95% Coverage Threshold
The 95% coverage threshold is Laufer's (1989) claim that a reader needs to know roughly 95% of running tokens for adequate reading comprehension. The figure predates Hu and Nation's 98% threshold by more than a decade and remains the working target for much graded reader publishing because the 98% figure is steep and pushes vocabulary-size requirements to 8,000+ word families.
The Laufer study
Laufer tested Israeli L2 readers on academic-style English texts manipulated for known-word ratios. She found that comprehension scores rose noticeably as coverage approached 95% and that readers below this level struggled to achieve a passing score on standardised comprehension measures. The 95% figure was framed as the minimal lexical condition for adequate reading, not a guarantee. Coverage above 95% was helpful, but 95% appeared to mark the lower boundary of viable independent reading for the population studied.
Position in the two-threshold framing
Nation (2006) and Laufer & Ravenhorst-Kalovski (2010) reconciled the 95% and 98% figures by treating them as complementary thresholds rather than competing ones. 95% coverage is now positioned as the minimal threshold for adequate comprehension under favourable conditions: dictionary access, glossing, teacher support, strong topic familiarity, or instructional purpose. 98% is positioned as the optimal threshold for unassisted reading for pleasure. Laufer & Ravenhorst-Kalovski put the 95% target at roughly 4,000–5,000 word families plus proper nouns, against 8,000 for the 98% target.
The 95% figure dominates classroom practice. Most published graded readers are pitched to deliver 95–98% coverage for their nominal level, with the lower end of that range used to keep page counts and vocabulary gates manageable. Carver (1994), working in L1 contexts, reached compatible figures by analysing the percentage of unknown words in texts at different relative-difficulty levels, and his work is sometimes cited alongside Laufer to support a 2%-unknown-word ceiling for productive instructional reading.
Status
Laufer's 95% threshold is empirically narrower than Hu and Nation's 98%, but together they bracket the operational range for graded input. Schmitt et al.'s linear-coverage result complicates both by showing no abrupt discontinuity at either point.
References
- Carver, R. P. (1994). Percentage of unknown vocabulary words in text as a function of the relative difficulty of the text: Implications for instruction. Journal of Reading Behavior, 26(4), 413–437. https://doi.org/10.1080/10862969409547861
- Laufer, B. (1989). What percentage of text-lexis is essential for comprehension? In C. Lauren & M. Nordman (Eds.), Special Language: From Humans Thinking to Thinking Machines (pp. 316–323). Multilingual Matters.
- Laufer, B., & Ravenhorst-Kalovski, G. C. (2010). Lexical threshold revisited: Lexical text coverage, learners' vocabulary size and reading comprehension. Reading in a Foreign Language, 22(1), 15–30.
- Nation, I. S. P. (2006). How large a vocabulary is needed for reading and listening? The Canadian Modern Language Review, 63(1), 59–82. https://doi.org/10.3138/cmlr.63.1.59