 |
Cleaning up the Mess: A Systematic Approach to the Interpretation of Continuous Two-Way Interactions
|
Author: Jeremy F. Dawson, U. of Sheffield
Author: Andreas Wilhelm Richter, U. of Cambridge
|
|
|
|
|
|
|
|
Common post-hoc probing for interpreting significant moderated multiple regression (MMR) findings include simple slope tests and the Johnson-Neyman technique: methods that rely on hypothesis testing, but can be arbitrary and even inappropriate. In this paper, we develop a typology of effects for continuous moderators, into which researchers can classify their moderation effects. Additionally we propose two effect sizes (one unstandardized, one standardized) that can be used to describe the main X-Y relationship being studied, and therefore should be more useful in interpreting effects than current alternatives. We advocate a series of steps that researchers can use to ensure their interpretation of two-way interactions is consistent, conducive to their intents, and not based on arbitrary choices or inappropriate use of significance tests.
|
|
|
|
Paper is No Longer Available Online: Please contact the author(s).
|
|
|
|
|
Disability and Socialization at Work: Causal Evidence from a Regression Discontinuity Approach
|
Author: Anna Brzykcy, U. of Sankt Gallen
Author: Stephan Alexander Boehm, U. of St. Gallen
|
|
|
|
|
|
|
|
Persons with disabilities remain an underutilized resource in the labor market, a situation with important implications for their social inclusion at work and in society. Stigmatizing labels (“mentally ill”) were shown to have detrimental effects for affected persons’ social support networks. We transfer this finding to the employment of individuals with disabilities. Specifically, we use the regression discontinuity design to test the proposition that persons officially labeled as “severely disabled” (holders of a severe disability identification card) experience less socialization opportunities at work than their counterparts with a similarly severe disability but without the label. We draw a sample of 512 persons registered as disabled from a large data set representative of the German workforce (N = 8,019). We test the treatment effect over and above the effect of individuals’ actual occupation. Finally, due to the post-stratification adjustment of our data, we are able to generalize the results to the population of interest, i.e. 4.1 million employees (and internet users) in Germany that are registered as disabled. As expected, being a holder of a severe disability identification card causes decreased socialization opportunities at work. Potential threats to the internal validity as well as theoretical and practical implications are discussed.
|
|
|
|
Paper is No Longer Available Online: Please contact the author(s).
|
|
|
|
|
Fractional Regression Models in Strategic Management Research
|
Author: Anders Ryom Villadsen, Aarhus U.
Author: Jesper Wulff, Aarhus U.
|
|
|
|
|
|
|
|
Many outcomes of interest to management and strategy researchers are fractional in nature. Fractional outcomes are attractive to researchers because they are easy to understand and facilitate interpretation of effect sizes that are meaningful and comparable across organizations. We review 10 years of research in five leading strategy and management journals. We find that approximately 5 percent of articles include statistical estimation of a fractional outcome. However, few studies use the best available statistical techniques. Using two empirical examples and a simulation, we demonstrate that linear regressions (incl. transformations) and Tobit models are inferior to fractional logit models and mixture models. The former are likely to produce biased estimates and unreliable significance tests. We present advice on how to estimate fractional outcomes in different situations.
|
|
|
|
Paper is No Longer Available Online: Please contact the author(s).
|
|
|
|
|
|
How Scaling Variables by Firm Size Limits Knowledge Accumulation in Strategic Management
|
Author: S. Trevis Certo, Arizona State U.
Author: John R. Busenbark, U. of Georgia
Author: Matias Kalm, Arizona State U.
Author: Jeffery LePine, Arizona State U.
|
|
|
|
|
|
|
|
Researchers in multiple disciplines have highlighted the problems that may result from using ratios in analytical models. In this study, we examine how ratios involving measures of firm size impact knowledge accumulation in strategic management. To examine this issue, we introduce an innovative technique whereby we run simulations based on real data involving R&D expenditures, debt, cash, and net income. We use simulation techniques to explore how parameter estimates may change when the models include ratios. Our simulations show conflicting findings with respect to each of these variables. Specifically, we illustrate that the OLS regression coefficients of ratio variables are not normally distributed across samples in the simulation, are both positive and negative depending on the ratios in the model, and result in low R-squared values. We provide a discussion of how this may impact strategy research and what scholars can do to resolve issues moving forward.
|
|
|
|
Paper is No Longer Available Online: Please contact the author(s).
|
|
|
|
|