THINK AGAIN: CONCEPTS TO REVOLUTIONIZE PEOPLE DECISIONS
AT A GLANCE
Scientific thinking can revolutionize the way we make decisions about our workforce. If
embraced, the following concepts can result in better people decisions:
1 - The Law of Large Numbers: The quantity of evidence needed to make a decision is a function
of the error variance for that attribute. The smaller the sample, the greater the error.
2 - Regression to the Mean: If measured over time, extremes in performance will average
out. In decision-making, consider the average, rather than single events.
3 - Opportunity Cost: Opting for cheaper or easier alternatives can come at an opportunity
cost. You can pay the price for accepting good enough instead of holding out for great.
4 - Sunk Cost: Past expenditures cannot be recovered and trying to do so can result in
wastage of time, money and resources.
5 - Self-selection: When we measure a target characteristic of a person (as opposed to
assigning the characteristic), many other characteristics associated with the target could
explain any given outcome.
6 - Control Testing: Randomized control testing and A/B tests can offer scientific method
to people decisions. The method avoids the self-selection problem and makes it possible
to be confident about causal relationships.
1. The Law of Large Numbers
Suppose a business leader interviews an applicant for the position of manager in
a major division of his company. The applicant comes armed with a great track
record, but the interviewer finds the applicant to be lackluster and somewhat
tongue-tied. As a result, he tells his colleagues that he doesn’t think the applicant
should be considered for the job. Reasonable decision or not?
And suppose a football coach goes to watch a practice to observe a possible
recruit—a striker who had earned great praise and had an impressive win-loss
record. On the day, however, the player puts in a weak performance. Disappointed
by what he sees, the coach tells his colleagues that the player should not be
selected for the team. Reasonable decision or not?
If you’re like most people, you may think the coach has been too hasty. One
practice doesn’t provide a great deal of evidence about skill in football. Again, if
you’re like most people, in the first example, the potential employer’s behavior
may seem reasonable at first glance. But it’s not. It is actually a worse judgment
than the coach’s.
Here’s why: Interviews, which typically take just over half an hour, predict future
competence in any business role to the tune of a correlation of .10. That’s the
equivalent of raising the odds of picking the better of two applicants to 54 percent,
up only slightly on the 50 percent chance you would get by tossing a coin.
As strange as it may sound, not meeting the interviewee at all could yield better
odds. The folder on any given candidate often contains enough information that,
if weighted properly, can increase the likelihood of choosing the better of two
candidates to the vicinity of 65 to 75%. So, in terms of choosing the right person
for the job, a brief interview is often less useful than selecting a candidate off
paper. Futhermore, interviews can actually be dangerous because we’re inclined
to weight them too heavily in relation to their actual evidential value.
So, why is it that people tend to make the right call for the coach and the wrong
call for the employer? The answer lies in the fact that we are well aware of
variation across occasions for athletic skill, but typically we don’t apply the same
logic to business employees.
We need to view every observation of a person as just a sample from that individual’s
behaviors. That will lead us to the need to apply the law of large numbers. This
means that the quantity of evidence we need in order to make a judgment about
an attribute of a person, place or thing is a function of the error variance for that
attribute. In simple terms, the smaller the sample, the greater the error.
72 Emerging Markets Business Summer 2016 • Issue No. 1
40 MINUTES
Interview Success Formula, a company
dedicated to helping job seekers ace
interviews, collated information from
multiple sources including the US Bureau of
Labor Statistics and The Wal l Street Journal,
to reveal useful stats for job hunters.
According to the company’s research, the
average job interview lasts 40 minutes.
Before reaching that stage, however, many
companies sift out up to 50% of applications
using talent management software.
REGRESSION
TO THE MEAN
Imagine you have a great meal at the new
Indian restaurant in town. You can expect
that the next meal is not going to be as
good. This is not because there’s a new
chef or because your expectations were too
high. Rather, there are more restaurants in
the world where you can get an excellent
meal some of the time than there are
restaurants where you can get an excellent
meal every time. That’s another way of
saying there are more restaurant meals
close to the mean than there are meals
that are far away from it.