Voodoo Correlations Are Everywhere-Not Only in Neuroscience

Area of Science:

• Psychological Science
• Neuroscience
• Research Methodology

Background:

• A methodological issue of inflated correlations between brain measures and behavioral criteria has been identified, particularly when selecting specific measurement points (voxels).
• This problem, initially highlighted in neuroscience research, is a symptom of a broader issue within paradigmatic research.
• Paradigms are conventional setups that can lead to idealized and inflated effect sizes by optimizing various research components.

Purpose of the Study:

• To identify the broader methodological problem underlying inflated effect sizes in paradigmatic research.
• To discuss the implications of selecting specific research parameters for empirical phenomena visibility and stability.
• To propose a reorientation of research practices towards genuine validity and scientific scrutiny.

Main Methods:

• Analysis of existing research practices within paradigmatic research frameworks.
• Examination of how various components (e.g., stimuli, variables, parameters) are selected to enhance effect sizes.
• Critique of the focus on maximizing effect size visibility in current research.

Main Results:

• The problem of inflated correlations is not limited to neuroscience but is a pervasive issue in all paradigmatic research.
• Researchers systematically select numerous components, including voxels, stimuli, and variables, to maximize the visibility and stability of empirical phenomena.
• This selection process results in idealized, inflated effects that may not reflect true validity.

Conclusions:

• Paradigmatic research conventionally optimizes for strong effect sizes, potentially at the expense of genuine validity.
• A viable remedy involves reorienting research from maximizing effect visibility to prioritizing scientific scrutiny and genuine validity.
• This shift is essential for advancing robust and reliable scientific understanding across disciplines.