| This study investigated the effects of L2 writing proficiency levels and thinking aloud on cohesive device use and composing behaviors of Filipino first-year college students (N = 30). Participants came from three intact classes and were grouped by the researcher into low (Level 1), mid (Level 2) and high (Level 3) levels of writing proficiency based on their performance on holistically scored writing tasks. Sixty essays written in non-think-aloud versus think-aloud conditions on two different writing prompts were analyzed for cohesive device use. Two raters coded the think-aloud protocols for reported composing behaviors.;Repeated-measures analysis of variance compared the frequency of use of grammatical and lexical cohesive devices per 100 words across writing levels, between writing conditions, and tested for any interaction between the two independent variables. One-way analysis of variance compared the frequency of occurrence of composing behaviors, i.e., Metacomments, Retrieving, Writing, Reading, Organizing/Deciding, and Revising, reported in the think-aloud protocols.;Results showed no significant difference in cohesive device use across writing levels, except for hyponymy. Results also showed no significant difference in the frequency of occurrence of most cohesive devices in the two writing conditions, except for demonstratives, comparatives, temporals, repetition, synonymy, and naming. There was no significant difference in the frequency of occurrence of reported composing behaviors across writing levels.;The results suggest the following conclusions: (1) advanced-level Filipino ESL student writers were comparable in terms of their cohesive device use and composing behaviors; (2) there was little evidence of reactivity from thinking aloud and/or the writing prompt having a significant effect on cohesive device use in L2 writing; (3) holistically rated writing tasks may not accurately distinguish differences in the frequency of occurrence of linguistically discrete textual features such as cohesive devices; (4) cohesion analysis was only marginally useful in characterizing the L2 writing proficiency of advanced ESL writers; and (5) advanced ESL writers showed deep commitment to their written product and low involvement with evaluative composing behaviors. |
