… the more likely the error in attributing significance to any one variable.

This is the downfall of many poorly designed studies and the Achilles heel of even well designed studies. Statistical models are just that, models. At some point they just can’t keep up with inherent study limitations and all the nuances that biology throws their way.