January 10, 2017 | Written by: Nigel Guenole
Categorized: Talent Analytics
Share this post:
Over the last few years at IBM my eyes have been opened to the powerful research methods and designs being used in workforce analytics. I’ve met economists using instrumental variable regression approaches, computer scientists working with neural nets, and sociologists deep into social network analysis. It can be hard to keep up with all of the advances in my own area of industrial-organizational psychology, let alone monitor the approaches of others from different backgrounds. And I’m not alone. Many HR practitioners are looking for guidance when it comes to workforce analytics research methods and designs.
This is definitely not a time to stick our heads in the sand. Human resources is now attracting greater numbers of analytically capable workers. These workers come from fields that have not previously shown much of an interest in human resource problems, but they are rightly excited by the opportunity of working on some of the most interesting challenges in business. So, our more seasoned HR professionals, who may lack a deeply analytical background, need to stay on top of analytic methodologies and understand what these new analysts are trying to achieve. In a recent paper I wrote with Dr. Sheri Feinzig, we set out a framework to provide this needed clarification.
Our framework starts from the premise that everyone working in workforce analytics requires a reasonable level of familiarity with research designs and analytical methods, but they do not need to be methodology experts. With that premise in mind, let’s get started with a couple of important definitions:
- Research designs refer to the ways that data are collected – who the data will be collected from, how the data will be collected, when the data will be collected, and so on. Decisions about how data are collected impact the conclusions you can draw from subsequent analysis. If you are developing business strategies on the basis of workforce analytics you need to start with a solid research design. For a hierarchy of research designs, including quasi-experiments and correlational studies, check out our white paper, Decoding Workforce Analytics.
- Analytical methods describe the analytical techniques applied to the data you collect. While there are many new methods emerging from different disciplines, a close look reveals that they generally aim to achieve a small handful of goals. Knowing your goal enables the ideal selection of analytical method. Are you looking to explore your data, examine an association, make a prediction, make a classification, segment a data set, or reduce the complexity of a data set? Once again, see the white paper to understand enough about these objectives for informed conversations and to enable you to map new techniques on to one of the basic analytical objectives.
The IBM white paper, Decoding Workforce Analytics, is based on a chapter from our forthcoming book from Pearson FT Analytics Press, The Power of People: Learn how Successful Organizations use Workforce Analytics to Improve Business Performance. I coauthored the book with Jonathan Ferrar and Sheri Feinzig to provide practical advice on developing workforce analytics capability. The content is informed by interviews with more than 60 workforce analytics experts in many of the world’s most successful organizations. Pre-order the book on Amazon.