Page 61 - 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

11

Root Out Bias in Decision‐making

Biased decision‐making is the enemy of data science. We have all experienced the disappoint-
ment when an important decision does not go our way.* The feeling is far worse when you feel
that the decision was somehow rigged – that the decision‐maker did not pay your results their
due or only used some of the data. You can accept a fair decision that goes the other way, but
a rigged decision feels much worse. And the ill will festers.
We have all experienced rigged decision‐making in our business, civic, and personal lives.
And we are not just victims. We are also perpetrators, letting bias creep into our own decisions,
even if we may not realize it. Data scientists make lots of decisions – about which data to
include, who to talk to, how to pose problems, which analyses to conduct, how to present
results, and so forth – that impact the course of their analyses and, in turn, what decision‐
makers see. Even a hint of bias can have profound impact. Further, we may be complicit in
ignoring, even assisting, in the biased decision‐making, shading our results based on what we
think decision‐makers want to hear.
Rigged decisions come in many forms. Here we’ll consider what to do about the most vir-
ulent, which features the following steps:

1. Make the decision based on some or
all of the following: intuition (see Intuition and Rigged Decisions
“Intuition and Rigged Decisions”), ego,
ideology, experience, fear, or consul- In Chapters 8 and 10, we highlighted uncer-
tation with like‐minded advisors. tainty as a dominant feature of all decisions
2. Find data that justifies your decision. and that intuition has an important role.
3. Announce and execute the decision. We further noted that some decision‐makers
Defend it to the minimum degree have great intuition, but that others rely on
necessary. it far too much. When this happens, the
4. Take credit if the decision proves potential for rigged decisions grows.
beneficial or assign blame if not.

*This Chapter is based, in part, on a Harvard Business Review digital article by Redman (2017b).

The Real Work of Data Science: Turning Data into Information, Better Decisions, and Stronger Organizations,
First Edition. Ron S. Kenett and Thomas C. Redman.
© 2019 Ron S. Kenett and Thomas C. Redman. Published 2019 by John Wiley & Sons Ltd.
Companion website: www.wiley.com/go/kenett-redman/datascience