22 hours of bottom-up skills instruction helped some listeners and not others, depending on their prior listening style. Bottom-up work needs to be combined with top-down and metacognitive work.

Study

Yeldham, M., & Gruba, P. Toward an instructional approach to developing interactive second language listening. Language Teaching Research, 18(1), 33–53. (The reference list says 2014, the body says 2013. Check before citing; vol. 18 corresponds to 2014.)

  • 5 Taiwanese university EFL learners; a 22-hour bottom-up course over ~1.5 semesters.
  • Multiple case study: tests, recall, verbal reports, interviews.

Findings (qualitative)

  • The top-down listener improved substantially, as bottom-up skills curbed over-guessing. Segmentation still didn’t improve.
  • One bottom-up listener gained a lot (possibly confounded by outside test prep). Another still struggled.
  • The interactive listener’s total rose, but top-down use fell.
  • Conclusion: bottom-up instruction is “inadequate” alone; bottom-up and top-down processing need to interact.

Classroom moves

  • Pair segmentation, stress and discourse-marker work with prediction and inferencing.
  • Teach the metacognitive cycle: plan/predict → monitor → evaluate → problem-solve (Vandergrift & Goh). Have learners reflect on what misled them.
  • Gap-completion: finish sentence endings, guess masked segments (noise over the recording).
  • Diagnose each learner’s tendency (top-down guesser vs bottom-up decoder).

Limits

n = 5, one setting, no comparison group, a possible confound.

Apply in my context (Claude)

  • IELTS Listening punishes over-guessers in form/note completion (exact spelling, word limits). Diagnose with a quick dictation: guessers produce plausible but wrong words, decoders produce fragments.
  • A ready-made DELTA listening LSA: problem = top-down over-reliance; solution = Scaffolding dictation for beginners style segmentation plus metacognitive reflection. Background essay citations: Field 2008; Vandergrift & Goh 2012; this study.

Connections