Fixed vs Random Effects (Finally Explained)
Hierarchical Linear Modeling (HLM) is often described as “advanced statistics,” but at its core, it’s about something very simple: real-world data has structure — and ignoring that structure leads to misleading conclusions.
In this video, I explain hierarchical linear modeling through the lens of linear mixed effects models, focusing on the distinction between fixed effects and random effects — and why separating them matters for both theory and practice.
Using intuitive examples from classrooms, teams, leadership, and performance, I walk through:
Why most organizational and social science data is nested
What fixed effects actually estimate (and what they don’t)
Why “random effects” aren’t random at all — they capture contextual variation
How ignoring group structure inflates effects and overstates confidence
Why mixed models give us cleaner, more honest estimates of the relationships we care about
This explanation is designed to be accessible even if you’re not a methods specialist — while still being rigorous enough for researchers who work with regression models regularly.
If you’ve ever wondered why results look “too strong,” why findings don’t replicate, or why context seems to disappear in analysis, this video will help you see the logic behind hierarchical modeling more clearly.