Page 62 - The Real Work Of Data Science Turning Data Into Information, Better Decisions, And Stronger Organizations by Ron S. Kenett, Thomas C. Redman (z-lib.org)_Neat

50 The Real Work of Data Science

Data scientists must be ready to tackle this chain of events. And they need to be smart
about it.
Understand Why It Occurs

Data scientists know that rigged decisions are antithetical to everything they stand for. So,
approach rooting it out using data science – first try to understand it. Start with the first step:
make the decision. Why do so many people make the decision first?
As we have noted, making good decisions involves hard work. Important decisions are
made in the face of great uncertainty (informally, it appears to us that the more important the
decision, the greater the uncertainty) and often under time pressure. The world is a complex
place – people and organizations respond to any decision, working together or against one
another, in ways that defy comprehension. There are way too many factors to consider. There
is rarely an abundance of trustworthy data that bears directly on the matter at hand. Quite the
contrary; there are plenty of partially relevant facts from disparate sources, some of which can
be trusted, some not, pointing in different directions. With this backdrop, it is easy to see how
one can fall into the trap of making the decision first. It is so much faster! Don’t discount this
benefit.
There are other reasons: Decision‐makers may be motivated by how their decisions will
appear to their superiors, to increase their personal power, and to pay back a favor. They may
have grown overly confident in their own capabilities, or their past experiences with data and
data scientists have gone poorly. There are dozens of possible considerations, and data scien-
tists are well advised to understand the motivations of those they advise.
Once people take the first step (deciding in advance), the second step (seeking data to jus-
tify the already‐made decision) comes easily enough. Decision‐makers know that those
impacted may ask how the decision was made, complain about it, even act to subvert it.
Decision‐makers know they will have to explain themselves, so getting the data needed to
defend themselves is only natural.
This route is common both in business and in the world at large – so much so that Stephen
Colbert coined the term truthiness (Wikipedia 2018c) to roughly mean the preference for con-
cepts or facts one wishes to be true. There has always been plenty of data to support whatever
decision one wants to make. And doing so has grown progressively easier with the penetration
of the Internet, social media, and special interest groups. Further, it is all too easy to fall victim
to confirmation bias (McGarvie and McElheran 2018), where one pays more attention to data
that supports a decision and dismisses what does not.
Steps three and four (announcing the decision and either claiming credit or assigning blame)
also come easily enough.
Take Control on a Personal Level

Before decrying rigged decisions made by others, we recommend that data scientists first
work to improve their own decision‐making. How can you avoid this trap? The first part of the
answer lies in simply admitting your lack of confidence. None of us like to admit we are
biased (Kahneman et al. 2011) – after all, the word carries negative connotations. But the best
decision‐makers we know freely admit their preconceptions. What values or beliefs may be
coloring your thinking? Taking such a hard look in the mirror forces you to acknowledge other
perspectives, softens your knee‐jerk reaction to make a quick decision, and forces you to seek
a broader view.