Reading fluency — guided repeated oral reading, and whether repetition matters

Verdict

Guided oral reading — reading connected text aloud with feedback — reliably builds reading fluency, which is a real, teachable domain skill and a plausible bridge between decoding and comprehension. Two important qualifications: (1) the widely cited large effect sizes for repeated reading are substantially a practiced-passage measurement artifact, and when exposure is held equal, repetition itself adds nothing over reading different text of the same volume — the mechanism is volume/practice, not rereading; and (2) sustained silent reading (just giving kids time to read on their own, or exhorting them to read more) has no experimental support for improving fluency or achievement.

What the evidence shows

The Therrien values look impressive until you notice they are measured on the same passage the child practiced. Transfer to novel text is roughly half as large for fluency and small for comprehension. Ardoin's RCT is the clarifying design: holding reading volume constant, rereading the same passage does no better than reading varied passages — so the fluency gain comes from the practice volume, and the big repeated-reading numbers are inflated by proximal measurement.

Hereditarian-lens assessment

Risk: low. The core contrasts (guided oral reading vs control; repeated vs wide reading) are experimental, and the decisive Ardoin comparison is an RCT, so selection/heritability cannot drive the between-method results. As everywhere in reading, the trainable domain skill (accuracy, rate) moves most and the g/language-loaded outcome (comprehension) moves least.

Fadeout & durability

Suggate 2016 puts fluency interventions in the fade-prone group (d≈0.47 → 0.28 at ~11 months) — better-maintaining than phonics but well short of full persistence. No long-horizon RCT isolates fluency-instruction durability.

What critics say / limits

• The "repeated reading is a uniquely powerful technique" framing is undercut by the equal-volume RCTs: it's the reading, not the repeating.
• The credible group-design base for fluency is thin (WWC found only 2 group-design studies met standards for students with LD), so the moderate grade reflects a modest evidence base, not a contested one.
• Comprehension gains from fluency work are small and, like everywhere, presumptively fade-prone.

Practical guidance

• Use guided oral reading with feedback (model the passage, correct errors, set a goal) rather than silent independent reading time as the fluency-building activity; an adult tutor beats a peer.
• Prioritize volume of connected-text reading at an appropriate level over drilling the same passage many times.
• Don't count on sustained silent reading as an instructional lever — there is no experimental evidence it raises fluency or achievement (which is not an argument against reading for pleasure, only against treating it as a proven intervention).

Open questions

• The exact transfer-to-novel-text effect sizes (fluency ~0.50, comprehension ~0.25) are commonly cited but under-verified from primary sources.
• Whether fluency practice adds value beyond the decoding and volume it delivers — i.e. is "fluency" a separate lever or just decoding × practice — is not cleanly identified.