This qualitative study explores the cognitive processes of 10 6th-grade students of varying Spanish and English proficiencies while writing persuasive letters in both languages. The participants who were purposefully selected included: three high Spanish/high English proficient students (high performers), three high English/low Spanish proficient students (high/low performers), two low Spanish/low English proficient students (low performers), and two monolingual English speakers.

The following research question was posed: How do sixth-grade students of varying Spanish and English proficiencies engage in the writing process while composing persuasive letters in both languages? In particular, (a) How do students engage in the cognitive processes of writing in L1 and L2, and do the processes vary by language and/or ability? (b) Are there patterns across participants' writing behaviors in L1 and L2 that indicate cross-linguistic transfer? If so, do the patterns vary by language or writing proficiency? (c) Do bilingual students differ from monolinguals in their use of strategies and resources?

The researcher audio-taped and video-taped participants thinking aloud as they responded to persuasive letter prompts in both Spanish and English (monolinguals responded to two English prompts). Recall protocols and student interviews were also collected. All data were transcribed.

Data were analyzed in three phrases. First, think aloud sessions were coded for three general writing processes: composing, reading, and selecting. Second, recall protocols were coded for specific behaviors within each general process (e.g. attending to text generation, reading the text produced, considering or changing an idea). Finally, interviews and recall protocols were analyzed for student strategy, bilingual strategy, and resource use as well as general strategy and bilingual strategy awareness.

Findings suggest: (1) The writing process did not vary for bilingual and monolingual writers across languages; (2) Low performers focused attention almost exclusively on idea generation without attention to topical importance; (3) For high/low performers' knowledge of discourse features in L1 appeared to mediate writing in the weaker language; (4) Topically important ideas articulated in L1 during the selection process were abandoned if students did not have the vocabulary to express the idea in L2; (5) High performers seemed to intentionally separate their language resources while writing.