Observation of student teachers and graduate teaching assistants in the second language classroom has long-been an essential component of teacher preparation and assessment. However, the guidance and support given to novice teachers during the teacher-preparation program is unfortunately a limited commodity. In all too many cases, the tremendous support surrounding the student teacher throughout his or her practicum comes to a screeching halt as soon as a degree is conferred. For this reason, the present article suggests that instead of focusing upon in-class observation as an assessment tool, observation combined with teaching diaries can provide student teachers with a unique type of feedback, as well as give them practice with a mechanism of auto-assessment and reflection that can be used throughout their careers. A review of traditional observation techniques provides the point of departure.
For decades, systematic observation of pre-service teachers was the technique of choice. Indeed, the ESL and language-teaching profession's "need for objectivity" (Allwright, 1988, p.44) during the 1960s and 1970s gave rise to a number of systematic category systems such as Flanders' interaction analysis (1967) and Moskowitz' Flint system (1976).
Category systems were designed to produce accurate records of classroom events by recording and categorizing teacher and student actions at regular intervals--three second intervals in the case of Flanders' interaction analysis. While developing the FLint system, Moskowitz augmented the number of categories from Flanders' ten to an elaborate twenty-two. However, as objective as a detailed, and possibly exhaustive catalog might appear, observation systems are not without flaws. In fact, Allwright (1988) believes that category analyses have failed to provide a "sufficiently illuminating account of the data" (p.193). For Bailey (1975), the "ambiguities of the category definitions, the problem of making rapid decisions, [and] the inability to overcome certain prejudices are all [difficulties] which the [observer employing a category system] must confront" (p.337).
Probably the most compelling problem with Flanders' system is the assumption that an external observer will be able to successfully detect and record a teacher's "understanding" or "acceptance," as is required for the first and third categories of the system. In other words, category systems attempt to observe the unobservable and to count the uncountable; when in fact, the "flow of classroom life is actually under the surface [...] and we must doubt the integrity of the observable" (Breen, 1985, p. 149).
A Reaction to Category Observation
Given that the language-teaching and teacher-preparation professions have grafted asocial research techniques onto a distinctly social situation--and that many teachers and researchers have equated valid and useful research to large-scale comparative studies--attempts to examine the individual learner or teacher have been met with skepticism (Allwright, 1988; Bailey, 1990; Reseigh Long, 1982). Moreover, many important classroom issues have been ignored simply because they do not fit into the narrow confines of what is researchable according to traditional techniques (Reseigh Long, 1982).
Opposed to the rigid, systematic, and numeric vision of the classroom espoused by proponents of category systems, or of those belonging to the hard-core quantitative tradition, are the teachers and researchers who treat the classroom not as a laboratory, but as a unique, viable culture. (Bailey, 1990; Breen, 1985). According to Breen (1985), there are too many aspects of the classroom language learning situation that are neglected by so-called objective or quantitative research techniques. Instead, the classroom is a unique social situation that must be treated as such. Clearly, "no human social institution or relationship can be adequately understood unless account is taken of expectations, values, and beliefs that they engage" (p. …