Content uploaded by Deniz S. Ones
Author content
All content in this area was uploaded by Deniz S. Ones on May 17, 2021
Content may be subject to copyright.
IDEA WATCH
Stat WatchHBR.ORG
O
Temperature at which online
shoppers were Uicelier to
go
to a ^'purchase" page
Online shoppers were 46% likelier
to
go
to
a "purchase" page when
the average daily temperature was 77°F (25X) than when
it
was
68"F (2O'C), according
to a
study by Yonat Zwebner, of the Hebrew
University
of
Jerusalem, Leonard Lee,
of
Columbia, and Jacob
Goldenberg,
of
the Interdisciplinary Center
in
Israel. And people
in a warm room were generally willing
to
pay more
for
items than
people
in
a cool room. Physical warmth activates emotional warmth,
eliciting positive reactions
to
products and higher estimates of their
worth,
the researchers say.
DAILY STAT
I's
Dailv Stat by email, sign up
at
hbr.org/dailystat.
TALENT
by
Nattian
R.
Kuncel,
David
M.
KUeger,
and Deniz
S.
Ones
In
Hiring,
Algorithms Beat Instinct
Y
ou
know your company inside out.
You
know the requirements of the
position you need to
fill.
And now
that HR has finished its interviews and
simulations,
you
know
the
applicants,
too—
maybe even better than their friends do.
Your wise and experienced brain is ready
to synthesize the data and choose the best
candidate for the
job.
Instead, you should step back from the
process.
If
you simply crunch the appli-
cants'
data and apply
the
resulting analysis
to the
job
criteria, you'll probably end up
with
a
better hire.
Humans are very good
at
specifying
what's needed for
a
position and eliciting
information from candidates—but they're
very bad at weighing the
results.
Our anal-
ysis of
17
studies of applicant evaluations
shows that
a
simple equation outperforms
human decisions by at least
25%.
The
effect
holds in any situation with
a
large number
of candidates, regardless of whether the
job is on the front
line,
in middle manage-
ment, or
(yes)
in the C-suite.
Moreover, in our research, conducted
with Brian
S.
Connelly, of the University
of Toronto, we looked at studies in which
the people making the call were highly fa-
miliar with the organization and often had
more information about
the
applicants
than was included
in
the equation. The
problem
is
that
people are
easily distracted
by things that might be only marginally
relevant, and they use information incon-
sistently. They can be thrown off course
by such inconsequential bits of data as ap-
plicants' compliments or remarks on arbi-
trary topics^thus inadvertently undoing
a
lot of the work that went into establishing
THE PREDICTIVE POWER
OF
NUMBERS
The bars below show the percentages of above-average employees (as gauged by
three different measures) hired through algorithmic systems versus human judgment.
The numbers represent improvement over chance.
jALGORITHM^S
SUPERVISORS* RATINGS
HUMAN
JUDGMENT
22%
29%
14%
NUMBER
OF
PROMOTIONSABILITV TO LEARN FROH TRAININO
parameters for the job and collecting
appli-
cants'
data.
So
they'd be better off leaving
selection
to
the machines.
Needless to
say,
there would be strong
resistance
to this
idea.
Surveys
suggest that
when assessing
individuals,
85%
to
97%
of
professionals rely to some degree on intu-
ition or
a
mental synthesis of information.
Many managers clearly beheve they can
make the best decision by pondering an
applicant's folder and looking into his or
her eyes—no algorithm, they would argue,
can substitute for
a
veteran's accumulated
knowledge.
If
companies
did
impose
a
numbers-only hiring
policy,
people would
almost certainly
find
ways to
circumvent it.
So
we don't advocate that you bow out
of the decision process altogether.
We
do
recommend that you use
a
purely algo-
rithmic system, based on a large number
of data points, to narrow the field before
calling on human judgment to pick from
just a few finalists—say, three. Even bet-
ter: Have several managers independently
weigh in
on the
final
decision, and average
their judgments.
In this
way,
you can both maximize the
benefits offered by algorithms and satisfy
managers' need
to
exercise their hard-
earned wisdom—while limiting that wis-
dom's harmful effects.
Ü
HBR Reprint F1405C
KSÊ Nathan
R.
Kuncel and Deniz S. Ones are
WÊa professors
of
psychology
at
the University
of Minnesota. David
M.
Klieger
is a
research
scientist
at
the Educational Testing Service.
32 Harvard Business Review May 2014
Copyright2014HarvardBusinessPublishing.AllRightsReserved.Additionalrestrictions
mayapplyincludingtheuseofthiscontentasassignedcoursematerial.Pleaseconsultyour
institution'slibrarianaboutanyrestrictionsthatmightapplyunderthelicensewithyour
institution.FormoreinformationandteachingresourcesfromHarvardBusinessPublishing
includingHarvardBusinessSchoolCases,eLearningproducts,andbusinesssimulations
pleasevisithbsp.harvard.edu.