Suppose the treatment is a new way of teaching writing to students, and the control is the standard way of teaching writing. We co two groups on 40 different attributes related to writing, eg. run-on sentences, comma splices, spelling, and so on. Hno difference in the scores for the two groups H: the treatment (new teaching method) group has better scores We find that the treatment (new teaching method) produces statistically significant improvement on two attributes: run-on sentenci Why should we be cautious in believing the conclusions from this study? The statistically significant improvements are probably due to Type 1 error. The statistically significant improvements are probably due to Type I error. The statistically significant improvements are probably due to sampling error The treatment is a better teaching method on these two attributes.