Archeologists can be very particular when choosing equipment. We prefer Marshalltown trowels, Brunton compasses, Sharpies, Trimble GPS units, and Munsell color charts. We spend a lot of time using this equipment, so such preferences are understandable. I have also thought a bit about the software that I use to analyze data once I’ve gotten out of the field and have processed my finds. A good software package for graphics and analysis can greatly speed report production and open many possibilities for the rigorous documentation of variability in the data.

Among commercial statistical software, I am a big fan of Statistica. I particularly appreciate the graphical capabilities of Statistica. Statistica can produce a wide array of graph types out of the box, and these options are all easy to customize. I have gotten considerable mileage from its options for categorized histograms, for example. Statistica also offers a wide range of analyses, available in different optional packages. Lately, however, I have begun the transition to R.

R is a statistical computing program. As such, it is extremely flexible and powerful. R is command-line driven, which makes it very easy to document and replicate the steps that you undertook during an analysis. If you’ve ever had to conduct an analysis several times, you will understand the appeal of this feature. Graphics can be generated in R, and they can be extensively customized. A lot of packages for it have been developed, providing functions by which you can conduct an extremely broad range of analyses. R is also free, which is another compelling reason to adopt it. R has an avid group of users, so you can have confidence that the program will continue to be supported in the future. The downside, however, is the steep learning curve.

To aid novices, lots of guide books to R are available, and some of them are quite good. I recommend starting with Introductory Statistics with R. With this book, you should be able to jump right into generating basic graphics and running common statistical tests. Another really useful book is Ecological Models and Data in R. This book covers model-building and maximum likelihood methods for evaluating the fit of data to those models. While the book’s examples reflect the interests of an ecologist (obvs), the book focuses on developing skills that transcend any particular discipline. The book is quite well-written. The presentation is also sufficiently detailed that you can very easily use it as a point of departure for playing with your own data and models.

Whatever you choose, make sure that the program offers a lot of flexibility. The analyses that I have done varied quite a bit from one research project to the next. Any program that you obtain will require a considerable investment—both financial and intellectual—so you’ll want to have that investment pay off again and again.

© Scott Pletka and *Mathematical Tools, Archaeological Problems*, 2009.