As the interest in human capital analytics grows apace, I’ve been thinking again about why it’s taken so long for HR professionals to appreciate the value analytics bring to decision making related to human resource investment.
One factor that came to mind recently as I tried to convince a client of the benefit of displaying (salary) data distributions graphically for preliminary analyses is quite specific – an over-reliance on Excel as the tool of choice for analysis. While hard to argue against Excel’s utility and power in most circumstances, there are certain situations in which Excel simply fails to deliver. Hypothesis testing, regression analysis and graphing of data distributions are some examples. In some cases it’s just a matter of the way Excel handles the analysis – it’s just too cumbersome to use and therefore easily avoided.
What’s the answer? I would argue that it’s time for serious human capital analysts to use statistical packages such as Stata, S-Plus, SPSS, SAS or R. These packages provide power, flexibility and ease of use for conducting practical and important analyses. I’ve begun examining such packages with a view to advising clients which one would make sense for their analysts to adopt.
The problem, however, is that it takes some serious effort to learn how to use statistical packages – both the mechanics of the package as well as the underlying statistics. I’ve felt for a while that there is a huge gap in the market for training HR professionals in even elementary human capital analytics and have therefore developed a curriculum that would appeal to HR departments eager to invest in analytics.
As I examined the available training literature, I came across A Stata Companion to Political Analysis, 2nd Edition, by Philip H. Pollock III. This is a self-learning textbook that introduces students to statistical methods in political science using Stata as the tool of choice.
It is targeted at undergraduates and is a companion book to the author’s undergraduate text The Essentials of Political Analysis, 3rd Edition. The Stata Companion to Political Analysis is 237 pages in soft cover and comes punched with three holes for easy inclusion into a 3-ring binder (the pages are US letter size – 8.5×11 inches). It comes with a CD of data sets that are used in the exposition and review exercises. The book does not purport to teach all aspects of Stata – it describes the features that are needed to conduct the basic analyses and generate relevant graphics.
This book was a valuable find since it demonstrated that statistical methods could be described and taught to students that didn’t have a quantitative background, but who had realized that their chosen area of expertise now required quantitative skills.
The data sets were very practical and included results from polls; social, political and economic information on US states as well as hundreds of countries. Most of the data sets are based on open-source data sets from the US Census and the international data set is available from a Harvard University website. Of course, you need Stata to replicate the analyses and conduct the review exercises.
Stata’s simplicity and power are apparent from the commands that are covered in the book. Like most statistical packages, you can use a graphical user interface (GUI) to run your analyses, issue individual commands at a prompt, or write simple programs. One of the recognized advantages of Stata is its ability to get a good grasp of the data in a very hands-on way. This would be important for HR data which can often be nuanced and requires qualification.
The author’s exposition is simple and detailed, taking you through the logic of the analyses in a very systematic and easily intelligible manner. The second edition makes use of the latest version of Stata, version 11. Stata’s graphics capabilities have been significantly improved with this addition, allowing for detailed customization. Simple and powerful graphing capability is a strong factor in the selection of a statistical package, since pictures convey information readily and succinctly.
As I read through the book I felt that the learning approach would fit well with HR professionals. The statistics are applied in the pursuit of compelling problems and this makes the statistics more digestible and the solution of the problem both exciting and urgent. A college level course in statistics, no matter how long ago it was taken, would be a useful qualification for this book, but not necessarily a pre-requisite since the writing is so clear.
Is this book useful to HR professionals? I think the HR profession deserves a comparable book that deals with human capital analytics. Whether Stata is the right platform to teach human capital analytics is arguable. I believe Stata is a hidden gem for human resources professionals and has significant advantages over the other packages. It is widely used in the areas of healthcare and economics, but has not been adopted very deeply in business use. One reason, perhaps, is the lack of an explicit focus on data mining, which traditional business analytics relies heavily upon (and which medical researchers and economists disdain!).
For those HR professionals already engaged in analytics, I would recommend that you take a look at A Stata Companion to Political Analysis to get a glimpse of how you might be able to broaden the capabilities of your team through an introduction to analytics and a very practical tool, Stata.