Eight Statistical Tools Used in HR Management Research

Introduction

Researchers across many disciplines rely on statistical analysis to reach conclusions and present findings in published literature. These techniques are necessary to provide consistency and validity to data, as well as to establish grounds for comparison with other studies. It is from these needs for consistency and comparability that specific statistical methods and tools have been developed and put into common use by researchers and statisticians of all kinds, including those working in human resource management. Though this field of study contains a great deal of qualitative information, as with any good science there are measurable and quantifiable elements to the study and application of human resource management, and these lead to statistical analysis.

This paper examines eight different statistical measures and tools of analysis as they appear in literature pertaining to the study of human resource management. The application of these statistical tools in real-world contexts and practical situations adds clarity to their use, and also provides an opportunity to examine the different ways in which these tools can be and are used in current research. A brief critique of each statistical tool in the context examined is also offered, to provide a broader framework of understanding.

Regression and Time Series Analysis

A study concerning the perceived work climate and turnover rate for different occupational categories in a Korean hospital provides an excellent example of regression analysis. The study was conducted by comparing observed turnover rates to responses from a questionnaire given to employees, which measured a wide array of attitudes and perceptions classified into a total of 32 variables (Hwang & Chang, 2009). The number of variables necessitated regression analysis in order to determine which specific attitudes and perceptions were most correlated with turnover rates.

This is an instance where regression analysis was absolutely necessary to answering the central research question. Regressions are used to compare the effects of two or more independent variables on a dependent variable — in this case, the turnover rate — and with 32 identified independent variables, a regression was definitely called for. This is a simple, compelling, and straightforward example of an application of regression analysis.

The title of this set of statistical tools is entirely descriptive of its methods and applications. An examination of the effects of fear on worker morale over a relatively long period of time is an excellent current example of this type of tool. Research shows that using fear as a motivator is hugely detrimental to employee performance, with the effects increasing dramatically as exposure to workplace fear continues for longer periods of time (Edwards, 2009).

This use of time series analysis is not as clear-cut or as necessary as the regression example above, but it still serves as a useful indicator of the scope of applications that time series analyses can have. The effects of fear on employee performance were measurable in discrete instances, but the researcher's decision to perform a time series analysis revealed a different set of results — a changing effect — than discrete instances alone would have provided. This makes it a highly useful application of the analysis.

Double Blind Studies and Triangulation

In order to eliminate bias as much as possible, many research studies are designed so that researchers are unaware of precisely who or what they are measuring until after the study's completion. In a study of the role of employee competency in service profitability, for instance, researchers analyzed questionnaires from respondents without knowing who they were or where they worked, making these connections only after data had been compiled and variables analyzed for correlation (Xu & Heijden, 2005).

The postponement of full knowledge of a respondent's status is what makes the study double-blind; it makes it impossible for the researcher to unconsciously alter their interpretation based on criteria outside the scope of what the study tries to measure. Double blind studies are more common in medical research, but this is an effective example of the approach applied in the human resource management field. The researchers wished to eliminate any suggestion of bias, and by withholding certain personal details about their respondents, they did exactly that.

The use of several methods and/or approaches to the same research question in a given study is known as triangulation, and it provides both greater understanding and heightened validity to results. A recent literature review of nonprofit human resources practices covering a broad range of situations and effects necessarily incorporated several different methods of analysis in order to determine the most common and accurate ways of analyzing and addressing nonprofit human resources concerns (Ridder & McCandless, 2010).

The triangulation employed in this analysis of human resource management versus strategic management in their effects on nonprofit success allowed for a better understanding of the issue from multiple perspectives, making the case for appropriate human resource management much more clearly and strongly than any single method employed could have (Ridder & McCandless, 2010). This makes the example a very clear demonstration of triangulation as a method of analysis and presentation. Triangulation enhances all of the contributing methods of analysis, creating a whole that is greater than the sum of its parts in terms of analytical clarity and impact.

Conclusion

The examinations of the statistical tools herein have been brief, but hopefully effective. Understanding the different methods of analysis and presentation at one's disposal is essential for carrying out effective research in a way that is meaningful to other researchers and practitioners — in the field of human resource management and in all other areas of scientific inquiry as well.