When using a quasi-experimental research design, inferring causation is often hampered by the design's ability to demonstrate that:

Rival hypotheses are not likely explanations.

A cause came before the effect.

A cause covaried with an effect.