A correlation is a statistical association between two variables. A causal claim says that changing one variable changes the other. Most cross-sectional and observational studies show correlations.
What separates a correlation from a causal claim
Demonstrating causation requires a study design that can rule out reverse causation, confounding, and selection bias.
Anxiety-specific example
Studies showing a correlation between adolescent social media use and anxiety symptoms can't, on their own, establish that social media use causes anxiety. Reverse causation, anxious teens use more social media, is a live alternative. Confounding factors such as sleep disruption and family environment are also alive. Selection effects matter too.
Designs that support causal inference
Randomized controlled trials, natural experiments, and well-instrumented causal-inference studies are the designs that most strongly support causal claims.
Why this matters for reading research on anxiety
Many anxiety headlines turn a correlation into a causal claim in a single sentence. Reading the underlying study design quickly tells you whether that move was earned.
Sources
Related reading
Your next step in The Shrink Network
Each site in the network has one job. No matter where you enter, we help you find the next step that makes sense.
Your next step in The Shrink Network
You are here: AnxietyResearch, the evidence layer of The Shrink Network.
Each site in the network has one job. No matter where you enter, we help you find the next step that makes sense.
Get clinical care at shrinkMDWant to understand more first?