The Sources of Statistical Biases Series

Formerly the Violating Assumptions Series

(Entries 1-9 updated on 08/24/2021)

When teaching and discussing statistical biases, our focus is oftentimes placed on how to test and address potential issues rather than the sources and effects of statistical biases on the estimates produced by our statistical models. The latter represents a potential avenue to help us better understand the impact of researcher degrees of freedom on the statistical estimates we produce. The Sources of Statistical Bias Series is an endeavor I have undertaken to demonstrate the sources of statistical bias on the estimates produced across various statistical models.

The series will review assumptions associated with estimating causal associations, as well as more complicated statistical models including, but not limited to, multilevel models, path models, structural equation models, and Bayesian models. In addition to the primary goal, the series of entries are designed to illustrate how simulations can be used to develop a comprehensive understanding of applied statistics. I personally believe that simulations can provide knowledge of statistics without relying on the formal calculations that often scare individuals away. You, however, be the judge of that. If you are interested in the effects of violating a specific assumption please feel free to reach out to me at isilver@rti.org. This document will be updated with each new entry.

Important Update: After much thought, I decided that an entire section of the Sources of Statistical Biases Series should be dedicated to propensity score weighting and matching techniques. As such, these entries have been removed from the main text and will be re-added to the document when the section on propensity score weighting and matching is released. The entries on propensity score matching, however, will remain available as independent PDFs. Please let me know if you have any difficulties accessing these entries.

Silver, I. A. (2020). The Violating Assumptions Series: Simulated demonstrations to illustrate how assumptions can affect statistical estimates. DOI: http://dx.doi.org/10.13140/RG.2.2.13339.69921/9. (PDF)

Upcoming Sections and Entries:

Measurement and Statistical Bias
1. Measurement Creation: Dichotomies, Ordinal Measures, Aggregate Events, & Variety Scores
2. Measurement Creation: Aggregate Scale, Average Scale, Standardized Scale, Weighted Scale & CFA
3. Measurement Error
4. Distributional Assumptions for the Dependent Variable (Linear Regression Models)
5. Distributional Assumptions for the Dependent Variable (Mixed-Effects Linear Models)
6. Distributional Assumptions for the Dependent Variable (Non-Parametric Regression Models)
7. Distributional Assumptions for the Lagged-Endogenous Variables (Structural Equation Modeling)
8. Distributional Assumptions for the Endogenous Variable (Structural Equation Modeling)
9. Data Transformations: Non-Normally Distributed Constructs
10. Longitudinal Clustering
11. Spatial Clustering