How TOEFL Speaking Is Scored: What the Data Shows

How TOEFL Speaking is graded, band by band

What an average Take an Interview answer looks like at each AI score band. Higher bands run longer and faster and score higher on Language Use and Topic Development; from 2.5 up, they also pause less.

Per-answer averages from LingoLeap’s AI-scored practice, not official ETS grades; bands below 2.0 are excluded. Click a band for its full breakdown, or compare all the bands.

The three dimensions, in one line each

Méthodologie

Échantillon

55,614 answer records from 14,884 TOEFL 2026 Take an Interview practice sessions on LingoLeap; the headline analysis reports the 54,247 answers in the 2.0–4.5 AI score bands. The number of distinct learners is not recorded.

Période

From 1 January 2026. The end of the extraction window is not recorded in the data file.

Source

A read-only, aggregate-only analysis of LingoLeap's own practice records. No transcripts, recordings, user IDs or individual answers are published.

Production des scores

Delivery timing and the delivery sub-scores (0–100) come from automatic speech assessment (Azure). The answer score (0–5) and the Language Use and Topic Development signals come from LingoLeap's AI rubric grader, mapped to the public TOEFL Speaking dimension names. None of this is ETS scoring.

Exclusions

Bands below 2.0 are excluded: very short or truncated answers produce speech-rate readings of 200–380 wpm from one or two words. The 5.0 band (n = 29) is too sparse to report. The speech-rate view keeps buckets with at least 100 answers, up to 180 wpm. 3,922 records are marked unattempted; whether any of them fall inside the reported bands is not recorded.

Limites

Observational and descriptive: no significance tests or causal claims, and repeated learners and prompts are not independent. AI scores may differ systematically from human rating, and the three dimensions may correlate with each other.

Pattern: all three dimensions rise together with the score

On one shared 0–5 scale, the three dimensions move upward together across AI score bands; this analysis does not infer causation.

<figure style="margin:0;max-width:660px">
  <img src="https://lingoleap.ai/research/charts/toefl-speaking-scoring-dimensions.svg" alt="Figure 1. The three dimensions by AI score band" width="660" height="360" style="max-width:100%;height:auto" loading="lazy">
  <figcaption>Source: <a href="https://lingoleap.ai/research/toefl-speaking-scoring">LingoLeap, How TOEFL Speaking Is Scored: What the Data Shows</a></figcaption>
</figure>

Worth noting is Topic Development: often assumed to be hard for a machine to measure, yet here elaboration and relevance are measurable and climb clearly with the score. High answers aren’t just well-delivered — they say more.

Want to see what an answer at each band actually looks like? Compare the score bands (2.5–4.5) — real average words, pauses, pace, and elaboration for each.

The chapters (one dimension at a time)

What this study cannot show

• Dimension names borrow TOEFL’s public framework; the numbers are LingoLeap AI signals, not ETS scoring, and reproduce no rubric text.
• Observational, descriptive data: denominators and differences, no significance tests or causal claims.
• Scores are AI-assigned and may differ systematically from human rating; the three dimensions may also correlate with each other.
• Bands below 2.0 are excluded for short-answer artefacts.

Données

Comment citer

Questions we hear from learners