People & Strategy Summer 2017 Vol. 40 Issue 3 - 38

ines the relationships between or among variables. Advanced
statistical methods can find many types of shared relationships, including non-linear relationships. These methods
can be very illuminating and can uncover previously hidden
patterns.
The risk with correlational studies is that they are by definition not able to definitively identify causation. Identifying a
pattern does not necessarily equip the researcher to change
that pattern effectively or efficiently. In some cases, identifying
a relationship is sufficient to inform a business decision. To
return to our example about predicting the best hires, identifying the colleges that produce the new hires who are most
successful five years post-hire may be sufficient information to
adjust hiring priorities. However, similarly finding that male
gender is correlated with higher promotion velocity should
not inform a business decision to invest more development
resources on men as the most likely promotion candidates.
Rather, that finding would suggest potential bias in your
promotion decisions, or persistent biases in things like work
assignments such that men were more likely than women
to receive assignments with high visibility or development
potential.
Correlational data can often highlight interesting relationships, but it cannot answer those important "why" questions. Thus, it's often important to complement correlational
studies with other methods that can better answer those "why"
questions. Often, qualitative methods, such as focus groups
and interviews, can be very illuminating. However, questions
of causation are best addressed through experimental design.
In an experimental design, treatment conditions and
control conditions are carefully created. This can be more
difficult to do in organizational settings for a variety of reasons. However, for those seeking to determine causation, it
is quite possible to take advantage of natural segmentation
within organizations, and a natural process of rolling out new
processes to create experimental conditions. Natural segmentation could be at the level of a location or branch, a division
or functional area, a shift, or down to the level of individual
teams. The effect of a new method of shift scheduling could
be studied by implementing it in four locations, while keeping
four other locations on the old method and examining the
effect on absenteeism, for example. That would allow variables such as business conditions, time of year, or even a bad
flu season to be controlled, and better isolate whether the new
scheduling approach delivered the desired impact on attendance and absenteeism.
Similarly, to return to our research question on bias in
hiring, an experimental design might focus on changing
or removing the names on previously accepted or rejected
resumes and resending them to hiring managers to see how
often those altered resumes got the same reaction (accept or
reject) as the original resumes. If both diverse and non-diverse
original resumes were reissued in their unaltered, altered, and
nameless form, one could reasonably identify the prevalence
of bias at the resume screening step of hiring.
Longitudinal or time series designs can be immensely
informative, and yet are often neglected due to methodological challenges. Clearly business leaders are disinclined to
38

PEOPLE + STRATEGY

wait five, 10, or more years to for answers to their questions.
For organizations with good historical data, it is possible to
run longitudinal studies by starting in the past and working
forward to the current day. This can be very useful in, for example, identifying the factors present during a high potential
program five years in the past that predict leadership success
today. More closely understanding those factors can lead to
adjusting high potential selection criteria and/or program
content to produce a greater proportion of successful leaders
among program alumni.
More advanced data science and artificial intelligence
techniques, such as machine learning, random forest analyses,
Monte Carlo studies, or Bayesian analysis can be very useful
for answering some complex questions. As the computing
power required to run these more advanced analysis is becoming more available, and the skills required to code the analyses
are more common, these methods are being used more. To
return to our hiring example, an artificial intelligence study
might consist of training an artificial intelligence model on a
set of successful and unsuccessful resumes, and allowing the
machine to determine what attributes were most predictive of
a successful resume. The risk inherent in such a design is that
all the human biases that influence the original hire or nohire decision are being codified by the artificial intelligence
engine, resulting in those biases being applied consistently
going forward. Here, artificial intelligence might add efficiency, but at a cost of quality and fairness.
Finally, where methods are concerned, more is usually better. Researchers can be more confident in their results if they
take advantage of multiple methods to analyze the data.

Step 3: Locating or Generating the Data to Answer the
Question
In some charmed cases, the relevant data will already exist,
in a tidy database ready for analysis. Increasingly, organizations are investing in data infrastructure that captures data
in an analyzable way. Even when data exists, data preparation
and cleaning often takes much more time than the actual
analyses.
Unfortunately, many organizations do not have much useful history data in their human resources information systems
(HRIS), as historically many of these were implemented to
manage transactions, rather than to enable analytics. A good
deal of important data, such as high potential lists, coaching
relationships, or assessment results, never makes it into an
organization's HRIS to start with. Intrepid researchers can
often find such data with tenacious sleuthing, but data quality,
such as partial names or undated lists, can frustrate the best



Table of Contents for the Digital Edition of People & Strategy Summer 2017 Vol. 40 Issue 3

From the Executive Editor
From the Guest Editors
Perspectives
It’s Time for a Second Playbook: HR’s Leadership Role in Transformation
Industry 4.0: Preparing for the Future of Work
When Fast Is Too Slow: “Xcelerating” Leaders at Electronic Arts
Patagonia’s Journey into a New Regenerative Performance Approach
Getting Results with Talent Analytics
How Artificial Intelligence Will Change HR
The Internet of People Delivers New Ways of Learning
Executive Roundtable
In First Person
Linking Theory + Practice
Book Reviews
Leadership Insights
People & Strategy Summer 2017 Vol. 40 Issue 3 - Cover1
People & Strategy Summer 2017 Vol. 40 Issue 3 - Cover2
People & Strategy Summer 2017 Vol. 40 Issue 3 - 1
People & Strategy Summer 2017 Vol. 40 Issue 3 - 2
People & Strategy Summer 2017 Vol. 40 Issue 3 - 3
People & Strategy Summer 2017 Vol. 40 Issue 3 - From the Executive Editor
People & Strategy Summer 2017 Vol. 40 Issue 3 - 5
People & Strategy Summer 2017 Vol. 40 Issue 3 - From the Guest Editors
People & Strategy Summer 2017 Vol. 40 Issue 3 - 7
People & Strategy Summer 2017 Vol. 40 Issue 3 - Perspectives
People & Strategy Summer 2017 Vol. 40 Issue 3 - 9
People & Strategy Summer 2017 Vol. 40 Issue 3 - 10
People & Strategy Summer 2017 Vol. 40 Issue 3 - 11
People & Strategy Summer 2017 Vol. 40 Issue 3 - 12
People & Strategy Summer 2017 Vol. 40 Issue 3 - 13
People & Strategy Summer 2017 Vol. 40 Issue 3 - It’s Time for a Second Playbook: HR’s Leadership Role in Transformation
People & Strategy Summer 2017 Vol. 40 Issue 3 - 15
People & Strategy Summer 2017 Vol. 40 Issue 3 - 16
People & Strategy Summer 2017 Vol. 40 Issue 3 - 17
People & Strategy Summer 2017 Vol. 40 Issue 3 - 18
People & Strategy Summer 2017 Vol. 40 Issue 3 - 19
People & Strategy Summer 2017 Vol. 40 Issue 3 - Industry 4.0: Preparing for the Future of Work
People & Strategy Summer 2017 Vol. 40 Issue 3 - 21
People & Strategy Summer 2017 Vol. 40 Issue 3 - 22
People & Strategy Summer 2017 Vol. 40 Issue 3 - 23
People & Strategy Summer 2017 Vol. 40 Issue 3 - When Fast Is Too Slow: “Xcelerating” Leaders at Electronic Arts
People & Strategy Summer 2017 Vol. 40 Issue 3 - 25
People & Strategy Summer 2017 Vol. 40 Issue 3 - 26
People & Strategy Summer 2017 Vol. 40 Issue 3 - 27
People & Strategy Summer 2017 Vol. 40 Issue 3 - 28
People & Strategy Summer 2017 Vol. 40 Issue 3 - 29
People & Strategy Summer 2017 Vol. 40 Issue 3 - Patagonia’s Journey into a New Regenerative Performance Approach
People & Strategy Summer 2017 Vol. 40 Issue 3 - 31
People & Strategy Summer 2017 Vol. 40 Issue 3 - 32
People & Strategy Summer 2017 Vol. 40 Issue 3 - 33
People & Strategy Summer 2017 Vol. 40 Issue 3 - 34
People & Strategy Summer 2017 Vol. 40 Issue 3 - 35
People & Strategy Summer 2017 Vol. 40 Issue 3 - Getting Results with Talent Analytics
People & Strategy Summer 2017 Vol. 40 Issue 3 - 37
People & Strategy Summer 2017 Vol. 40 Issue 3 - 38
People & Strategy Summer 2017 Vol. 40 Issue 3 - 39
People & Strategy Summer 2017 Vol. 40 Issue 3 - 40
People & Strategy Summer 2017 Vol. 40 Issue 3 - 41
People & Strategy Summer 2017 Vol. 40 Issue 3 - How Artificial Intelligence Will Change HR
People & Strategy Summer 2017 Vol. 40 Issue 3 - 43
People & Strategy Summer 2017 Vol. 40 Issue 3 - 44
People & Strategy Summer 2017 Vol. 40 Issue 3 - 45
People & Strategy Summer 2017 Vol. 40 Issue 3 - 46
People & Strategy Summer 2017 Vol. 40 Issue 3 - 47
People & Strategy Summer 2017 Vol. 40 Issue 3 - The Internet of People Delivers New Ways of Learning
People & Strategy Summer 2017 Vol. 40 Issue 3 - 49
People & Strategy Summer 2017 Vol. 40 Issue 3 - 50
People & Strategy Summer 2017 Vol. 40 Issue 3 - 51
People & Strategy Summer 2017 Vol. 40 Issue 3 - Executive Roundtable
People & Strategy Summer 2017 Vol. 40 Issue 3 - 53
People & Strategy Summer 2017 Vol. 40 Issue 3 - 54
People & Strategy Summer 2017 Vol. 40 Issue 3 - 55
People & Strategy Summer 2017 Vol. 40 Issue 3 - 56
People & Strategy Summer 2017 Vol. 40 Issue 3 - 57
People & Strategy Summer 2017 Vol. 40 Issue 3 - In First Person
People & Strategy Summer 2017 Vol. 40 Issue 3 - 59
People & Strategy Summer 2017 Vol. 40 Issue 3 - Linking Theory + Practice
People & Strategy Summer 2017 Vol. 40 Issue 3 - 61
People & Strategy Summer 2017 Vol. 40 Issue 3 - 62
People & Strategy Summer 2017 Vol. 40 Issue 3 - 63
People & Strategy Summer 2017 Vol. 40 Issue 3 - 64
People & Strategy Summer 2017 Vol. 40 Issue 3 - 65
People & Strategy Summer 2017 Vol. 40 Issue 3 - Book Reviews
People & Strategy Summer 2017 Vol. 40 Issue 3 - 67
People & Strategy Summer 2017 Vol. 40 Issue 3 - 68
People & Strategy Summer 2017 Vol. 40 Issue 3 - Leadership Insights
People & Strategy Summer 2017 Vol. 40 Issue 3 - 70
People & Strategy Summer 2017 Vol. 40 Issue 3 - Cover3
People & Strategy Summer 2017 Vol. 40 Issue 3 - Cover4
http://www.nxtbook.com/ygsreprints/HRPS/hrps_41_3_2018
http://www.nxtbook.com/ygsreprints/HRPS/hrps_41_2_2018
http://www.nxtbook.com/ygsreprints/HRPS/hrps_41_1_2018
http://www.nxtbook.com/ygsreprints/HRPS/hrps_40_4_2017
http://www.nxtbook.com/ygsreprints/HRPS/hrps_40_3_2017
http://www.nxtbook.com/ygsreprints/HRPS/hrps_40_2_2017
http://www.nxtbook.com/ygsreprints/HRPS/hrps_40_1_2017
http://www.nxtbook.com/ygsreprints/HRPS/hrps_39_4_2016
http://www.nxtbook.com/ygsreprints/HRPS/hrps_39_3_2016
http://www.nxtbook.com/ygsreprints/HRPS/hrps_39_2_2016
http://www.nxtbook.com/ygsreprints/HRPS/hrps_39_1_2016
http://www.nxtbook.com/ygsreprints/HRPS/hrps_38_4_2015
http://www.nxtbook.com/ygsreprints/HRPS/d52272_hrps_summer2015
http://www.nxtbook.com/ygsreprints/HRPS/d49675_hrps_spring2015
http://www.nxtbook.com/ygsreprints/HRPS/d47867_hrps_winter2015
http://www.nxtbook.com/ygsreprints/HRPS/hrps_fall2014_teaser
http://www.nxtbookMEDIA.com