Second language (L2) writing research has demonstrated that young learners discuss linguistic issues, make use of feedback, and show a generally positive disposition toward writing tasks. However, many issues deserve further investigation. Regarding task implementation, few studies have been conducted with young learners writing individually, and few have compared different feedback types. Also, the focus of analysis has mostly been placed on students’ discussions; little is known about aspects such as draft quality and task motivation. To address these gaps, draft quality and task motivation were measured in 75 learners (aged 10–12 years) engaged in a three-stage writing task. They were divided into a task repetition group (n = 21), a group that received feedback via direct corrections (n = 30), and a group that received feedback via model texts (n = 24). Students’ drafts improved very slightly in complexity with task repetition; they significantly improved in accuracy with direct corrections, and also in lexical diversity and text quality with models. Task motivation remained high throughout the writing cycle in the task repetition group, where students found the task easy. However, it dropped slightly with direct corrections and – especially before the final draft, with models – which students deemed conducive to learning but difficult. Some pedagogical guidelines are discussed.

第二语言（L2）写作研究表明，年轻学习者讨论语言问题，利用反馈，并对写作任务表现出普遍积极的态度。然而，许多问题值得进一步研究。在任务执行方面，很少有针对年轻学习者单独写作的研究，也很少对不同的反馈类型进行比较。此外，分析的重点主要放在学生的讨论上；对草案质量和任务动机等方面知之甚少。为了解决这些差距，我们对 75 名参与三阶段写作任务的学习者（10-12 岁）的草稿质量和任务动机进行了测量。他们被分为任务重复组（n = 21），一组通过直接纠正获得反馈（n= 30），以及通过模型文本收到反馈的一组（n = 24）。随着任务的重复，学生的草稿在复杂性上略有提高；通过直接更正，它们的准确性显着提高，模型的词汇多样性和文本质量也显着提高。在任务重复组的整个写作周期中，任务动机仍然很高，学生们发现任务很容易。然而，直接更正后它略有下降——尤其是在最终草稿之前，有模型——学生认为这些模型有助于学习但很困难。讨论了一些教学指南。