I would say the pro-cheating attitude is pretty widespread among young adults tracking into white collar careers, and computer science is no different. It usually starts in school - and these interview quizzes track this behavior pretty closely. Even without the cheating, you can pass an interview "honestly" by putting in the time to cram for it. That does not justify cheating in any way, though.
The fact that people can just cheat in these interviews and not get PIP'd within 6 months, at an organization notorious for PIPing its under-performers, does illustrate how bad the screening process it. Not only do they need to shift back to in-person interviews, they also need to fix their process to actually test skills.
> they also need to fix their process to actually test skills.
If we're being honest, LC style interviews aren't even about testing skills, it's trying to proxy for testing raw intelligence (or at least the ability to remember and regurgitate lots of facts). If companies could just IQ test people and hire by IQ, there'd be a lot less weirdness in interviews, but that's illegal (and probably rightly so).
If a LC interview effectively serves as a proxy for an IQ test, then I think the courts might treat it the same as a literal IQ test. I wonder if we will see lawsuits testing this hypothesis.
The precedent for the illegality of IQ tests for employment is Griggs v Duke Power Co. [1] if I'm not mistaken. The court ruled:
> The Act proscribes not only overt discrimination, but also practices that are
> fair in form, but discriminatory in operation. The touchstone is business
> necessity. If an employment practice which operates to exclude Negroes cannot
> be shown to be related to job performance, the practice is prohibited.
(apologies for the dated and offensive language, starring it out seemed like it would imply a different and more offensive term so I left it as is)
My reading of it is that IQ tests aren't broadly illegal, but the employer would have to demonstrate that it's somehow related to job performance. E.g. one employer refused to hire people with too high of an IQ, and the courts ruled that was a reasonable (if unwise) justification [2].
I'm doubtful any case challenging LeetCode tests on the same grounds would succeed. I think it would be exceptionally difficult to convince a court that solving coding problems is unrelated to a programmer's job performance.
If they can proven to be so. What employees do is then build a proxy for IQ tests that has some job related stuff. Math or logic puzzles could be used if they relate to CS topics. It's not just "find the next element in the pattern" but "find the next sorted item in a list" and so on. They could argue that "Well the employee could in principle be inverting binary trees all day at their job..."
I bet many involved in designing and setting the requirements know exactly how it works but they'd use veiled language and never put anything down in written communication.
IQ is imperfect and has many issues that haven't been accounted for, nor should it be extended to represent all of intelligence, and many tests claiming to test IQ are completely pseudoscience, if even that, but the general concept isn't pseudoscience. A standardized IQ test has correlation with a number of other indicators of intelligence. It isn't a prefect test and popular culture has greatly oversold its importance, but our society misusing the science, and even scientists historically misusing the science, isn't reason enough to call it pseudoscience.
Sure, but is that causal? For example, maybe there's some other factor (like SES) that's driving both. And switching to IQ tests doesn't really solve the problem, as your avenues for cheating on that are identical to LC.
You could probably screen candidates based on the question "Did any of your parents go to college? And if so, did they study or work in a STEM field?" and be roughly as predictable.
identical twins reared together vs apart, their IQs correlate more with their twin's than with the SES of the adoptive family.
high quality studies have found that the additive heritability of IQ is somewhere around 0.8-0.9, that's just totally incompatible with parental SES as the primary upstream cause. it's simply not debatable. and it wasn't debatable ten or twenty or fifty years ago either; it is one of the most well-established facts in human quantitative genetics.
ironically one of the things having a high IQ is bad for is having accurate beliefs about IQ; ask anyone who didn't go to college and odds are they'll correctly tell you that smartness runs in families.
The fact that people can just cheat in these interviews and not get PIP'd within 6 months, at an organization notorious for PIPing its under-performers, does illustrate how bad the screening process it. Not only do they need to shift back to in-person interviews, they also need to fix their process to actually test skills.