In this quasi-experimental case study, we compared five sections of a basic undergraduate technology course. Within an asynchronous web forum, pre-service teachers wrote short critiques of websites designed by their classmates. This peer feedback was provided anonymously by students in two classes (n = 35) whereas providers and recipients of peer feedback were identified by their real names in three other classes (n = 37). Computer-mediated discourse analysis methods (Herring, 2004) were used to code student written comments according to substance and tone of feedback. Next, we estimated likelihoods of specific feedback patterns through Analysis of Patterns in Time (Frick, 1990). Results indicated that students who were anonymous were approximately five times more likely to provide substantively critical feedback than were those whose identities were known to their recipients. When feedback was given anonymously, students were approximately four times more likely to provide reasons for needed improvement to a website, and then to suggest design alternatives. In light of advantages afforded by this form of pseudonymity, we conclude with a discussion of pedagogical prescriptions for supporting learners' production of feedback. (Contains 1 figure, 3 tables and 1 footnote.)