What’s the difference between correlational study and causality experiment? When is it safe to conclude there is a casualty relationship between two variables?
It is safe to assume a casual relationship when you specifically are manipulating the variables. A correlational study allows us to identify the relationship between variables, so we might notice that people are happier when they get outside. By specifically manipulating a variable, such as introducing light therapy, we can push it over the edge to an experiment.
An experiment is the only type of study that can prove causation. Look for action verbs.
A correlation study “predicts.” One variable “predicts” the presence of the other in a positive correlation, and one variable “predicts” the absence of the other in a negative correlation.
Correlational study is looking at whether A often appears (predicts) B.
For example, did you know that 100% of people who drink water (A) experience death (B)?
It’s true! But that certainly doesn’t mean that drinking water will kill you.
To determine if one variable is directly responsible for another variable researchers perform a causality experiment. In these studies, specific variables are defined precisely with operational definitions.
To answer your question about safety, you’d need to do a little research on correlation coefficients and statistical significance, but there are already a lot of posts on our forum tonight for these