3 ms·

The interview is also selecting for a single thing - GPA. You can be an utter arsehole and have a high GPA. You can have personal hygiene that stinks out any ro

by vacri 10y ago

The interview is also selecting for a single thing - GPA. You can be an utter arsehole and have a high GPA. You can have personal hygiene that stinks out any room you're in and have a high GPA. Basically, you can be completely impossible to work with and have a high GPA. The research they've done is suspect, because they weren't interviewing for an ongoing role, but for a single KPI. Similarly, the people doing the prediction were other students rather than teachers who know better what to look for. The research would better fit HR people hiring for roles they know nothing about than experienced team members hiring for roles similar to what they do. The research results are massively over-applied, and in no way whatsoever is sufficient to use the term "utter uselessness of X". Unsurprising, given it is research by a 'professor of marketing'.

tptacek 10y ago
Like everyone else, I agree a face to face interview is the best way to verify a candidate doesn't smell bad. I assume that interview could be pretty short.
- vacri 10y ago
  Well, I suppose you could fish out the most trivial part of the comment and refute that. Why not set up a battery of short unit tests for the candidates? "Follow me, Jones, and we'll see how you handle the 'coworker with chewing gum' test next..." The study was done on students, who almost universally have zero experience selecting people to work under them in an industry setting (or at all). Drawing conclusions from this particularly inexperienced subject pool and then extrapolating out is bogus, particularly given the extremely certain language of the article. The subject pool is at an age (18-22) where people are still figuring out what to make of themselves and others; they have extremely little adult experience of the workplace and judgment of character - indeed, at this age, people are notorious for making bad decisions in their personal lives. When you look at the actual paper they link, out the window goes the declarative language, and instead the article is unusually full of weasel-words ('can', 'may'). There's a major difference between "utterly useless" and the actual conclusion of their paper: "interviewers probably over-value unstructured interviews".
  - tptacek 10y ago
    In the experiment the authors ran, interviewers felt more confident after interviewing a "candidate" who had secretly been instructed to randomize the answers. Later, they informed the researcher that so great was the value of the information they extracted from interviews, they'd prefer a randomized interview to no interview at all. I think the authors made their point quite well.

by vacri 10y ago

tptacek 10y ago

Like everyone else, I agree a face to face interview is the best way to verify a candidate doesn't smell bad. I assume that interview could be pretty short.

vacri 10y ago
Well, I suppose you could fish out the most trivial part of the comment and refute that. Why not set up a battery of short unit tests for the candidates? "Follow me, Jones, and we'll see how you handle the 'coworker with chewing gum' test next..." The study was done on students, who almost universally have zero experience selecting people to work under them in an industry setting (or at all). Drawing conclusions from this particularly inexperienced subject pool and then extrapolating out is bogus, particularly given the extremely certain language of the article. The subject pool is at an age (18-22) where people are still figuring out what to make of themselves and others; they have extremely little adult experience of the workplace and judgment of character - indeed, at this age, people are notorious for making bad decisions in their personal lives. When you look at the actual paper they link, out the window goes the declarative language, and instead the article is unusually full of weasel-words ('can', 'may'). There's a major difference between "utterly useless" and the actual conclusion of their paper: "interviewers probably over-value unstructured interviews".
- tptacek 10y ago
  In the experiment the authors ran, interviewers felt more confident after interviewing a "candidate" who had secretly been instructed to randomize the answers. Later, they informed the researcher that so great was the value of the information they extracted from interviews, they'd prefer a randomized interview to no interview at all. I think the authors made their point quite well.