Project 2: Features of Student Writing in English 100

For your second project, you’ll work in small groups to analyze a corpus of written texts and develop an argument about discourse features you see therein.

Your data will be timed, impromptu, handwritten essays written by first-year UW-Madison students during the first weeks of English 100 in Fall, 2004.  You are not required to use all the texts in your analysis.  Nor, obviously, are you required to deal with every discourse feature of those texts.  Your group is free to make choices about data, method, and claims as you see fit.

You may want to proceed in the order that Huckin (1992) recommends, the first step of selecting an initial corpus of texts already done for you.  The second step is to scan those texts holistically, reading impressionistically and searching inductively for salient patterns therein, what Barton calls “rich features.”  The third step is to bring to the foreground those particular patterns that seem most “interesting” to you, most relevant to the questions you have about student writing, most useful to others in the field.  The fourth step (optional when the initial corpus is small) is to select a “study corpus,” some portion of the original data set that you believe is representative of the whole but still capable of being closely analyzed given your time constraints, etc. (for example, you might choose to analyze only the first 100 words of each text).  The fifth step is to systematically gather data from your study corpus to verify your original impressions, attempting to define with precision the “interesting” features or patterns you see, coding the data for instances, etc.  The final step is to develop a plausible argument about the data, establishing the significance of your analysis, illustrating it with examples, etc.  (Some methods, e.g., critical discourse analysis [CDA], might proceed very differently.)

There are several broad approaches you can take to this assignment, which we might categorize as qualitative descriptive, quantitative descriptive, and quasi-experimental.   The first two attempt to describe or interpret the data, or some subset thereof, through either numerical or verbal means.  Here, you might focus on a single problematic text, compare texts within the corpus according to some feature, develop an argument about patterns common to all or most of the texts, establish averages or means for the corpus, etc.  The third approach, by contrast, would attempt to relate discourse features in the data to some “external” set of variables, e.g., independent ratings of the quality of the texts or some other corpus (e.g., the writing of published authors or graduate students in English).  As for specific methods of analysis, you might try one or more of the following: content analysis; critical discourse analysis; “error” analysis; functional or pragmatic analysis; genre analysis; information structure analysis; linguistic analysis (psycholinguistic, sociolinguistic, etc.); rhetorical analysis; or text analysis.

For March 3, write as a group a single 5-10 page paper and prepare an oral presentation of around 15 minutes about your study; a few slides or handouts can be used.  Unlike your presentations for Project 1, this one should be more of an analysis, albeit tentative, and less a reflection or narrative about your experiences.