WCPM Calculator Logo
WCPMCalculator

Words Correct Per Minute

Guides

Good Reading Accuracy Percentage

A good reading accuracy percentage depends on the source: 95% (running records), 96% (DIBELS 8), 99% (Betts). Includes an errors-by-passage-length table.

Published October 20, 2026

By Last updated

A good reading accuracy percentage is usually 95% or higher, but the sources do not agree. A running-record guide treats 95% as independent. DIBELS 8 sets its core-support line at 96%. Betts (1946), as summarized in a review, asked for 99%. To check your own numbers, use the reading accuracy calculator, then read on to see which convention fits your situation.

The formula

Reading accuracy is the share of words a student read correctly:

Accuracy % = (words read − errors) ÷ words read × 100

DIBELS 8 words it as words read correctly divided by total words read, including those that were wrong, times 100. Count self-corrections according to your manual. The Hertfordshire guide below does not count them as errors, and DIBELS 8 scores a word corrected within three seconds as correct.

The conventions, side by side

Teachers get conflicting numbers because several traditions exist. Here is what each source says, with the numbers as printed. Hertfordshire is a UK local-authority service, so no US running-record guide is quoted here.

ConventionTop bandMiddle bandLow bandSource
Running-record styleIndependent: 95% to 100% (5 errors or fewer in 100)Instructional: 90% to 94% (6 to 10 errors in 100)Frustration: below 90% (more than 10 errors in 100)Hertfordshire SPLD Outreach Service, “How to ensure a pupil is reading the right level book” (September 2018)
DIBELS 8 benchmark (ORF accuracy), grade 2 mid-year through grade 8Green, core support: 96 and aboveYellow, strategic support: 91 to 95Red, intensive support: 0 to 90DIBELS 8th Edition Administration and Scoring Guide, Appendix E
Betts (1946), as reported by Allington, McCuiston and BillenIndependent: at least 99% word readingInstructional: 95% to 98%Frustration: below 90% word recognitionAllington, McCuiston & Billen, “What Research Says About Text Complexity and Learning to Read”

A few things stand out.

The DIBELS 8 numbers are screening categories, not reading levels. They describe how much support a student at a given point in the year is likely to need, so a 93% in the yellow range is a flag for a closer look, not a verdict. Appendix E also has separate cut points for earlier grades and benchmark periods, so check the table for your student’s grade and time of year.

The Betts numbers, as reported by Allington and colleagues, leave a gap: the instructional range stops at 98%, and the frustration level starts below 90%. The review does not give a label for 90% to 94%. Shanahan’s blog post reports the same 99% to 100% and 95% to 98% bands for Betts, so the two secondary sources agree with each other. Both are secondary sources; the 1946 original was not consulted.

The 95/90 running-record bands and the 96/91 DIBELS 8 bands look close but cut at different points. A student at exactly 95% is independent in one system and in the yellow range in the other.

Reading accuracy by passage length

Percentages depend on how many words were read. One error costs 1 point in 100 words, but only 0.5 points in 200. Use this table to read a score without a calculator. Each cell is (words − errors) ÷ words × 100, rounded to one decimal.

Errors100 words150 words200 words
199.0%99.3%99.5%
298.0%98.7%99.0%
397.0%98.0%98.5%
496.0%97.3%98.0%
595.0%96.7%97.5%
694.0%96.0%97.0%
793.0%95.3%96.5%
892.0%94.7%96.0%
991.0%94.0%95.5%
1090.0%93.3%95.0%
1189.0%92.7%94.5%
1288.0%92.0%94.0%

Six errors is 94.0% in a 100-word passage and 97.0% in a 200-word passage. The same number of mistakes can land in different bands, which is one reason to compare accuracy across passages of similar length.

Worked example: accuracy vs. WCPM

A student reads for a full minute, reaches word 112, and you mark 6 errors.

  • Words correct: 112 − 6 = 106
  • Accuracy: 106 ÷ 112 × 100 = 94.6%
  • WCPM: 106 × 60 ÷ 60 = 106 (the WCPM formula is covered in its own article)

The two scores answer different questions. WCPM says how many words were read correctly in the minute. Accuracy says what share of the attempted words were right. Another student could also score 106 WCPM by reading 120 words with 14 errors; their accuracy would be 106 ÷ 120 × 100 = 88.3%, in the lowest band of every convention above. Same speed score, very different reading.

Where does 94.6% land? Reported as a whole number, 94.6% is 95, which falls in the DIBELS 8 yellow range of 91 to 95. The Hertfordshire guide says to round to the nearest whole number (its own example turns 92.7% into 93%), so 94.6% becomes 95% and counts as independent. You can enter the counts in the WCPM calculator for the speed score and the reading accuracy calculator for the percentage. The running record calculator shows the same counts as an error ratio, and it is the better choice if you are scoring a full text with self-corrections.

What to do when accuracy is low

A low score is a starting point. A few practical steps:

  1. Re-check the count. Confirm the word total, the errors, and whether your manual counts insertions and self-corrections.
  2. Try a second passage. One reading can be thrown off by a hard text or a bad day. Compare passages of similar length so the table above applies.
  3. Look at the errors. Do they cluster on multisyllable words, vowel teams, or sight words? The pattern tells you what to teach, which the percentage cannot.
  4. Check the text level. In the running-record and Betts bands, the lowest band marks a text that is too hard for independent reading, so move down a level or give more support. Shanahan, whose post is cited below, argues the Betts cutoffs lack solid research support, so treat all of these bands as rules of thumb.
  5. Look at speed too. Accuracy and WCPM together give a fuller picture. The oral reading fluency norms list WCPM by grade and season, and what is reading fluency covers accuracy, rate, and expression as parts of one skill.

For an IEP or intervention decision, accuracy is one data point. The DIBELS 8 guide itself says no single test should drive high-stakes decisions.

Frequently asked questions

What is a good reading accuracy percentage? It depends on the convention you follow. A running-record guide from Hertfordshire’s SPLD Outreach Service calls 95% to 100% independent and 90% to 94% instructional. DIBELS 8 lists 96% or higher as core support from mid-grade 2 through grade 8. Betts (1946), as summarized by Allington and colleagues, set 99% or higher for independent reading. Use one convention and apply it every time.

How do you calculate reading accuracy percentage? Subtract the errors from the total words read, divide by the total words read, and multiply by 100. DIBELS 8 words it as words read correctly divided by total words read, times 100. For 112 words read with 6 errors: 106 ÷ 112 × 100 = 94.6%.

How do you calculate accuracy in reading from a running record? Count the words in the text the student read, count the errors (self-corrections are not errors in the Hertfordshire guide), subtract, divide by the words read, and multiply by 100. A common shortcut is to express it as errors per 100 words: 5 errors or fewer in 100 words is 95% or better.

What is the difference between reading accuracy and WCPM? Accuracy is the share of words read correctly. WCPM is the number of words read correctly in one minute. A student can have high accuracy and a low WCPM if reading is slow, or a high WCPM with weak accuracy if the student rushes. DIBELS 8 reports both for oral reading fluency.

Is there an ORF accuracy calculator? Yes. Our free reading accuracy calculator takes words read and errors and returns the percentage. The running record calculator uses the same counts if you also want an error ratio and a self-correction ratio.

Sources