Hacker News Reader

Comment by lumost | Hacker News Reader

I'm skeptical of the tests core result that reading performance has dropped to 1970s levels among thirteen year olds. 13 year olds today almost exclusively communicate through text, whereas in the 70s it was quite plausible for a 13 year old to simply not spend time reading.

I think its worth questioning the test and whether there has been a shift in reading/math skills since the last iteration which de-emphasizes skills the test pushes. Mental math in particular is a skill which was borderline relevant with the advent of calculators and is highly questionable when everyone has a calculating machine in their pockets 24/7.

We should expect that we need to adjust our educational standards from time to time less we continue to teach alchemy, or meta-physics.

View on HN

If you deep dive into demographic groups (https://www.nationsreportcard.gov/reading/nation/groups/?gra...), you get a bit of a different story, showing a lot of the drop might just be Simpson's Paradox at play and the educational outcomes getting worse for the lower performing students.

For instance, non English language learner's reading scores are unchanged over the last 3 decade. It's crashing lower for ELL students.

Dive into other metrics and you see the pattern of the lower performing groups performing worse and better performing groups roughly holding.

Simpson's Paradox isn't really relevant here. All you're describing is the average going down because one group goes down a lot. But that's how averages generally work.

Oh the Simpson's Paradox is that the relative portion of the lower performing groups is increasing as well. For instance, ELL students have increased from 8.1% in 2000 to 9.6% today (https://www.edweek.org/leadership/the-nations-english-learne...)

Simpson's paradox is that several (most, all) groups show an increase, but the average is decreasing.

Imagine a simple chart 1980:

-----------| students | avg-score

poor ...... | 10 ....... | .2

non-poor | 100 ...... | .8

If the chart in 2020 looked like this:

-----------| students | avg-score

poor ...... | 100 ...... | .25

non-poor | 150 ...... | .85

Both groups increased their average score.

But the total average went from (.2x10+.8x100)/110=~.75 to (.25x100+.85x150)/250=~.61

Which is not actually surprising because the dominant change is the migration from non-poor to poor, not the change of the average score - 1980 has only 9% poor results, 2020 has 40% poor results.

When scores change, at first the average scores within groups will change, but if the scores change enough they will eventually cross group boundaries at which point the group averages become more or less irrelevant and the changes are mostly reflected by changes in group sizes. The group averages are bounded by the range of each group and can only indicate if the group is moving towards one of the boundaries, but as scores cross boundaries they start contributing to different group averages and at opposite ends, e.g. decreasing scores will at first contribute at the lower boundary of one group and then at the upper boundary of the next lower group which may increase the average score of both groups, in one case by loosing a low score relative to the group range, in the other case by gaining a high score within the group range.

I was inclined to agree, but the questions seem to be as simple as "add 32 and 14" [1]. I doubt students failed to answer this question on the basis of what field the numbers belonged to :)

That said, the emphasis on rote memorization through multiplication tables has long been utterly useless. So, there is fault in the purpose of instruction as well. As an aside on this notion of rote learning, I recall some of my high school's top students (by GPA) could not solve a simple ODE, despite getting the highest score on the AP Calculus exams (granted, this doesn't mean much when such a score can be attained with a 70%)

[1] - https://nces.ed.gov/nationsreportcard/subject/field_pubs/sqb...

Maybe it's just me, but multiplication tables are one of the few useful things I got out of primary school. To this day I use this often, almost daily. I mean to call them USELESS is pretty severe mischaracterization of them.

I often find myself wishing my mental maths was better in general, I'm pretty dreadful at it and while I'm never more than five seconds or so away from a calculator it's still sometimes a slowdown and context switch I could do without.

(20% serious) Play Sid Meier's Alpha Centauri. A useful thing to do there is to rush-buy improvements, but to spend just enough energy credits that it gets finished the next turn. For example, if it would take 52 credits to purchase a Recycling Tanks, and the base produces 3 minerals per turn, then you want to spend 52 - 3 * 2 = 46 energy, and the game would let you do that. One does quite a lot of arithmetic that way...

Yeah, I'm really confused how someone could see multiplication tables as useless. I'm not sure how anyone could become adept at multiplying without memorizing the relationships between common numbers. Are there grown people out there who add up 8 8's when they need to solve 8*8?

There are grown people who plug 8*8 into their phones

EXACTLY!

"You will need to learn this because you won't always have a calculator with you"

Turns out, you do always have a calculator with you.

A calculator, for me, really doesn't replace the utility of being able to do some basic mathematics in my head. Even if you don't run into a lot of uses for mental math, I'm in the camp that thinks learning 1-10*1-10 and building a sense for numbers will aid in understanding of other concepts and in developing other cognitive abilities. Some carefully designed experiments could convince me otherwise.

It's not about having a calculator. It's about making quick connections when encountering new facts or numbers. Without sizeable core of memorized facts, results and numbers you will miss a lot of those. It's the very definition of being not that smart.

Being able to easily calculate e.g. $/100g to a rough degree without reaching for your pocket is super handy.

I think we could gain from emphasizing math with orders of magnitude, and knowing 1/n's would be super handy too.

I often leave my phone at home to clear my mind and get away from things. I try to be phone free for about 4 hours a day. And during this time I almost always go out. And often I use my multiplication table during this time.

You can skip the whole memorization pain by rigging at most 3 additions together. You just have to know that adding a number to itself is doubling and adding it again is tripling and multiplication by 10 can be done by simply writing zero on the end.

2x = double it. 3x = triple it. 4x = double it twice. 5x = ... 6x = triple it then double it. 7x = ... 8x = double it thrice. 9x = from 10x subtract it once. 10x = put "0" on the end.

That gets you everything but 5x5, 5x7, and 7x7. That's just 3 things to memorize instead of 55. I usually get the 5s by just halving 10x, which then handles every case but 7x7. The overhead of doing 2 (sometimes 3) mental additions isn't much worse than memorizing everything as a singular operation and this way of doing things makes it a bit clearer what multiplication represents rather than being an arbitrary thing to memorize because teacher said so.

To hell with primary school. Memorizing is for suckers.

Congratulations, you now do multiplication like my 8 bit 6502 used to do..

BTW 5x is "double it twice (4x) and add it (1x)"

Personally, for up to 10x10 - I prefer a lookup table (both mentally and in code)

P.S. It's been 30 years? since I had to implement it, but as I recall multiplication could be done by any number with just doubling and adding the initial value.. 2x = double 3x = double +x 4x = double the double, etc. any even multiple be done just by repeating the doubling, and odd multiples are gotten by repeated doubling + adding the starting value.

This worked great for cpu's as 'doubling' was just a bit shift left 'x' number of times - very fast.

This worked great for cpu's as 'doubling' was just a bit shift left 'x' number of times - very fast.

Except that you'd get a summation with the number of terms equal to the number of bits in the source operands. Multiplication in a CPU only seems very fast because the theory is simple. Building a fast multiplier in silicon requires a lot of area and a lot of interconnect, because you don't require "just" a massive addition table, you also need carry lookahead logic to perform fast carry propagation otherwise you get no speed increase.

A naive multiplier in a CPU requires 1 cycle per result bit (multiplying two 32-bit numbers requires 32 additions, then you need to wait another 32 cycles for full carry propagation). That's very slow, in CPU terms.

not quite sure what your getting at here...on a simple 8 bit like the 6502, you were generally dealing with plain binary, not some IEEE standard, so you get at most a single addition (if the multiplicand was odd), and just threw an 'overflow error' at the end if the carry/overflow bit was set... Now arbitrary length BCD...THAT was a PITA

But again, it was 30 years ago so my memory may be fuzzy.

Edit: Hmm - if I recall, it was something like: and the value with x01 to check LSB,

asl value, (multiplicand masked with LSB=0),

if LSB check was true, add value to result,

check overflow flag ?

Again, this was simple binary data usually of a fixed 1 byte or 2 byte length. we used BCD for 'real work' and I can't even remember how that was done...

How do you get to "at most a single addition"? Even when multiplying any N with a 4-bit number like 14 (0b1110), you'd get at least two additions: (N << 1) + (N << 2) + (N << 3). And that is assuming you already have spent the additional lookahead logic to skip the addition step if the source operand has a 0 bit (which most implementations don't do, because the zero detection logic requires even more logic gates in addition to the adder tree).

Ahh - yeah, my bad...

I always remember it being much simpler then it appears it was. Guess I'm getting rusty from using fancy languages with built in multiply instructions :-)

Don't know if I'm the weird one but I always do 5x as "multiply by 10 then divide by 2" you just need to put a 0 at the end and get half of that result which is not a complicated mental calculation.

That actually sounds easier...

I know. And 7x can be done from taking 3x off the 10x, and 8x can likewise save an addition by taking 2x off the 10x I intentionally left the 5s and 7s blank because they are 3-addition operations and also require you to mentally hold on to the very first number. Instead of the overhead of juggling an extra value and doing an extra step, I find it easier to switch around the multiplication terms and handle them using a different rule. The remaining cases of 5x5, 5x7, 7x7 are small enough to effortlessly memorize. Especially since 5x can be done with counting by 5 (which is just normal counting but with 0 and 5 appended to everything). 6 ends up being the only thing in the 10x10 table that requires 3 additions to reach. All the others can be done with just 2 additions/subtractions. For 8 you get a choice between 3 additions or 2 subtractions but with temporarily juggling one extra number.

If you add division by two into the list of primitives, 5 and 6 become accessible within 2 steps and so the system can reach the whole 10s table within a cost of 2 (times 10 is considered a zero cost step). That makes the cost one step worse than traditional memorization but you end up being faster overall because you only need to focus on being fast at the three primitives.

Recognizing the general principle that you can change A into BC or (B+C) where B and C are both easier to multiply with than A is far more useful than memorizing the table. The best part is except for the (optional) division by two, all the primitive operations are things you already learned how to do at that stage. You can short circuit a tedious, long and wasteful part of the early math curriculum AND walk away with genuine understanding of why the rules are and how to figure things out for yourself. While everyone else struggles to multiply by 17, you can be the smarty who doubled 4 times and added once. Or put zero at the end, doubled, and subtracted 3x. Or put zero at the end, subtracted 1x, doubled it, subtracted 1x. With a bit of thought, you can extend the entire multiplication table to 20x20 such that very few rarely does anything cost more than 3 steps.

Tractenberg rules can be used, although I haven't tried it.

1: https://warbletoncouncil.org/metodo-trachtenberg-11215

My job involves 80% rote memorization including a lot of mental math. No being able to calculate simple stuff would slow me down at least 2X. It's a high-pay job.

What do you do, if I may ask? The only roles that I'd characterize as you describe are fixed income management and some actuarial roles

Anesthesia. I calculate dosage, dilutions, hemodynamic parameters, etc. Many, many times a day.

So if someone elbows you accidentally when you're working out a dosage, will you end up giving the wrong one?

Have you elbowed an anesthesiologist?

Yes, especially to distracting and vindicative patients.

"Anesthesia". Well, add a new one to your list.

> What do you do, if I may ask? The only roles that I'd characterize as you describe are fixed income management and some actuarial roles

What? As a software engineer, I'm constantly doing mental math to determine things like "how much memory will this take up when the input size is X". Sure, Big O notation is nice for limits, but in real life, there are typical input ranges, and you have to do the calculation. Can I use a calculator? Maybe sure, but why would I?

I wouldn't really expect them to be able to solve a simple ODE with just Calculus BC. It's been about a decade since I took it, but to my memory, it concluded with basic elements of real analysis after getting pretty far along with single variable integral calculus techniques. We never discussed ODEs as such, but some clever students would probably recognize how to solve them on sight without being acquainted with the broader fields of differential equations

"Add 32 and 14" is from the directions page of the test -- it's not an actual question. It's intentionally easy so as not to distract from the point, which is to tell students how to fill in answers.

You can find sample questions here: https://nces.ed.gov/nationsreportcard/nqt

* 13 year olds today almost exclusively communicate through text, whereas in the 70s it was quite plausible for a 13 year old to simply not spend time reading.*

Comparing the way teenagers text to reading an actual book seems a bit ridiculous to me. Have you read a 13 year old's text thread? The standardized ways of measuring reading performance have much to with vocabulary and comprehension skills. Are you suggesting that if kids can read each other's texts, and maybe Instagram comments, they are in good shape?

No I am not, however I'm old enough to briefly recall childhood life before computers were commonplace. There was a time when I was in elementary school where getting children to read at all was an activity.

They might be in fine shape, I remember back when we had to learn cursive in school because if you couldn't read or write cursive then you might as well have been illiterate. Does cursive have any relation to literacy now-a-days? Absolutely not. Whats to say that being able to read or write instagram comments doesn't fully encapsulate what literacy entails in the 21st century?

I love to read fiction, but for me its just a hobby. Is the ability to read an old book actually all that important? I never actually read any of my textbooks in high school or college and I even stopped buying them because I realized how pointless and expensive they were. I'm well into my career and haven't ever touched a book related to my field. Communicating with my boss through email or chat is far more important than any formal reading comprehension.

I'd say if anything, modern day literacy is more about understanding fragmented conversations, reading between the lines and conceptualizing meaning, understanding context and being able to converse with colleagues including familiarity with memes, recognizing clickbait and being able to sort through misinformation. All of which is are skills the older generations exceptionally lack. Perhaps they are the ones who are actually currently illiterate.

Mental math in particular is a skill which was borderline relevant with the advent of calculators and is highly questionable when everyone has a calculating machine in their pockets 24/7.

As someone with kids, I’d say the public school educators and teachers we dealt with very much mirrored your attitude, and that’s a big part of why scores are down. Lots of soft curriculum, math games, but no emphasis on learning things like multiplication, grammar, vocabulary, spelling.

Reply on news.ycombinator.com