Software | Advice to doctoral students

Update to "Tools of the trade 1 -- Spelling and grammar checkers"

Shortly after I posted Tools of the trade 1—Spelling and grammar checkers, version 3.0 of the Hemingway Editor became available. The changes are evolutionary, rather than revolutionary, but provide some useful new features. The update is free to owners of version 2, which is generous for an update of this magnitude.

Read more

Tools of the trade 1 -- Spelling and grammar checkers

This is the first of a series of posts on "Tools of the trade." It will focus mostly on software, although you can see my thought on using a Mac rather than a Windows computer here .

Many, perhaps most, of us choose our tools fairly casually. Inertia may be the primary determinant of what we use. However, scholarship places different demands on our tools, so a more deliberate consideration of options can be worthwhile. The word processor or presentation software you've used for years may or may not be the best choice for your scholarly work. Also, some useful tools like bibliographic information managers (Zotero, EndNote, etc.) are probably outside of your prior experience.

This first post in the series is about grammar & spelling checkers. Most word processors include at least a spell checker and some, e.g. Microsoft Word, also include a grammar checker. Still, I see two reasons to consider a separate grammar & spelling tool. First, we do much of our writing outside of word processors, including emails and web-based forms. Second, while all the spelling checkers I've encountered are competent, the grammar checkers are more limited. I've experimented with two grammar checkers, Hemingway and Grammarly. Both are useful but fulfill different roles.

Read more

Projects feature of Stata 13

One of the features of Stata 13 is “Projects”, which are meant to provide easier access to multiple files related to a, well, project you are working on. The files can be do files, data, logs, graphs, etc. In fact, they don’t even need to be Stata files. One advantage I have found is that they make it possible to maintain a strict organization of certain types of files going in certain directories, while still having access to all of those files from one pane within Stata. Read more

UPDATE OF "What statistical package should I use?"

Technological progress continues. In an older posting, I mentioned the role of specialized packages that addressed models not available in the general purpose software, such as LISREL for structural equation modeling (SEM). That example is now somewhat moot, as Stata 12 has an extensive SEM capability and new add-ons for R allow modeling of SEMs. I suspect that if I were a power user, I would find limitations in Stata/R relative to the dedicated packages, but at my level, I haven’t found them. Read more

Excel is evil

Excel has caused more trouble for more doctoral students than I care to think about. Doctoral students can hurt themselves with Stata in at least two ways (there may be more).

  • Using it to clean, combine and otherwise manage data
  • Cutting and pasting results into Excel (or worse yet, Word) and then formatting them for presentation

Both of these a very inefficient uses of time. The first is a disaster for data integrity, because it is hard to document, almost impossible to revise, and very easy to mess up (sort only have the variables, be one row off when pasting, etc.) I briefly touched on data management in another posting and will probably write more in the future.

The second use of Excel is also prone to mistakes, although they are probably more easily corrected than butchering your data in Excel. Fortunately, there are many better approaches. Read more

A template for Stata .do files

There are many different approaches to writing and documenting the many steps that go into an empirical project. J. Scott Long has a great book, The Workflow of Data Analysis Using Stata, which I strongly recommend. He recommends developing a series of small, highly focused do files, which are run in sequence as needed. I take a different approach, which is keep all of a project’s code in one honking large do file, which is divided into sections. This posting provides more details about my approach. Read more

What statistical package should I use?

This is an amazingly contentious question. My first answer is "If you are comfortable with a package and it is serving your needs, keep using it." That can be complicated, of course, if you have a co-author dedicated to a given statistics package. If your only need to is pass data back and forth with that co-author, I strongly recommend Stat Transfer, which can convert from pretty much any statistical format to any other. Another consideration is the package most frequently used in your field. Read more