A reference for key terms used in our discussions on methodology, statistics, and neuroscience.

fMRI (functional Magnetic Resonance Imaging)

A neuroimaging technique that measures brain activity by detecting changes associated with blood flow. This is the technology at the center of the "Voodoo Correlations" debate.

Non-Independent Analysis

Also known as "circular analysis." A methodological flaw where the same data is used twice: once to *select* a subset of data (e.t., "significant" voxels) and again to *test* a hypothesis on that same subset. This artificially inflates results.

p-Hacking

The practice of re-analyzing data in multiple ways (e.t., adding variables, removing outliers, changing models) until a statistically significant p-value (typically p < .05) is achieved. This is a "questionable research practice" (QRP).

p-value

A p-value is the probability of observing a result at least as extreme as the one found, *assuming the null hypothesis is true*. A low p-value (e.t., < .05) is traditionally used to reject the null hypothesis, but it is often misunderstood and misused.

Publication Bias

A bias in the scientific literature where studies showing positive, "significant" results are more likely to be published than studies showing negative or null (non-significant) results. This creates a skewed view of the evidence.

Replication Crisis

A widespread methodological crisis in science (especially psychology) where many well-known, published findings have failed to be replicated by independent researchers.

Statistical Power

The probability that a study will detect an effect that is genuinely present. Studies with low power (often due to small sample sizes) are more likely to miss real effects (false negatives) and have a higher rate of false positives.

Voxel

A "volumetric pixel," or the smallest 3D unit of measurement in an fMRI scan. A single fMRI scan contains hundreds of thousands of voxels, creating a massive multiple comparisons problem.