Comparing data sets (raw vs. not raw data)

#1
I have multiple data sets that I am trying to run regression analyses on. I have the raw data for some of the data sets. For others, I only have the mean, standard deviation, and number of samples. What is the best way to compare these data sets?
 

Dason

Ambassador to the humans
#2
How would you run regression when you only have the summaries? Maybe you could provide a concrete example of what you have and what you want to do.
 
#3
I think I misspoke. If I wanted to compare the data sets in general, whether that's with t-tests or regressions (which I can't do) or other methods, how would you compare these data sets if I only have raw data for some of them?