Wednesday, May 13, 2020

Work samples better than interviews?

Interesting part of a recent New York Times piece on job interviews:
...A work sample can also be a live simulation of the job in real-time. In orchestras, women finally got a fair chance when blind auditions were introduced. Suddenly, managers started making decisions based on the quality of music being played behind a curtain instead of the demographics of the person playing it. At General Electric, to identify aircraft engine mechanics who work well with others, managers dump a pile of LEGOs on the table and ask a half dozen candidates to work together to build a helicopter, and score their teamworking skills. 
At Momofuku, the chef, David Chang, asks job candidates to make him an omelet. “You can tell a lot about an individual if they’re cracking the eggs and they’re trying to get every bit of the albumen out of the egg,” Dave said on my TED podcast, WorkLife. “I’m not trying to see perfect technique. I’m trying to see the intent of the individual first and foremost. I’ll take hungry and eager over super talented any day of the week.”
It would be interesting to understand how various chemical companies do interviews, and if they have a lab practical portion. So far as I understand, there aren't a lot of them (and I suspect that US safety/labor law has something to do with it.

Overall, I agree with the author's perspective that interviewing is a blunt and not very useful tool, but I am not sure what a better one is.  


  1. I like how NYT always seem to be able to get Dave Chang to throw in his two cents, no matter what the topic is.

  2. Hiring is a total crapshoot, even hiring people you've known for some time (though I think that derisks things). The Lego thing is clever, though.

  3. I had a lab practical as the final step of the hiring process at a water treatment plant. It was fairly innocuous stuff, doing some weighing, measuring in graduated cylinders, pipetting water into microbio plates, things like that. All while the lab manager was talking to me about the job!

  4. In theory, a publication track record should be like a work sample, but we all know those can be a crapshoot, with lots of false negatives and false positives. Some of the most brilliant, hardworking, and knowledgeable chemists I know got zero publications during their PhD's, and I know the complete opposite types of people who got JACS/ACIE/PNAS/Science publications (sheer luck). I also know a person who got zero publications in his/her PhD, had an absolute disaster of a defense, and yet managed to get a postdoc with Prof. Grubbs. So, *shrug*.

  5. I feel like this just trades one set of issues for another.

    If the observation is not blinded, as in the Momofuku example, there are abundant opportunities for bias to come in ("I didn't like the way he weighed out reagents"), and these will be harder to spot/correct than the biases we're used to.

    Also, I don't know how big an issue this is, but in the engineering community, there have been complaints that "work samples" were actually work that the company wanted done, but didn't want to pay for.

    Bottom line? Commit to a procedure, and think through all the implications as best you can, I guess.

  6. About 10 years ago, I was visiting a fast-growing company that said they were hiring a large group of entry-level chemists. Like more than a hundred - you can guess the locale. I want to say 400? but it was the late 00's, so my memory may be faulty. It was a lot. They said the hirees would all work for 6 months, and at the end, only the top 2/3 would be retained. That's a system - rigorous and/or ruthless.

    1. Part of my point was that at least the candidates WERE getting paid for the trial work product.

  7. I've had plenty of coworkers who were good at their jobs and valuable teammates, but would have bombed this kind of interview. It's a measure of how well you can think on your feet and resist getting flustered under pressure, which is more of a nice-to-have quality than something the entire hiring decision should be based on.

  8. At my company for the PhD hiring process we have both a behavioral type interview and a technical interview, which is weighted more heavily. For the technical we ask questions about the candidate's seminar and 2-3 papers that they submit. While still not perfect, it does become pretty clear which candidates were just data collectors for their advisors and which ones really engaged with their projects. Candidates are not considered further if the technical team doesn't approve. Then we discuss behavioral. This is a little tricky, but again, you can get a decent idea of how a candidate works. Finally we consider communication and motivational fit. Does the candidate actually want to work here? Do they play nice with others (good)? Do they run with scissors (bad). These are collated from all of the various interactions.


looks like Blogger doesn't work with anonymous comments from Chrome browsers at the moment - works in Microsoft Edge, or from Chrome with a Blogger account - sorry! CJ 3/21/20