The problem is that very few companies know how much false negatives are costing them, so it is very difficult to evaluate whether the false negatives are—cough—a negligible tradeoff against weeding out the false positives.
False positives are right in your face. But it is nearly impossible to systematically keep track of the candidates that aren't hired to see if they are successful elsewhere.
So people lean towards eliminating false positives at all costs. It's another variation on "Looking for the watch where the light is better."
It's another variation on "Looking for the watch where the light is better."
These situations aren't at all symmetric. A false positive always costs you (the company). A false negative is only a cost to you if you fail to hire well for that position.
Also the fact that a candidate you passed on does well elsewhere doesn't mean you were wrong to pass on them (it could, but it's not clear).
That applies in the case of caring more about false positives than false negatives, since nobody (or close enough) even knows how many false negatives they have.
Your statement that someone being successful elsewhere doesn't prove we have a false negative underscores the uncertainty involved. It doesn't invalidate what I'm saying in the least.
But if we want to be rigorous, we could do some direct testing. Let's say we give a telephone screening and four interviews to each candidate. Simpkifying somewhat, let's say that our policy is that a fail on any one eliminates the candidate.
If we hire lots of people, we can collect false positive and false negative stats by randomly passing someone who fails exactly one test.
If a lot of people who fail a particular test end up succeeding, we can infer that the test provides false negatives.
On the other hand, if everyone who gets the random "free pass" on one particular test ends up failing, we can presume it has low false negatives, and maybe abandon the random passes for that test.
The original joke/anectdote works because it leads to clearly suboptimal search - the guy knows his watch isn't near the streetlight, but he's looking there because he can see.
This is a very different situation. To abuse the analogy further, the guys best bet to find a watch seems to be just looking near the streetlight. After all, not only can you see better there, but most of the watches are near there too.
Your saying "false negative is only a cost to you if you fail to hire well for that position" could be easily disproved with examples of people, who were rejected and started their own competitive business (e.g. [1]). IMHO, one of the goals of having 10Ks of engineers on staff is to keep them from the competition. But since nobody is tracking this and the extreme cases like WhatsApp can be hand-waved away it's easy to pretend that the false negative's cost is miniscule.
The guy with the watch analogy only works if there is good reason to believe he’d be better off searching elsewhere. But we don’t appear to have that here....
Let's be generous and say the WhatsApp was a 10x unicorn so FB overpayed only 9B for it. Which makes 4.5B loss per false negative on each founder. If a false positive was, on average, 1M loss (which is an extremely generous in an organization such as FB in my recollection), it means you could have had 4000 false positives for each of these negatives and still come ahead.
And, it brings us back to the Type I and II errors - it's always cheaper to make one, which has a limited downside than the unknown (potentially very high one). Typical example given is that while you are on the way to an expensive cruise you realize that you might have forgotten to turn off the stove. If you go back you might miss your 10K cruise. If you cary on, you might lose you 1M house. On average, people who miss cruise in such circumstances fare much better than ones, who risk their houses.
I get what you guys are saying, there just doesn’t seem to be real evidence that is an issue nearly all cases. Whatsapp notwithstanding
So in order for it to make sense to spend that money, you want some evidence there will be a positive return. Far from contradicting my position, what limited data I have seen does not support that effort, at least in the general case. Certainly whatsapp and the like don’t speak to it - they are very much outliers.
By all means try it. If your false positive rate explodes it will be an expensive lesson, but that’s always a risk.
One last point, the idea that nobody thinks about this and try’s thinks is laughable , which is worth noting because some comments in this thread almost suggest that it’s true,.
With my example of sacrificing a 10K cruise for a chance to not lose a 1M home, I figure?
> Far from contradicting my position, what limited data I have seen does not support that effort, at least in the general case.
Indeed, the data does not contradict a statement "on average, false positives are 4.5B each", which nobody has argued. However, it does contradict the statement "false negatives have any cost only when you are unable to fill the position as the result", which you have made in this thread.
>One last point, the idea that nobody thinks about this and try’s thinks is laughable , which is worth noting because some comments in this thread almost suggest that it’s true,.
Again, it would be silly to say that nobody thinks about it. Evidently, a number of people think the same way you do.
At first, yes. But when one has had enough experience looking for scarce commodities (of whatever kind: genuinely attractive apartments, interesting music / restaurants, romantic partners, etc) one learns that the "where the light is" search technique only gets you so far.
Certainly in this scenario - the false positive rate is only important if the true positive rate is too low. It's not at all clear that is a factor.
That's the whole point -- in order to find evidence that it isn't working, you have to start looking at the false negatives --
or that is to say, where the "light is not".
LinkedIn could offer this as a service, if there was sufficient demand that they could monetise it. Every company will routinely check out every candidate’s profile anyway. They just need an automated way to check again in 6 months and see if everyone they reject is going to a competitor who is beating them in the market.
I would agree with you if the supply and demand weren't so skewed towards top developers. Do you think you are choosing between A and D people, or are you really choosing between C and D people and A and B people never walk through the door?
It's pretty hard for a hiring manager to be called out as too selective. However, if you let in a few bad apples then you look really bad.
The incentives are for hiring managers to be too careful.
It is, oddly enough, the case that if you are not hiring any bad engineers that you probably are being too selective.
Then again, probably one of the worst bad apple I have seen would ace whiteboard as he was quite good at trics. He just did not like the actual work involved in software development and tended not to do work he don't like except bare minimum.
Also you can fix some false negatives - allow people to interview again for other positions at the company. I've seen multiple people get hired on the 2nd or 3rd time around and ultimately be successful.
When we first got married my wife was really hestitant about me working on contract. Over the years, she's seen how much larger my checks are when I do.
Besides, there are some opportunities where companies just want someone to come in, fix a problem and then after that it's just boring maintenance work. If it's a contract, I can come in, fix the problem, and look for a more interesting full time job without being labeled a "job hopper". I can just put on my resume that the 6 month job was a contract.
I prefer contracting but I found that more companies are willing to pay me my minimum salary than the commensurate hourly rate I ask for to make up for the possible gap in employment, unpaid time off, and self employment taxes.
But, my wife has a stable government job with decent benefits that allows me to go back and forth between contracting and permanent and allows me to take risks working at companies that pique my interest without concern about the long term viability of my job.
People complain whiteboarding because they've had bad multiple bad experiences with hiring processes that involve whiteboarding and they came to a conclusion that whiteboarding itself is the problem.
Unfortunately, you can't fix a bad hiring process by simply changing the technique. Each technique has its own pros and cons, but no technique will guarantee success.
These are the most common problems I've seen with hiring processes:
1) The people doing the hiring don't really know what they're looking for. Instead, they go by a vague, handwavy notion of whether the candidate is "a fit" or not. You have to define what "a fit" means in such terms that all of the people involve understand it, agree on it and know how to test for it.
2) The people doing the hiring don't know how to test for their criteria. Whether you're using whiteboard or a take-home project or some other testing technique, you have to design the test so that it helps you observe the traits you're looking for. [1] And you have to actually train your people to interview candidates.
3) Lack of transparency. This is a huge problem and it infests all aspects of hiring. Some recruiters won't even tell you which company they're hiring for. A lot of companies misrepresent what you would be doing for them, sometimes intentionally, but usually not. Some companies won't give you perfectly reasonable information about the job. [2] And, thanks to the litigious nature of our society, no company -- anywhere, ever -- will give the candidate honest, detailed feedback of why they were rejected.
4) The company is unreasonably demanding when it comes to candidate's personal time. The typical two-stage process -- a phone screen followed by on-site interview round -- is demanding enough, because it involves more than just the interviews. The candidate has to spend time and effort on other aspects, too: multiple phone calls with the recruiter, preparation for the interviews, travel, and juggling all of this with their personal and professional lives. Some companies don't understand this, so they think there's nothing wrong with a process that involves a take-home project, a debriefing phone call, a face-to-face conversation to test for cultural fit and a full-day on-site interview round.
Most companies don't even know that their hiring process sucks. Of those who eventually realize something is wrong, many end up cargo-culting whatever they read or heard about "more successful" cases. As things are right now, the whole situation is eerily reminiscent of the fuss and hype the industry went through with software development processes. I wouldn't be surprised to see yet another wave of parasitic consultants pop up to prey on companies by flogging their hiring process improvements and best practices.
----------
[1] This is where a lot of people's complaints go wrong. The most frequent grumble is: why does this company's interview process focus on algorithms and data structures, if you'll just end up coding some services in Java? If the company is doing things right, they're using these topics as a tool to observe and measure whatever it is they're looking for; most often, it's the candidate's understanding of computational complexity. Just because you won't be writing your own implementations of common data structures and algorithms, it doesn't mean you shouldn't know how to pick the one that won't blow up your memory or execution time.
[2] I had a case where I was interviewing with a company that gave me a print-out overview of their benefits, which included the price of each health plan, but not what it covered; when I inquired about it, they replied they wouldn't be able to give me that information until I signed the New Hire NDA.
The employer (or school, etc) wants to make successful hires and to avoid bad ones. False positives are always a big problem. False negatives are usually much smaller one. A process is good if it produces good hires, and no bad ones.
In principle, a process that hires 3 good candidates from a pool containing 5 good candidates and lots of bad ones is good. Avoiding bad hires is paramount. Missing 2 good hires is not as bad as making one bad hire, not near as bad usually.
Most employers don't care about false negatives at all. If they do, they probably care most about missing really good outlier hires, not mistakenly selecting the 5th best candidate instead of the 4th. The only reason to care is fairness. People instintively care about fairness so this affects practices but it's not nearly like false positives.
Employers that need huge volume (eg, the army) are the only ones that care about false positives as much as candidates do.
The candidate (student etc.) generally feels the process is "broken" when it is unfair to them. Unfair means that a person could have done the job well, but cannot get hired. Ie, if the candidate is a "false negative" then the process is broken.
Anyway... When things become systemic, the "brokenness" becomes acute to a candidate. If whiteboards (your architecture version, Google's CS version or any other version) are prevelant and you suck at them... you can't get hired. There are people who suck at your test (whatever it is) but would be good at the job. Your proxy is a biased against some people, every proxy is.
Like I said, a lot of things are like this. One side is working the numbers game, the other side isn't. Insurance markets, college entrance, executive appointments, PayPal's fraudster detection (bastards!), credit scores, investment (even startup investment)...