R or SPSS

Posted by Martin Holt on Aug 09, 2011; 8:40pm
URL: http://spssx-discussion.165.s1.nabble.com/R-or-SPSS-tp4683616.html

R or SPSS ?

 

Whichever one you opt for, when you're starting out it would be a good idea to check that the results of using that package generally fit what you expect to get. From the point of view of checking that you have followed the correct procedues accurately, you could do the analysis using your preferred package, save the results, and then repeat the analysis....if you get the same results then you've checked yourself (unless you've made the same mistake(s) twice . I find I need to do this more with R simply because it's more of a computer language and my background is not computer-techie. In doing this it gives you more of a chance to understand more about how the programs work.

 

SPSS has the GNU interface and Syntax, allowing you to check that you get the same results with both. If you have access to another package you can check that R is doing what you'd expect, but I'm guessing that you don't ? Some people might use Excel, but this is vehemently opposed by others, and anyway Excel couldn't handle the more complex procedures that R and SPSS can. So many analysts use R and SPSS, and if you look at the R website procedures are validated by the R team, and SPSS is sold as being validated that it's probably pedantic to search for this double-check of R and SPSS.

 

Both softwares are well-supported by help lists, although my experience has been that the SPSS list is more user friendly and more open, although quite fairly the R lists do expect you to try to solve the problem yourself if you can before mailing them. 

 

Martin Holt
Medical Statistician