Spss with mathematical tools in RockWorks or ArcGIS are proper.

I like JMP. Very powerful, fast, and pretty user friendly.

R works very well and is easy to use. It can be downloaded for free at:

cran.us.r-project.org

There are so many tools available in MATLAB to deal with statistical as well as nonlinear parameters. Moreover, LABVIEW is also Very powerful, fast, and pretty user friendly.

it depends: for the "light" user there is very nice add-on for Excel, called "WinStat", and for the advanced user of course Spss!

I use SPSS by IBMfor (Principal Component Analysis-PCA and Hierarchical Custer Analysis-HCA to process geochemical data. The results are good but the graphical output quality (especially dendrograms from cluster analysis) is generally poor. I sow the graphical output of Statistica by StatSoft and it is a lot better.

try WinStat, it has good graphics, and of course also PCA and HCA (Ward)!

ArcGis is the best with its geostatistical tool. One more is Statistica 10 Statsoft, where you can easily make descriptive statistics, graphs 2d-3d and multivariate analisys.

Dear Nelson Campos, Sure SPSS will be very useful for you. Since you are doing your research in Geology I would suggest for using MATLAB too. MATLAB and EXCEL has interfacing hence by exploiting this situation you will enjoy.

Hi, I think, that depends, because, some time could be use full SPPS, or Minitab, but in some case is necessary to use Matlab; this one have a very useful library for statistical analysis.

I guess that statisticians are very keen on R. It is free-software and very powerful. Matlab is not free but as far as I read its Linux free-software equivalent QtOctave is not being maintained anymore, I would go for Matlab as well. Cheers

Take R if you do not want copy protection and pay an awful lot of money. In R you can do a lot of of geotstatistics, there are many packages for raster/vector GIS and you can interface with qgis if needed. Forget commercial software if you are not working in an environment where money does not matter. If you need a replacement for MATLAB you can use SciLab. We switched years ago from SAS/SPSS to R and it was the best decision we ever made. The user interface us a little bit harder to master, but (hey - we are scientists) the gain in flexibility is priceless. Just try to get a new software license if your notebook crashes in the middle of nowhere and your license is coupled with the MAC address of your broken computer. If we replace ArcGIS with Qgis we have a complete open source scientific workplace (which also runs on Linux).

To make the start in R easier you can use Rcmdr or Rstudio as a user interface.

I would definitively not use Excel for statistics....

http://www.decisionmechanics.com/when-spreadsheets-go-bad/

http://www.practicalstats.com/xlsstats/excelstats.html

Use Minitab. It is friendly and has a complete menu of statistical tools.

I completely agree with David and Burns. R is a very good software but it is a complex one and you need to use it on regular basis and not from time to time. I never tried GUI release but I will. Thank you both for the suggestion.

For your information: the tutorial for our course (environmental sciences), the script is work in progress...

http://www.hydrology.uni-kiel.de/mitarbeiter/Statistics

You can use Graphpad statistical package

I think everybody is not telling you that R, apart from all the stuffs mentioned yet, have a good number of libraries for geostaticstics. In the last week, we're been using R to interpolate information with Delaunay Triangles.

Here you could find excelent tools for you geostatistics jobs.

http://cran.r-project.org/web/views/Spatial.html

once again: there is an excellent add-on for Excel, called WINStat, and I think it gives you all you need, including multivariate statistics (cost less than 100€).

Seems that R is what most people use, however the Excel WINStat add-on that Chirstian mentions sound interesting.