Correlation vs causation

[akimhaneul]

---

Members do not see this ad.

In psychology studies, when can you actually conclude causation?


Sent from my iPhone using SDN mobile

 

[theonlytycrane]

When it's an independent variable -> dependent variable setup. Like I explicitly change X and observe Y.

 

[aldol16]

Basically only under experimental conditions. You have to be able to manipulate one variable while seeing the effect on another. Otherwise, it's correlational.

 

[bobeanie95]

You need three conditions to determine causation:

1. Correlation between the two variables
2. Temporal order, so that the independent variable is always temporally prior to the dependent variable
3. Determine non-spuriousness (no confounding/3rd variables are involved).