Preface

Purpose of the Issue

Theory building is a must if a field is to move forward. In that vein, a call for creative theory building has been strong in the social sciences in general (Haig, 2013; Hesse-Biber & Leavy, 2008; Maxwell, 2004) and HRD as an emerging field in particular (e.g., Holton, 2003; Reio, 2009, 2010; Torraco, 2004, 2005). There are quantitative, qualitative, and mixed-method ways to generate and build theory (Newman, Hitchcock, & Newman, this issue; Torraco, 2004); this issue of ADHR narrows its focus on some of the most promising innovative and cutting-edge quantitative approaches to doing so. Acknowledging that indeed there are many ways to build theory, the authors in this issue take a more deductive logical approach to discussing how to test and validate, and build theory. Inductive approaches begin by exploring authentically open research questions versus a deductive approach where theoretically derived hypotheses are tested (Haig, 2013). Of course, there is some overlap present because the principal components analysis (PCA) that Reio and Shuck (2015) discussed is inductive, in that they are data and not theory driven. Findings from the described PCA (a secondary data source in an organization) could be used to identify preliminary relationships among the variables, where theory could begin to be developed. Notwithstanding, the majority of the quantitative methods highlighted in this issue feature how theory guides the formation of hypotheses and their subsequent testing through quantitative means. The new information gained from analyzing the quantitatively derived findings subsequently enriches what we know about the theory being tested and provides validation evidence. A key emphasis of this issue is the accurate use of these quantitative approaches to reduce misunderstandings about their application as a means to support better theory building, research, and practice in the social sciences.

The purpose of this issue is to describe a number of emerging kinds of quantitative data-analytic techniques that scholars and scholar-practitioners can utilize to test hypotheses specific to HRD settings. This issue is an acknowledgment that better understandings of advanced quantitative techniques and their precise applications are required if HRD researchers are to meet the theory-building needs of the field. The articles in the issue present these statistical techniques within the context of HRD and demonstrate how each can be used preliminarily to advance HRD theory. Each contribution also makes explicit links to how the statistical technique is useful for testing hypotheses in future research and how the information gained, in turn, would assist empirically based HRD practice. Finally, the issue furnishes pragmatic information that would also aid research methods faculty for use in their advanced research methods classes.

Contribution of Articles

The following articles each introduce a quantitative research method and link it to theory building, with careful attention to the method’s use for guiding future research and empirically-based practice. The respective articles are written to keep the HRD audience in mind; extensive use of concrete examples is provided to make the complex material more interesting and immediately applicable for gaining new insights and solving daily workplace exigencies. In the first article titled “Exploratory Factor Analysis: Best Decision-Making Practices,” Reio and Shuck linked their work to Torraco’s (2005) and Reio’s (2009, 2010) calls for more theory-building methods articles and theory building in the field of HRD. Exploratory factor analysis (EFA) was selected for examination because of its extensive use in the social sciences like HRD and Fabrigar, Wegener, MacCallum, and Strahan’s (1999), Conway and Huffcutt’s (2003), and Henson and Roberts’ (2006) compelling evidence that EFAs often were conducted incorrectly. The authors argued that the lack of analytic precision in conducting quantitative analyses such as EFAs led to errors which threatened theory building, empirical research, and HRD practice. Five best EFA decision-making practices were identified that could lessen the likelihood of introducing systematic error into HRD research: selection of observations, factor extraction method, factor retention, type of rotation, and interpretation. The article helped clarify through concrete examples how inattention to each EFA decision-making practice could make published research unsound for use in theory building.

In the second article, “Secondary Data Analyses From Published Descriptive Statistics: Implications for Theory, Research, and Practice,” Nimon presented the use of secondary data analysis as a largely overlooked means to support theory building in the field of HRD. The author introduced statistical software that can be used for conducting secondary data analysis within the general linear model (GLM) framework by using published descriptive statistics (Ms, SDs, ns, and correlation matrices) as the data source. The author noted how following conventional descriptive statistics reporting guidelines (e.g., American Educational Research Association [AERA], 2006) strongly support secondary data analyses as a source of future research. One important contribution of the article is that it alerts HRD researchers to the exciting opportunities for theory building through secondary data analysis by “interpret[ing] old data in new ways” (Torraco, 2004, p. 174). Another contribution was that the article highlights the utility of the innovative R software for theory building because it permits different or advanced statistical analyses on descriptive statistics in published organizational research without the need for collecting new data.

In the next article, “Methods for Analysis of Social Networks Data in HRD Research,” Yamkovenko and Hatala illustrated that even at the individual level, social context matters. In the workplace, individuals and groups can be influenced greatly by formal and informal networks that can affect learning, performance, and change; thus, there is a relational influence on organizational outcomes that must be considered when conducting HRD research. Developed in the field of sociology, the authors introduced social network analysis (SNA) as an approach to studying networks and patterns of relationships within organizational settings. SNA has been applied extensively to refine social network theory. As a relatively unknown research method in HRD, the authors demonstrated how SNA as a rigorous research method can be useful for looking at data in new ways to build theory, and inform research and practice.

In the fourth article titled “Mediating Analysis Approaches: Trends and Implications for Advanced Applications in HRD Research,” Song and Lim promoted mediation analysis as an important methodological means to advancing HRD theory. The richness that mediation analysis affords theory building is that it is a way for investigating not only hypothesized relations between independent and dependent variables but also the role of theoretically relevant mediating variables that influence the relation between the two. Importantly, they distinguish mediation from moderation. Mediation in its simplest form can be thought of as “the independent variable causes the mediator which then causes the outcome” (Shadish & Sweeney, 1991, p. 883). In contrast, moderation differs from mediation, in that the strength and direction of the relation between an independent and a dependent variable can be altered by interacting with a moderator variable (e.g., age, ethnicity). The authors reported on general trends of mediation research, the primary analytic approaches being used, and the importance of using more advanced mediational methods such as structural equation modeling (SEM) and bootstrapping to support all the time more precise and sound HRD theory development efforts.

Next, in the fifth article, “Using the Q Methodology Approach in Human Resource Development Research,” Bartlett and DeWeese introduced Q methodology as a means to study opinions, beliefs, and attitudes, or human subjectivity. Patterns and themes underlying data garnered from interviews and naturalistic observations, for example, can be identified through Q methodology. Instead of identifying what people do, the method supports revealing the processes of how and why people think the way they do. The method does not use traditional R-factor analysis where correlations are between items or constructs; rather, it uses Q-factor analysis where correlations are between participants. Noting that Q methodology has largely been ignored in the field of HRD, the authors presented a cogent case that as an innovative research method, it has impressive potential for exploring data in new ways and subsequently building theory.

In the sixth article, “Hierarchical Linear Modeling: Testing Multilevel Theories,” Turner explored the notion that nested structures are inherent in organizational settings, but HRD researchers tend not to account for such nested structures in their research. The upshot of this is that ignoring that hypotheses should be tested not only at one level, but across levels can comprise the interpretation of research findings. As with the other papers in this special issue, the author calls for utilizing more advanced empirical methods such as hierarchical linear modeling (HLM) to foster theory building in HRD. The advantage of HLM is that it supports the use of predictors at both lower and higher levels to test hypotheses. One important part of using HLM is that it allows combining a number of levels of analysis into one study, extremely useful for conducting increasingly sophisticated HRD research and refined theory building.

Seventh, in “Propensity Score Analysis: A Secondary Data Analysis of Work-Life Policy and Performance Outcomes,” as with a number of the previous authors in this special issue of ADHR, Lane and Gibbs presented a relatively new statistical tool for aiding theory-building (enrichment theory in this case) efforts and research in HRD, that is, propensity score analysis (PSA). Because ANCOVAs and hierarchical regressions are frequently used in non-experimental research to control for theoretically relevant covariates when testing hypotheses, the authors tout PSA as a way to improve theory-building efforts because it overcomes the misuse of both. Researchers tend to overlook, for instance, the necessity for participant random assignment to groups with ANOVAs and the requirement for a lack of covariate interactions with group assignment in hierarchical regressions. Inattention to each can muddle group comparisons and distort theory development. PSA uses covariates (synthetic composite variable) to match participants on their likelihood of group assignment; the statistical matching thereby adds analytic precision to group comparisons, which ultimately supports HRD theory building. The authors’ work also provided new insights into the discussion of work–life balance programs and policies.

In the final article, “The Use of Research Synthesis and Nomological Networks to Develop HRD Theory,” Newman, Hitchcock, and Newman noted that despite the considerable utility of research syntheses such as meta-analysis for advancing theory, research, and empirically-based practice in the social sciences, the procedures have been used too sparingly in the field of HRD. The contribution of this article is that it introduced the umbrella term “systematic review” and described how specific types of research syntheses can be used to systematically aggregate literature and build theory. Specific types of research syntheses include meta-analyses, qualitative syntheses, and integrative literature reviews. The authors focused primarily on quantitative syntheses in the form of meta-analysis, yet discussed qualitative and mixed-method synthesis approaches. The authors, in turn, described how research syntheses can be used to build theory through forming nomological networks, networks that explain how constructs and propositions are linked. Nomological networks consist of several component nets that can be tested systematically and depending on the strength of the evidence, a net can be modified as new evidence and learning expands, and aid further theoretical development.

Conclusions and Implications

The issue responds to the calls of Holton (2003), Torraco (2004, 2005), and Reio (2009, 2010), among others for additional research about theory-building methods and theory building in the field of HRD. Because it is an emergent field, one goal of this issue was to alert HRD researchers to the need to continue testing and building theory to meet its research and practice needs. These indeed are energizing times for HRD researchers as they investigate the links among theoretically relevant, diverse, and untested variables and their organizational outcomes. Along with the novelty and thrill of conducting such research though, the need for theory or theories to guide hypothesizing becomes explicit and unassailable. We must remain mindful that available and future theory requires evidence of cross-validation if its precepts are to remain tenable (Sue, 1999).

Each of the contributors in this issue goes to great lengths to present an advanced quantitative research method and address how its accurate application could be useful for hypothesizing, testing, and building theory. The methods are cutting-edge and underutilized by HRD researchers currently and the new information they present as to its correct and practical use adds considerably to the quantitative theory-building knowledge base. Best practices related to each are identified, the implication of which is that attention to them will support more refined theory-building efforts in the field. Enhanced and enriched new theory will serve HRD researchers well as they generate new hypotheses and design increasingly sophisticated research to test these hypotheses. The findings generated by the new research, supported by theory, serve to inform the process of new HRD theory building. Finally, the practical implications of being introduced to these cutting-edge methods and their appropriate use cannot be overlooked. The new knowledge produced from researchers using the innovative research methods presented in this issue could support empirically-based best future workplace practices and provide opportunities for HRD practitioners to gain possible new insights into multifaceted organizational phenomena in a wide range of settings.