Both can be used with secondary data I believe. Is the data cross-sectional? You may need to be careful with assumptions or conclusions of causality with a mediation analysis using cross sectional data. Is there a reason you think they could not be used to analyze the data?