June 28, 2024

Commonly Used Statistical Models in Psychology are Not Equipped to Deal with Real-World Conditions: A Simulation

View Full Paper

Ask AI

Bookmark

Authors

MSMartina SladekovaUniversity of Sussex AFAndy P. FieldUniversity of Sussex

Discussion

Member takes

Implication

Simulation study reveals that ordinary least squares models fail under common non-normal data conditions in psychology, highlighting the necessity of adopting robust statistical estimators.

Key Points

Ordinary least squares models underperformed across most simulated conditions, yielding accurate estimates only within a very narrow range of symmetric error distributions.
Simulation of typical psychological research varied skewness, kurtosis, and sample size to benchmark standard linear models against robust trimming and bootstrap techniques.
Applying an mm estimator or robust trimming protects against severe skewness, whereas bootstrap methods effectively counter heteroscedasticity in psychological data.

Cite This Study

Sladekova et al. (2024) studied this question.

synapsesocial.com/papers/68e62d5fb6db6435875bfa22 https://doi.org/10.31234/osf.io/xb4at

View Full Paper

Ask AI

Bookmark

Implication

Simulation study reveals that ordinary least squares models fail under common non-normal data conditions in psychology, highlighting the necessity of adopting robust statistical estimators.

Commonly Used Statistical Models in Psychology are Not Equipped to Deal with Real-World Conditions: A Simulation

Discussion

Member takes

Implication

Key Points

Cite This Study

Commonly Used Statistical Models in Psychology are Not Equipped to Deal with Real-World Conditions: A Simulation

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider