| Autor: |
Certo, S. Trevis, Li, Chi Hon, Semadeni, Matthew, Raney, Kristen |
| Zdroj: |
Academy of Management Annual Meeting Proceedings; 2024, Vol. 2024 Issue 1, pN.PAG-N.PAG, 1p |
| Abstrakt: |
Researchers in strategic management often use independent variables operationalized as dummy variables to test hypotheses. A unique property of a dummy variable is that its standard deviation is directly determined by its mean. We use simulations to show how this unique property influences the results of OLS regression models that include dummy variables as predictors. Using a baseline of dummy variables with an equal number of 0s and 1s, our simulations show that while holding all else equal, increasing the imbalance of 0s vs. 1s increases standard errors, decreases statistical power, and decreases R-squared values. We also include Bayes factors in our simulations as a supplementary technique that scholars can employ in hypothesis testing. Bayes factors are particularly useful in helping to understand nonsignificant findings. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
