Statistical software offers of course a wide range of statistical analyses. Even without knowing a particular software it will be quite easy to produce let's say a logistic regression or a discriminant analysis, browse the menus, look up the command and your done.

But before you can apply statistical tools to data, data needs to be collected and prepared for analysis. And results (tables and charts) produced will have to be integrated into your research reports. These tasks will be in practice much more complex and you will usually spend much more time setting up the data and write your reports than producing statistical analyses.

Statistical software helps you with all this and needs to be integrated into your work flow. Therefore my advice is to acquire data management and data preparation skills with the software you know best and that is easily available to you (and support is easily available as well) and learn to integrate the output produced by that software into your writing process. If required add another statistical package to your toolbox, but only for performing analyses not provided by your favourite software; keep data management/preparation with your main software.

To analyze data, you need a dataset prepared for it: Sometimes you will need to create a new dataset, i.e. data needs to be entered (data entry facilities)

Software packages like SPSS; Stata, SAS, R and several others have strong data preparation, data management and database integration features, and provideeverything needed to transform your variables, i.e. perform tasks like recoding variables, creating complex composite scales from variables, creating subsets of variables, aggregating data sets and so on. With the exception of some trivial tasks this kind of problems are not simply solved by calling up some menu and selecting a few items or write a simple command; there you will have to invest much more time to learn all this. Therefore my recommendation to concentrate on a particular software for this task and learn it seriously.

• Today is quite easy to go from one statistical package to another, for instance SPSS can read and write without any problems SAS, Stata and other formats, i.e. you can easily prepare your data for analysis in SPSS, then write the dataset to a Stata file and perform a particular analysis, not found in SPSS, in Stata....
• Spreadsheets (e.g. Excel) are often more useful than statistical software to prepare data from the Internet, presented in tabular form and meant for publication; it is worthwhile to acquire some basics skills with spreadsheets as well.

All statistical software produces output of course, but there is a wide range of facilities offered; between software like SPSS that integrates a full fledged output editor and viewer that lets you not only change elements like colour in a chart, but lets you rearrange tables, move columns and rows around, apply styles and could be used to produce reports for publication and software, like Stata or R for instance. where the only way to integrate output into your reports is by cutting and pasting text ouput and exporting and importing charts.

Here of course, if SPSS is you choice you can spend some effort on the SPSS Output viewer, but you will be dependent on SPSS to use these teatures in the future. An alternative is of course to continue to rely on your favourite word processing program for this, especially if you already have some skills there...the choice is yours. However, as always a middle way combining both is often the best solution

• [Advanced stuff! Requires time and special skills!] Regarding R: If you think that in your future you will use statistics very frequently for your work and are not afraid to learn to use Teχ or Lateχ there is a feature that lets you integrate the production of statistics into the writing process (i.e. the commands producing statistical output are embedded in your documents and before viewing/printing the document R is called up to produce the tables and charts required.