As previously reported, we here at Open Logic Central have run surveys to gather some data on the relative success our open textbooks have with students. The first survey was done in several sections of Calgary’s Logic II course, where some sections of the course used Boolos, Burgess, and Jeffrey’s *Computability and Logic*, and some used the OLP-derived *Sets, Logic, Computation*. We’ve now run these surveys also in Logic I, our first course in formal logic. One section, taught by Aaron, used *forall x: Calgary Remix*, four others used either Goldfarb’s *Deductive Logic* or Chellas’s *Elementary Formal Logic*. *forall x* was provided free as a PDF; the hardcopy retailed in the bookstore for $12, Goldfarb for $54, and Chellas for $40. The highlights are below. But first a

**HUGE CAVEAT:** As encouraging as they are, of course, the result must be taken with a large grain of salt. The sections surveyed were taught by different instructors and with different TAs, at different times of day, in different terms, with different modes of delivery, and different modes of evaluation. Students’ attitudes towards a text used in a course depend not only on the text itself, but also on their experience in the course overall. If the material covered and tested in lecture doesn’t match up with the text, they may see it as a failure of the text. The quality of instruction may influence the perception of the quality of the textbook as well. Which direction? I think either is possible. If lectures are clear, students find it easier to understand the textbook, and so find it clearer. *Or,* if lectures are *unclear*, the textbook may look clearer by comparison.

As in the previously reported results in Logic II, students consulted the open textbook much more frequently than the other two textbooks.

They also consistently used *forall x* more for specific purposes.

*forall x* also ranks significantly better on every measure of textbook quality we asked about.

Last time I made the plots laboriously in Excel and Plot.ly. This time, the plots were done in R. (Still laborious, since I knew and still know pretty much zero about R.) If you’re curious exactly how, I wrote it up. Data and code on Github.