Hacker News new | past | comments | ask | show | jobs | submit login

Using data like this to inform your hiring decision might be a good idea, but only if you acknowledge that the data you are comparing against might be a local maxima. If everyone you already hired is mediocre and then you only hire people who have the same personality, you're potentially throwing away the people who will make everyone better.

Not to mention diversity of experience is one of the best ways to improve both the workplace and the products that come out of it.

Used to work on a product doing OCEAN pre-hire testing.

The way these systems have classically worked is a company would have some researches put together a corpus of around 200 questions, most directly mapping to an OCEAN category, a few attention-check questions to make sure the user isn't just hitting "Very Likely" all the time, as well as a few questions that are the same as earlier questions but phrased differently to try and catch liars.

This corpus would then be surveyed against ideal candidates, how they determine that is a mystery to me, for a given job, e.g. "Insurance Salesman" or "Grocery-Store Cashier." Then when an applicant comes in they will be tested against this model of an ideal. The population surveyed might be "All cashiers in the country" or "All cashiers in your company," depending on where you got your software and how much you were willing to spend.

Newer software in this space will provide tools like employee analysis, e.g. Bob is more (A)greeable than Joe but also less (E)mpathetic. Usually these will also allow for arbitrary groupings and analysis, so you could compare two existing or hypothetical teams, etc. This would allow a manager to take their best performers and look for comparable candidates, personality-wise, or alternatively look for similarities to the bad-apples.

Personally, I think there is something to it from a "culture fit" perspective. If you take someone who is very (C)onscientous, very (N)eurotic, and throw them into a team filled with the opposite, you'll run into issues. At the high-end, low-applicant jobs like software these systems might be used to spotlight possible problem areas so that you can further assess the candidate. In the low-end, high-applicant jobs like a grocery-store clerk it becomes an automated HR tool to filter out candidates based on some statistics.

At the end of the day it's a tool with some amount of scientific backing and legal defensibility, how a hiring manager uses it can be good or bad and says more about them than anything else.

Flexing corporate power to make prospective employees submit to invasive things like this seems a good way to only hire those who are desperate or don't know any better.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact