Monday, February 16, 2026

The harder-problem fallacy (which is about to turn into related once more)


If you happen to’re making an attempt to differentiate between totally different college students’ ranges of understanding—notably in conditions the place data can presumably substitute for reasoning and comprehension—merely making questions tougher will seldom assist and can typically do exactly the alternative. For instance, if a Math Olympiad fashion take a look at switched from geometry inquiries to trigonometry questions, the examination would primarily be good at figuring out which college students had taken pre-cal. 

In these instances, a well-designed take a look at will discover a manner of leveling the enjoying area in order that extra info and coaching is not going to give one individual a bonus over one other. The most effective examples of that is the outdated SAT reasoning take a look at, earlier than David Coleman—The New York Occasions darling—“fastened” it.

An outdated English professor of mine (who, not fully coincidentally, launched me to Raymond Smullyan) precisely described it because the hardest ninth-grade math take a look at you’ll ever take. By way of data, it didn’t require something past Algebra I and some actually fundamental geometry ideas that had been helpfully supplied on the primary web page of the take a look at. On high of that, types of notation had been invented in order that the coed who hadn’t taken a math course for a 12 months or two was on a kind of equal enjoying area with the child who was effectively into the primary semester of calculus. 

Again in 2014, we talked about how the SAT labored across the harder-problem fallacy (although not by that identify) and about how the reporters overlaying the take a look at (which was on the day trip of trend with the NYT et al. earlier than shifting once more) saved lacking the purpose.

As you may have guessed, we’ll be connecting this to our AI thread in a couple of days.  

Maybe we should always add “opaque” to the checklist of journalists’ vocabulary questions  

Final week, Andrew Gelman criticized Todd Balf for choosing phrases and phrases for his or her emotional connotation relatively than for his or her precise that means in his New York Occasions Journal article
on the adjustments within the SAT. ‘Jeffersonian’ was the precise time period that
Gelman choked on. I might add ‘opaque’ to the checklist although the blame right here
primarily goes to David Coleman, president of the School Board and fairly
presumably probably the most highly effective determine within the training reform motion:

For the School Board to be a fantastic establishment, [Coleman] thought at
the time, it needed to come clean with its vulnerabilities. … “It’s a drawback
that it’s opaque to college students what’s on the examination.”

There is a double irony right here. First as a result of Coleman has been a
long-standing champion of some very opaque processes, notably together with
these involving standardized checks,
and second as a result of take a look at makers who routinely publish their outdated checks
and who attempt to preserve these checks as constant as potential from 12 months to
12 months are, by definition, being clear.

This results in yet one more irony: although the contents of the checks are
available, nearly not one of the numerous articles on the SAT
particularly point out something on the take a look at. The one exception I can suppose
of is the latest piece by Jennifer Finney Boylan, and it is value noting that the precise matter she talked about is not really on the take a look at.

Being only a lowly blogger, I’m allowed a bit leeway with
journalistic requirements, so I’ll break with custom and discuss
about what’s really on the mathematics part of the SAT.

Earlier than we get to the questions, I need to make a fast level about
geometry on the SAT. I’ve heard folks argue that prime college geometry
is a prerequisite for the SAT. I do not purchase that. Taking the course
definitely would not damage, however the sort of questions you will see on the examination
are primarily based on very fundamental geometry ideas which college students ought to have
encountered earlier than they acquired to highschool. With one or two extraordinarily
intuitive exceptions, all of the formulation you want for the take a look at are given
in a small field on the high of the primary web page.

As you’re going via these questions, remember that you do not
have to attain all that prime. 75% is an efficient rating. 90% is a good one.

You will hear so much about trick questions on the SAT. Most of this comes
from the take a look at’s deliberate avoidance of simple algorithm
questions. Algorithm mastery is all the time merely an middleman step — we
care about it solely as a result of it is typically a obligatory step in drawback
fixing (and as George Pólya noticed,
when you perceive the issue you’ll be able to all the time discover somebody to do the
math) — however when college students are used to being instructed to issue this and
simplify that, being as a substitute requested to resolve an issue, even when the
algorithms concerned are quite simple, can appear difficult and even unfair.

There are another points of the take a look at that contribute to the status for trickiness:

Questions are written to be learn of their entirety. One frequent type
breaks the query into two elements the place the primary half makes use of a variable
in an equation and the second asks the worth of a time period primarily based on that
variable. It is a easy change but it surely does job distinguishing
those that perceive the issue from those that are merely doing
Pavlovian arithmetic the place the stimulus is a phrase or image and the
response is the corresponding algorithm;

Phrase issues are additionally extensively used. Typically the two-part type talked about above is said as a phrase drawback;

One approach that very most likely would strike most individuals as ‘difficult’
really serves to extend the equity of the take a look at, using
newly-minted notation. Within the instance under, use of normal operate
notation would give an unfair benefit to college students who had taken extra
superior math programs.

One factor that jumps out when us math sorts is how easy the algebraic
ideas used are. The one polynomial factoring you’re ever prone to
see on the SAT is the distinction between two squares.

A fundamental understanding of the properties of actual numbers is required to reply lots of the issues.

An excellent grasp of exponents will even be required for an ideal rating.

There will likely be a couple of issues in fundamental statistics and likelihood:

I’ve thrown in a couple of extra to make it a extra consultant pattern.

We will and may have a lot of discussions in regards to the particulars right here —
I am positively planning a submit on Pavlovian arithmetic (easy
stimulus/algorithmic response) — however for now I simply need to squeeze in
one fast level:

Regardless of the SAT’s faults could also be, opaqueness is just not amongst them. In contrast to
a lot of the devices utilized in our metric-crazed training system, each
this take a look at and the method that generates it are extremely clear.
That is a normal that we ought to start out extending to different checks as
effectively.

 

Related Articles

Latest Articles