In a subject that is near and dear to my heart, Jin-Soo Huh from Alpha Public Schools has created a guide for teachers who are looking into statistics about the usefulness of a particular edtech tool.  Three main points: make sure the results were actually statistically significant (p<.05, and they had better show you some data), watch out for caveat statements that could discredit the findings, and beware of studies that use experimental methods that call the results into question (for example, small sample size, discarding a lot of outliers).  And always remember when dealing with statistics of any kind, correlation does not imply causation.