Best books for learning STATA?

1
1465
Posted by Cory Schires, community karma 1465

I'd like to become more proficient with STATA (so not SPSS or SAS), and am looking for book recommendations that have:

  • Exercises that use real datasets (government datasets, open source data, etc.)
  • Designed for Mac STATA
  • Focus on regression analysis (versus longitudinal, MLM, etc.)
  • This is a pie-in-the-sky characteristic, but if the book came with access to instructional videos, that would be ideal

Any book (or, heck, any sort of learning medium) recommendations would be most appreciated!

about 12 years ago
Tags: stata

2 Comments

1
91
Michelle Naffziger-Hirsch, community karma 91

In case you haven't already come across them, my go-to for STATA has always been the UCLA Academic Technology Pages for STATA: http://www.ats.ucla.edu/stat/stata/default.htm. They're searchable which is a big plus. I must admit, I've never used Mac Stata, so I can't speak to whether it is helpful there or not.

about 12 years ago
login to leave comment
0
227
Will Hauser, community karma 227

Cory, 

I love Stata as much as a man can love a statistical package, so it's always great to see someone interested in learning it over less powerful (i.e. SPSS) or more confusing (i.e. SAS) programs.

The UCLA web resource Michelle directed you to is great.  Their examples are useful and always use freely available datasets so you can replicate their results.  I am not sure but I suspect that many (or all?) of the pages are written by graduate students so I do sometimes find errors or ambiguities but not often.  Not every subject is covered but the basics, such as OLS regression (model fitting, diagnostics, etc.) are.  Sometimes you do find pages that are outdated - i.e. written for Stata 8 which isn't terrible since anything you can do in 12 you can do in 8 but not vice versa (the very useful factor notation and corresponding margins command come to mind).

There is also the Statalist, you can try searching through the archives for an anwer to your question or post your question.  They're big on etiquette there so make sure you follow all the rules to the letter to ensure the best chance of getting a reply.  I find the listserv is best for simple technical sorts of questions like how to use a foreach loop for a particular problem or what to make of an unexpected error message.  You can see the listserv (and rules) at http://www.stata.com/support/faqs/res/statalist.html 

What's great about Stata is how open it is.  Users write their own programs and share them *freely* with the community.  Sometimes simple but irritating problems like reverse coding an ordinal variable or mean centering variables are easily solved by a user written program.  Such programs are easily found and installed from within stata (the findit and net search commands).

The documention is really great and the syntax is intuitive.  You can usually execute commands that you might not yet fully understand just by reading through the documentation for that command.  Experimenting is a great way to learn.  The documentation for a command always concludes with a few examples using freely available data.

As for books, see Stata Press, they maintain a nice list of books on various topics that are written specifically for Stata and have examples of code/syntax use freely available data that can be easily accessed from within Stata.  You can see Stata Press at http://www.stata-press.com/books/

I haven't read a Stata specific book on OLS regression so I can't make a specific suggestion for you.  I do know that "A gentle introduction to Stata" has been well-received and is widely used in introductory statistical courses.  Books by J. Scott Long are great reads - the categorical dependent variable book is excellent (if you have a categorical dependent variable).  Long's [free] SPOST package is a must for categorical data analysis.  There are some great tips on data management in Long's book on Workflow.

 

edited to say:  While I haven't used Stata for Mac, it should have precisely the same functionality as the Windows version and the syntax will be the same - you will be doing everything using syntax.

almost 12 years ago
login to leave comment