Blog

Novel methods, research best practices, and tools for behavioral scientists.

Methodology & Open Science

The Replication Crisis: What Actually Happened

From Bem's precognition paper to Many Labs and registered reports — a scholarly retrospective on the events that reshaped psychological science and what reform looks like in 2026.

Causal Inference Without Randomization

DAGs, propensity scores, instrumental variables, regression discontinuity, difference-in-differences — a graduate-level tour of how to extract causal claims from observational data.

Multiple Comparisons and the Garden of Forking Paths

Why running 20 tests at α=.05 yields a 64% chance of a false positive — and what Bonferroni, Holm, FDR, and pre-registration actually do about it.

Mixed Models for Nested Data

Students within schools, patients within hospitals, repeated measures within persons — when independence assumptions break and how to model what's actually there.

When to Use Bayesian Methods (And When Not To)

An honest, non-zealot guide. Bayesian inference shines for null evidence, sequential designs, and small samples — and isn't always the right answer.

Power Analysis: How Much Data Do You Actually Need?

The 0.80 convention came from somewhere. So did the underpowered-study problem. A practical guide to sample-size planning for t-tests, ANOVA, regression, and beyond.

Effect Sizes Beyond Cohen's d

Hedges' g, Glass's delta, omega-squared, Cliff's delta, NNT, CLES — a reference-quality tour of effect-size metrics and the field-specific benchmarks that replaced "small/medium/large."

Novel Methods

Convergent Core Analysis: Extracting What Your Data Truly Says

Multiverse analysis shows you 200 results. CCA tells you what all 200 agree on. Here's how it works and why it matters for robust research.

The Complexity Navigator: Finding the Simplest Model That Actually Works

Every model is a tradeoff between simplicity and accuracy. The Complexity Navigator maps that tradeoff and finds the sweet spot automatically.

Dialectical Inference: When Frequentists and Bayesians Finally Talk

What happens when you run both a p-value and a Bayes factor, then algorithmically synthesize the conclusions? A stronger, more nuanced finding.

Scale-Persistent Features: A Durability Test for Your Findings

You found a correlation. But is it real, or did your sample get lucky? SPF stress-tests your relationships across scales and resamples to find out.