1) Research design, This study used a two-group quasi-experimental design to examine whether app-based spaced retrieval improves short- and medium-term retention of academic collocations. The independent variable was study condition (spaced-retrieval app vs self-study). The primary outcomes were gains on an immediate post-test and a delayed post-test. A brief questionnaire provided supporting data on learner perceptions and study behaviour., 2) Participants / data sources, Twenty-eight doctoral students from a large European university volunteered for the study (16 female, 12 male; mean age 29.4). Participants represented engineering, life sciences, and social sciences, and all reported writing in English at least monthly. Eligibility criteria were: (i) currently enrolled in a PhD programme and (ii) no prior systematic training in spaced retrieval. All participants provided informed consent and were free to withdraw at any time without penalty., 3) Materials / instruments, We used (i) a 40-item vocabulary test consisting of target academic collocations (e.g. pose a challenge, yield results), (ii) a short background questionnaire (discipline, L1, prior exposure to academic writing instruction), and (iii) a study log exported from the spaced-retrieval app showing total minutes studied and number of review sessions. The app delivered retrieval prompts and scheduled reviews automatically. All test items were piloted with a separate cohort (n=8) to check clarity and difficulty., 4) Procedure, Participants completed a pre-test and background questionnaire in Week 1. They were then assigned either to an app-based spaced-retrieval condition or to a conventional self-study condition, and both groups received the same list of target collocations. Over the following 14 days, the app group completed daily 10-minute review sessions, while the self-study group was asked to review the list independently for the same total time. An immediate post-test took place on Day 15 and a delayed post-test on Day 29; both tests were administered online under timed conditions., 5) Data analysis, Responses were analysed in two stages. First, we computed a retention score for each participant (immediate post-test minus pre-test; delayed post-test minus pre-test). Second, we compared mean scores across groups using an independent-samples t-test and reported effect sizes (Cohen’s d) with 95% confidence intervals. For the questionnaire, Likert-scale items were summarised descriptively and open comments were coded for recurring themes (e.g. workload, usability, perceived transfer to writing)..

Methods Section Structure

by
Print
Embed

Leaderboard

Visual style

Options

Switch template

)
Continue editing: ?