give real problems. She recommends this expii site, but the first problem I clicked on (the most recent one) seems ridiculous: https://www.expii.com/solve/11/1
Then it should break down the skills that are needed to solve each of these problems, and cover those skills and the smaller steps of those problems. Finally building to a cohesive solution to the initial problem.
One of the difficulties here is termed "cultural competency". Textbook writers sometimes try to write in "every day scenarios" that end up being much more familiar to middle-class suburban white kids than to lower-class inner city latin-american immigrant kids. Math problems that are supposed to reference familiar everyday scenarios sometimes end up turning into lessons about language or culture.
The idea of "cultural competence" is about figuring out how to eliminate unnecessary cultural barriers -- to make sure that a kid trying to solve this type of word problem is working through the problem, and not getting sidetracked by the fact that they don't have the context to make sense of what it's even about.
It was about budgeting for a couple. The couple had a difficult-to-read name in a language not spoken in that province. Most of the kids couldn't even read past the first sentence without stumbling, because of this. The couple in the problem probably earned about as much as all of the students' parents combined. Both of the people had their own car - while in this village there were about four cars total. The couple had no children, and didn't look after any family, while many of the students had children of their own, and anybody in that village who earned an income had a large family to support on it.
[1] https://s-media-cache-ak0.pinimg.com/736x/be/55/08/be5508f4c...
Good word problems are science or engineering. Once you abstract abstract those away, there is just not a lot of anything actually useful or interesting left.
"12 - 3 = 15"
EDIT: to fix my own math typo. :P
Google search link, sorry the URL stinks:
https://www.google.com/search?tbs=sbi:AMhZZis86j61bP06Kk5rlrdY2JVS0H-D1Xsn5sK08f05G1wSqOdLQ5g8T9U_1yF-w7sv38paVFKOdSVHysQNKwWbK4YDdxZm6MqKqCux3W0dwtWLAZKHw1U8MhpBvinHzxDzAyIouinbahNg99ZAqVBD1jJV-GM--c4QnTZdrhFnOqEBzoEaeHDQuDtqrAudr4p7CHLUuhVGEn-XZYm3Ns4GHlGIvrR5s8bhkocwdS0z0vxdtaZo96q6jygbSB8hh9jQuCMFrPYUY4_1I0R-D8Ie79Kx35Yyy_1MUq_1zPo38apyjf67JWzxPIdyqEmm8Vjhl_1Yn-rj_1Bm9_1JfERtUpViA9XMmn5x8S1e-Zqte1UCnXDVmpA0lUGw2GKyJ4ljKlwBblXwsGEzVbnM1mky-unilzl8UJxXE1lfTstS2StECOyckCd6Q1sribDT5_1QCLcpAbQ6HR1ZKKYk74m9UInbq9Q-oydbkgXu617TJRTkboBgzo3tdXcqEsj5EK-djbD7iCeTWggLgSKyQWZv9wragMzWzMr0N6fPIxwal-V9Aj3zCKkYHxkFXPVPqNHA9RqRqCLnmCLSHDIAL1LgPArxqL97Nc4N9O0KLgyqIkDDc-BizjT4M_1PvoIiwdiAsKiJDj_1QjoYqssEazUalAs_1jwJbj1h3k2qbF-u5ddWVuqb-Xfh98I8G_1nJjZIBsaWTBZo35aS2yT7z2N5NLbZU2GITFtnKQoZeyS0TkRivDhrVUx_1NzlMbnYZwW_1K4Wu5jLKWSgCPztn4TSLWVKfFREc2zdvVX2vsQLKmxkg1M_1mjFEoOWmmp8BHT6qC6_122vjSbT5iPbq5Nx35a-n0fIRhfDTYmcfzTWBqCO4pfp5u3lmFSpSJc4BUf4Zg6ey9o3nEMzpCynNaudsn7JcNvBq17PtZ0rwpKtChCeHoDSqnZszvf0t97tblltV2Qt35Q8Panot0ns2ko5jZBlZQQDZs6kpmo4J1P0Od4dKym-w14G54poln_1SQiIcqx4VLJbSLpkhBa4dYUnmkbYNXCWGjoczmmRR6NRYb8D--or9tvaJyjbqTDEBBzWihYEuDPKShccwNamxoxt0PqFddSAlVEC4V8vGoEyOqluHxzElKqxL3cXCYoajTT8CU1wdU7ALhrAvTX797WOqSuSWLNu8QBW3SSsz9ekMX6MOTD-gFplMBSemfYuD5QBEt_1sUD-scqHz4cwUab6Pv4JajRCUQkNBhyGmYDN8STXCNprUaXZwPEGsQTyW1GjTepCnbsKyA8Hl1TMwiCjVYP5g7QHtxu2853xJMsAEYpjoMf_1F6fvClxcDfYx8DqdPGq5JGBEQpwsCcHqUwyWI-Tzw7xni19StRMKcFlIIUo0SxUn3aSmUgV23sCFBV7pnTgzl16cKopXBlBkFPGq-nu93jfVJvLmBAVE_1ZSOhIiOsfY8eQKBEtCxeZycUJ9zMO0b5CkmwKTGj43UYYyjsercPXWaal6GwheFSgs4iND33oU7Vp8XPl4Bw121R5G6ZzdxSJ0pHEc_1x3xhy1IPV2Hp7HI-iQo8z0GQ01lYp77yr8pQ&btnG=Search%20by%20image&hl=en