Why can't you make confident cause and effect statements about the results from a quasi-experimental design?

Subject to concerns regarding internal validity because the treatment and control groups may not be comparable at baseline