Hacker News Reader

Comment by civilized | Hacker News Reader

Exams may sometimes be disruptive to learning, but they exist for a compelling reason: they help ensure that students exiting the class have actually learned the material.

All students want the credential of having passed the course, but not all students are willing and able to learn the material. If we care who learned the material, we have to verify who learned the material.

To eliminate exams, we would have to either (1) change the world so it doesn't matter who actually learned what they earned credentials for in school, or (2) make all students so intrinsically motivated that they all reliably learn the course material without any need for verification.

Does anyone think that either of these goals is realistic?

View on HN

It's also conceivable (theoretically) to just decouple teaching from examination. Anyone may participate in any class (no entry requirements), but there's a certain bar for passing exams and getting a degree.

To a certain extent, this possibility already exists with textbooks, openly accessible video lectures, etc., but there's also a lot of people who would still benefit from in-person teaching even if they'd never get to the same level of understanding as people who are going to earn a degree.

That's actually a very interesting idea.

What if it changed the game so that instead of exams causing stress and discouraging learning, people were incentivized to learn at their own pace and eventually pass the exam?

This situation already exists with many professional credentials, such as CPA, CFA, FSA, etc. They tell you what will be on the exam, provide practice problems, lots of study materials are available from different sources, and you take it when you're ready.

There's also college exams that have a similar function (CLEP is the acronym I think?) but they're not really highlighted as a preferred path to students.

> There's also college exams that have a similar function (CLEP is the acronym I think?) but they're not really highlighted as a preferred path to students.

That's correct. But not all schools accept it, and if a school accepts it they may not accept all the different exams. (CLEP exams mostly cover general-ed subjects.) Most schools also have an in-house option to test-out of classes, but it may be up to the individual departments to determine whether they want to allow that. Schools may also disincentivise testing-out in other ways... When I wanted to test out of the intro-to-computers course which was, at the same, required of all students, I had to pay the normal per-unit fee (as if I had taken the course normally) plus a special "challenge fee".

I wish this was how college worked; everything is free, except to sit for an exam and then to obtain the degree itself.

Colleges could structure degrees and majors by examination and interview, rather than by "credit hour", which itself is a meaningless assessment of comprehension.

As an example:

"The University of College offers a Degree in Electrical Engineering, offered at the following levels:

- Associate (At least a 60% in the following examinations: Classical Electricity, Circuit Design, ...)

- Bachelor (Meets Associate requirements plus at least an 80% in the folowing examinations: Signal Analysis...)

- Master (Meets Bachelor requirements plus at least a 90%...)

- Doctor (Meets Master requirements plus three original research publications...)

HOWEVER--this would require the total dismantling and reform of the student loan industry, academia itself, and the secondary market that's sprung up around college being a "part of life" for young adults. "College" is a many-billion dollar industry and very few peoiple want it disrupted to the degree that it would require to move to an examination-based format, with the exception of builders and technicians (engineers, etc.) who are actually trying to filter on skill when they hire.

The German system goes with a different take: it is almost free because much of the burden is placed on the student, no TAs to grade homework that isn’t graded anyways, just one final exam to determine your grade for the course. No fancy sport centers or dorms or other facilities that Americans find mandatory but don’t really matter that much.

I know a few Germans--the only issue I have with that system is that they're required to take the class before sitting for the examination. This has two effects; one, it makes a degree vanishingly difficult for anyone that doesn't have a lot of time, and two, it makes it impossible to use your degree to accurately reflect your skill level once you're out of college.

I think it ought to be appropriate for a degree to take decades, with examinations taken and passed at leisure. Most entry-level jobs that require degrees are far easier than college degrees, anyway.

> Most entry-level jobs that require degrees are far easier than college degrees

Germany handles that with their apprentice programs. But for a SWE, I think having gone through the class in a limited amount of time and forced to take an exam at the end, that can be reassuring to some people. Anyways, it still beats the American system, IMHO.

I live in Germany. Yes, education is (almost) free but still not anyone can just enroll, although I think some universities have programs where you can just attend lectures and not do exams. I don't think that's possible for some more prestigious universities (like KIT or TU Darmstadt), though. I might be wrong.

I am in disagreement with the statement that exams provides proof that the student has learned the topic.

In my opinion, exams show which students have figured out how to beat the exam.

There might be some that do this by learning the topic, but in my experience the path of least resistance to a good grade has been to study the exam - not the topic of the course that it's based on.

As an example, I was one of the three people (out of 250 students) in my physical chemistry class that got the highest possible grade. I gained no understanding of the topic at all during that time - what I learned was to recognise which formula to use for specific problems on the test. I don't know what any of the results meant or what said result could be used for, simply that it was correct.

This correlates well with the observations made by Google and other large companies that analyse how to identify valuable hires. Turns out that there's no correlation between grades and workplace performance.

So if the exam result offers no information of the students ability to do a job in "real life" and at the same time introduce incentives that conflict with actually learning a topic, then I think it's better to significantly limit their importance or outright get rid of them.

I haven't been able to find any positive gains from having exams, but if someone has some resources on this, I'd be very curious to see it :-)

I agree that exams don't always do a great job of measuring actual competency, but I'm not sure they are worthless either.

> This correlates well with the observations made by Google and other large companies that analyse how to identify valuable hires. Turns out that there's no correlation between grades and workplace performance.

These observations probably suffer from a confounding effect known as range restriction. Basically, if Google generally only hires people with a GPA of 3.0 or above, they may have some idea of how GPA variation above that level correlates with job performance, but they don't know about it below that level. So they won't know if people with GPAs of 1.5 will do just as well at Google as people they normally consider.

Suppose Google started hiring people with much lower GPAs than they do today. Anyone want to make a bet what the outcome of that experiment would be?

Well, Google knows, and that's why they don't do the experiment.

Google has no requirements for GPA - here's a quote from 2013 by Lazlo Bock who served as Senior vice President of People operations at Google for several years:

"One of the things we’ve seen from all our data crunching is that G.P.A.’s are worthless as a criteria for hiring, and test scores are worthless — no correlation at all except for brand-new college grads, where there’s a slight correlation. Google famously used to ask everyone for a transcript and G.P.A.’s and test scores, but we don’t anymore, unless you’re just a few years out of school. We found that they don’t predict anything."

This says nothing about whether Google has a range restriction effect. It just suggests that their HR doesn't know or care what range restriction is.

Google probably filters out low-GPA candidates based on their poor performance by other criteria. It doesn't mean GPA isn't predictive of how such candidates would perform at Google.

If this is untrue, let them bring out their parade of successful low-GPA engineers.

Where's your evidence that they do have a range restriction? Seems to me that you're moving the goal post without providing any justification for it.

There's plenty of sources online stating Google doesn't have a GPA requirement, that you're welcome to look up on your own time.

Besides those I can also tell from my own (and a couple of friends) experiences, interviewing with Google, that indeed they do not care about GPA - which is a plus when you consider, how many potential hires are not American and have used a completely different grading scheme.

The quote literally states the opposite of what you're claiming, according to Google - GPA does not in any way predict job performance, so why is it so important to you find a narrative, where it does?

What seems more likely - that several Google engineers all made an erroneous conclusion based on the available data or that they found that GPA has zero value in real life?

Ignoring range restriction is a very common and very consequential statistical fallacy. I'm not going to blindly assume that Google engineers are above statistical fallacies. Especially ones that embody universal human cognitive biases, like thinking that what you see is representative of everything that's out there.

There's plenty of evidence that grades matter in other contexts, so I'm not going to throw all of that out because some Google guy made a vaguely general claim based on his limited and very biased sample.

As I said, all Google has to do to refute my doubts is point out its huge population of very low GPA, very successful engineers. If GPA is meaningless, this should be a trivial thing for them to do.

Yes, I do think both of those goals are realistic, although certainly difficult to achieve.

I typed six or seven drafts of a comment trying to approach you from a convincing angle and realized that it was probably hopeless, because we likely have too many conflicting assumptions about society, morals, human nature, etc. to come to an agreement in a web forum.

I would like to extend to you an invitation, and that is to read the following essay:

> The Six-Lesson Schoolteacher by John Taylor Gatto: https://www.cantrip.org/gatto.html

If you find this essay accurate and intriguing, consider sending me an email (address in profile) and I'll send you a large set of essays I've collected that discuss the same principle (i.e. that modern education is harming society more than it's helping).

If you find the essay naive and inaccurate, then there's likely very little that we would agree on at all, and it's not worth a 10-message reply chain where we argue over semantics. (Although I admit I enjoy a comments section rumble, as my post history reveals.)

I'll have a look. We will definitely agree that there is a lot wrong with US schooling but the devil's in the details.

Broadly, there's a lot to like about human nature, but I don't believe in it to that degree.

I like the other commenter's idea to decouple learning from exams, but I'm not sure it would work at all ages, and it mostly just solves the problem that I, personally, had in school: being forced through things I already knew or learned extremely quickly. Others have different problems I'm sure.

I will say Montessori has solutions to some of the problems in this essay and my kids go to a Montessori school.

Reply on news.ycombinator.com