Search This Blog

Monday, December 26, 2011

Google - Bias in Hiring

Daniel Kahneman
I have been reading a book called "Thinking, Fast and Slow" - a book by Nobel laureate Daniel Kahneman.  If you are interested in how you and other people think in terms of preconceptions and biases then I highly recommend this book.

Now I am writing today not about this book per se but about what this book says about how we bias out thinking out in the real world.  So here is an example, from the book as described in this NY Times book review, called "the linda problem."

Participants in the experiment were told about an imaginary young woman named Linda, who is single, outspoken and very bright, and who, as a student, was deeply concerned with issues of discrimination and social justice. The participants were then asked which was more probable: 

(1) Linda is a bank teller. 

(2) Linda is a bank teller and is active in the feminist movement. 

The overwhelming response was that (2) was more probable; in other words, that given the background information furnished, “feminist bank teller” was more likely than “bank teller.”

The only problem with this is that set of feminist bank tellers in choice 2 is a subset of of choice 1 making it more probably than choice 2.

So to me what's interesting here is that the experiment appeals to specific aspects of people's biases.

Now, with this in mind, I was reading this WSJ article about hiring at Google.

Basically this article describes how Google interviews job candidates with tricky puzzle questions.  For example, according the linked article, one question asked by Google (as described by applicants in  post-interview discussions) is "You are shrunk to the height of a nickel and thrown into a blender. Your mass is reduced so that your density is the same as usual. The blades start moving in 60 seconds. What do you do?"

Now, as an employer and as someone who, in a job interview context, was asked similar kinds of questions I am intrigued by the relationship between "the linda experiment" and this interview model and what it says about the interviewing company.

Basically, at least superficially, the idea of behind these kinds of questions is to use them to judge how you will react to situation.  Are you "brainstorming," are you stymied, how well do you think on your feet, how well do you listen, and so on.

Google is not the only company to use these questions and, based on the article at least, the entire concept is being adopted by other companies as well.

But, after reading Kahneman's work, one starts to wonder what sort of biases might be involved in this sort of interviewing process.

For one I see that, at least in the mind of Google, the world is a small, simple place.  Everyone understands a blender - whirling blades of death and destruction.  You, the shrunken human, maintain your humanity and density in a smaller form.

Kind of like a story you might tell a small grandchild.

Google is interested in how you would "react" to this.

But I am not a small child taken into a magical (how I shrank) and dangerous (whirling blades of death and destruction) world.

If I magically shrank to the size of a nickle why can't I magically stop the blender?

There really isn't even the logical or probability-based aspects of the "linda experiment" here - just a bias toward a make-believe world where problems are disconnected from reality.

(Does that mean solutions have to be too - no, apparently not.  SO even if you're in this magical spot why must you use rational thought to escape it??)

So in my business, which touches many large companies across the face of the earth in may countries, expressing business or technical thoughts like this would make my customers frightened.  If my customer had a problem would they expect child-like wonderment about it?

Magical solutions?

Being a geezer I don't like a world run by "child-like intellects" - that's my bias.

I am a grown-up - child-like solutions are known to me not to work - so actually its not a bias.  I have tried wishing problems away - but it doesn't work.

Now let's look at Google's former boss - Eric Schmidt.  A while back Congress wanted to know if Google's searches favored Google's products.  (See this article.)

A simple premise and question.

If you offer "free searches" are they biased, i.e., if there is product A from some other company and product B from Google - does Google try and present product B (one which Google stand's to benefit from) ahead of product A?

Everyone uses Google and Congress is trying to understand if Google is dealing itself an advantage under the table.

Schmidt's response was "…the question of whether we “favor” our “products and services” is based on an inaccurate premise. These universal search results are our search service — they are not some separate “Google content” that can be “favored.”

From my perspective a "child-like argument" - no, no we don't offer video on YouTube - we just offer "Search Results."


To me this is reminiscent of the interview question - "child-like."

Except like a small child trying to lie his way out of his hand obviously in the cook jar, e.g., how could I be reaching for a cookie - the jar is empty - he tell's mom...

But mom doesn't buy it.

From what I read about Google over the years the entire place has this quality - toys for the adult employees to play with, and so on.

I wonder how it is they make adult decisions if their model of the world and interview bias is "child like?"

To be sure technical advances are often made by those who look at problems from a different perspective.  But I am not sure that a "make believe" ability is a good criteria for hiring.

I remember being in a similar interview related to some complex, high performance imaging I had worked on.  Two things stand out in my mind from this.

First, they wanted me to solve some sort of stupid puzzle questions in the interview.  I am not good at puzzles and puzzles don't reflect the real world which is far, far more complex.  To me, at least, puzzle questions test your ability to solve puzzles, not solve real problems.

(I am biased too, I have made a living for many decades doing what I do - putting children through college, etc.  So, at least in my mind, not operating with a "child-like" perspective seems to work.  And, by the way, I am still able to create fantastically original products - but for this I was motivated not by child-like fantasy but instead by children needing food...)

Secondly they wanted me to describe to them techniques I had developed to solve the problems they were interested in.

Sure, I'll tell you how to solve your problems and you won't need to hire me.

Except your bias toward looking at things in a simple, puzzle-like fashion will keep you from really understanding what the issues are.

I guess the bottom line is if you build a business on "child like" thinking you should expect mommy (Congress) to show up and slap your wrist when you are naughty...

No comments:

Post a Comment