Statistics

market research — statistics — database — psycholinguistics

During the past 6 years I spent a lot of time in acquiring how to analyze data properly. Most experience comes from the analysis of psycholinguistic data sets. I was involved in the design, analysis, and interpretation of roughly 40 experiments, including questionaires, reaction time experiments, and ERP (event related potentials) studies . As for the statistical procedures, they share a large number of properties with customer research (marketing) and clinical studies.
Corpus studies (about 10) sometimes need different data treatments due to the non-linear distribution of the data.
Descriptive statistics (price range, market share, etc.) was part of the economic projects.

Software

Simple descriptive statistics and some procedures can be done with MS Excel or OpenOffice.org Especially when the diagrams get a bit more complex, MS Excel is easy to use and offers a lot of possibilities.

I was using SPSS for most of my studies. It is simply *THE* standard software.

SAS was the software appropriate to analyze the massive amount of brain data for my master thesis.

More and more I am convinced to use R for all kinds of analyses. For me, there are three major advantages:

It is open source software (no painful expensive licence renewals) runs with Linux.
It is more powerful in a sense of which procedures are possible and I have complete control about what I am doing. This holds for graphs, too.
I can do mixed models which much more powerful than classical ANOVA-style analyses as they involve less constraints to the data set, they test significance and indicate the size of the effects, and allow for careful investigations of fixed and random effects for a maximum of targeting the research question.