The problems of classroom teacher-conducted research are examined with reference to a particular case study. The case study involved a field test of "A Practical Guide for Advanced Writers in ESL" over a semester in two sections of an intermediate English as a second language (ESL) writing class for college-bound students at the University of Houston. Both the experimental and control sections kept error charts emphasizing spelling, diction, and syntax. T-unit analysis was used to assess the relative gains of the experimental versus the control sections. A major difficulty with the study was attrition, since 50 percent of one of the sections was lost during the course of the research. Results showed no difference between the control and experimental groups, although both registered substantial improvements. In order to further investigate the nature of these improvements it would be necessary to have a control group of students who did not receive writing instruction. However, it is impossible to justify teaching an intermediate ESL class of university-bound students without writing instruction. This is one of many limitations on teacher-conducted research. A final difficulty is the tendency of a teacher to hope for a certain set of results. (RW)