Psychology
8 years ago

Beware: Longitudinal studies may lead to misleading research findings!

Journal of Speech Language and Hearing Research
Explaining and Controlling Regression to the Mean in Longitudinal Research Designs
Xuyang Zhang, J. Bruce Tomblin

Paper Summary

Related papers

Theme background image
Mathematics
8 years ago

Regression Analysis Unlocks Transformative Insights for Societal Change

Multicollinearity
Multivariate Analysis
2 views
Theme background image
Mathematics
2 years ago

New Statistical Models Revolutionize Decision-Making Across Industries

Multicollinearity
Chemometrics
1 view
Theme background image
Mathematics
2 years ago

Uncover Hidden Patterns, Predict Outcomes, and Transform Decision-Making Across Industries.

Multicollinearity
2 views

Regression to the mean can affect research findings in studies of clinical populations. This happens when participants are chosen based on extreme scores and then studied for changes in that trait. The error in estimating change is due to measurement mistakes and bias in selecting extreme scores. More unreliable measures and extreme score samples lead to bigger regression effects. Changes in dichotomous states from normally distributed traits show prominent regression effects. The study used simulations to demonstrate these features of regression to the mean.