Who is in control?: topic modulation in spontaneous L2 writing : interest, confidence, fluency, and complexity

The present study examined the differences manifest in the writings of third semester German language learners when topic selection was modulated between student and instructor. Students enrolled in third-semester German (four intact classes taught by three different teachers) wrote in an in-class journal for 10 minutes each week for an eight week period. One half of the time, participants selected their own topics, the other half of the time, they were assigned a topic by their respective instructor. To account for order of treatment, two of the four groups were counterbalanced with the other two. As an intermediate variable, participants were asked to indicate their level of interest in topics that they either selected or were assigned for each writing session. Additionally, participants indicated a general self-appraisal of the quality of each written product (referred to as confidence in one’s own writing). Each of these was indicated with a 6-point scale (1=lowest interest/evaluation, 6=highest interest/evaluation). Each written product was textually analyzed and results were categorized into a general fluency index and an overall grammatical complexity score. The results of these indices were correlated with the intervening variables of interest level and self-appraisal of written work, and an analysis of variance (ANOVA) was performed to see if topic control modulation influenced any dependent (fluency/complexity or intermediate (interest/self-appraisal) variable. ANOVA results indicate that topic control did influence participant’s written fluency but not grammatical complexity. Participants’ overall level of fluency was significantly higher when they selected their own topics. Interest in a given topic showed no significant correlation with complexity in writing, except in the only class taught by a native speaker of German. There was a correlation between participants’ confidence in written products and writings which were elicited from assigned topics (i.e., participants indicated higher levels of confidence in their writing when topics were assigned). The only exception to this finding came from one group which showed correlations with both assigned and self-selected topics.