I agree with your point, however I would like to highlight your use of confidence intervals:

"If two statistics have non-overlapping confidence intervals, they are necessarily significantly different but if they have overlapping confidence intervals, it is not necessarily true that they are not significantly different."

http://www.cscu.cornell.edu/news/statnews/stnews73.pdf