From Saylors and Traifmow (2020) "the probability that an obtained correlation coefficient is theorized for the “right” reason is not the same thing as the correlation itself. For example, imagine that a researcher obtains a strong correlation of 0.80 between A and B, thereby indicating a strong predictive relationship between A and B. This does not mean the correlation is theorized for the right reason. For example, the calculated .8 does not indicate that the arrow goes in the right direction, that there is no exogenous third cause, or that all other arrows or absence of arrows to or from A or B are properly theorized. Even more striking is the fact that actual models often involve a causal chain, making it even more improbable that the causality in the model is properly theorized*. Thus, the numbers we present represent the upper bound of the probability of the model being true."
*We would like to thank Organizational Research Methods Associate Editor Professor Louis Tay for this insight.
Saylors, R., & Trafimow, D. (2020). Why the Increasing Use of Complex Causal Models Is a Problem: On the Danger Sophisticated Theoretical Narratives Pose to Truth. Organizational Research Methods, 1094428119893452.
*We would like to thank Organizational Research Methods Associate Editor Professor Louis Tay for this insight.
Saylors, R., & Trafimow, D. (2020). Why the Increasing Use of Complex Causal Models Is a Problem: On the Danger Sophisticated Theoretical Narratives Pose to Truth. Organizational Research Methods, 1094428119893452.