Dedicated to providing information for learning assistance professionals.

  Barbara McLay

Reading Rate: Research versus Teaching

By Barbara McLay, University of South Florida


There are thousands of studies on reading rate, going back for decades and continuing to current times, yet college reading-improvement textbooks continue to be filled with misinformation based on folklore. Reading only a very few studies, an interested party would discover that most of the techniques for improving rate given in college textbooks are not at all helpful to poor readers, and may cause good readers to try techniques that significantly reduce their comprehension

Even the briefest, most cursory, look at good research studies on rate will prove that (1) no one can read an entire page of print, an entire line, or even half a line of print at one fixation—a fixation being the stop to take in information as your eyes sweep across the lines of print; proficient readers take in information one word at a time, (2) silent speech, inner speech, or subvocalization—hearing words in one’s mind—is not something we should be trying to eliminate, and (3) reading is not a uniform process across materials and purposes for reading.

Reading Processes

Reading is not one uniform process. So when commercial speed reading courses promise to have students reading at ridiculously inflated rates, they are teaching readers to skim—to skip some sentences or words as they read. Skimming reduces comprehension to fifty percent or less. Most college reading textbooks do explain this even though commercial programs often promise higher comprehension with their various skimming techniques. Other processes are given various names: scanning, general reading or rauding, study reading, analytical reading, memorizing. Scanning, however, is a search for a word or a phrase on a page and is not really reading since the comprehension level is nearly zero. With scanning, it is not necessary to hold the words in short term memory for any time at all. Words that are examined are dropped instantly from memory storage if they are not the target word. It is a yes-no process.
Measuring reading rate is difficult and confusing because it is not always how fast the words are being processed that makes the difference in rate; it is what must be done to retain the information at various levels of comprehension. With skimming, words are skipped, but the words that are examined must be collected into a complete thought in order for the words to have meaning. Learning and memorizing require rehearsal and take more time still, as readers look longer at the words and spend more time going back to re-read words.

Researchers may tell their subjects to “read normally,” but the students may be using skimming, learning, or memorizing if the instructions or the materials do not explicitly and purposefully induce normal reading.

Eye Span and Fixations

Contrary to a plethora of scientific research, many college reading textbook continue to advise students to lengthen their eye spans and decrease their fixations in order to improve reading rate. Unfortunately, the ability to read complete lines, phrases, or thoughts is not supported by research. Taylor’s (1965) review of research reports that “…the average college reader has a usable span of recognition of only 1.1 words per eye stop and…even the most superior readers, trained or untrained, seldom achieve a usable span greater than 2.5 words” (p. 188). That every word must be processed and that we cannot read in complete phrases has long been supported by research.

Carver (1990) examined a 1937 study by Sisson that found no difference in numbers of fixations for poor and good readers, a 1946 study by Tinker that showed there was no evidence that eye training movement was helpful, and that such training could be harmful (pp. 174-177), a 1985 study by McConkie and Hogaboam that showed that the last word the students reported seeing was the word on which their eyes were fixated when the computer blocked out the following words (p. 107), and numerous other studies that showed that good readers read word-by-word or almost word-by-word. Carver concluded “Thus, it appears that it is virtually impossible for a reader to fixate upon the center of one standard length word and be able to report accurately another word that is further than one standard length word into the periphery during the operation of the rauding process” (p. 108). Carver (1990) also reported that a 1986 study by Rayner “found that the size of the perceptual span for the college students was limited to about 14 character spaces to the right of the letter in the word being fixated; this means that it was limited to the word being fixated upon plus the next two words to the right” (p. 109).

Brysbaert, Drieghe, and Vitu (2005) found that reading of literature did not involve a smooth continuous flow of processing words from left to right along lines of print. Words were skipped and many regressions occurred. Skipped words were predictable words. Regressions occurred when the predicted words did not occur, indicating that the periphery vision may have taken in the first letter or some letters of the non-fixated word. This finding does not conflict with research showing that fixations can take in about fourteen letters.

According to Sereno and Rayner (2003), neuroimaging has shown that fixation time is a large factor in reading rate. Processing time is directly related to length of fixation. “For example, words that are short in length, regular in their spelling—sound patterns, frequent in their occurrence, or semantically or syntactically predictable from a previous context are fixated for less time than those that are not” (p. 490). Rayner and Well (1996) also reported that highly constrained (predictable) words have shorter fixations than do unconstrained words and constrained words are also more likely to be skipped—assuming the words have fewer than five letters. Long words are unlikely to be skipped.

In a review of the literature, Rayner (1998) reported on studies by Just & Carpenter and by Just, Carpenter, and Masson that monitored “speedreaders” who claimed to read at 600-700 wpm. and compared these fast readers with normal readers. While speed readers could answer “questions about the gist of the passage,” they “could not answer questions if they had not fixated on the region where the answer was located. The normal readers, whose fixations were much denser than the speedreaders, were able to answer the detail questions relatively well” (p. 393). The eye movements of the speedreaders, as well as their level of comprehension, showed that they were skimming the material.

Subvocalization or Silent Speech

Legge, Pelli, Rubin, & Schleske (1985) presented text to college students at rates from 10 wpm to 300 wpm and asked them to read orally. The study reported that “For very rapid scans, the observers often finished speaking well after the scan was terminated. However, none reported any difficulty remembering the text to be spoken” (p. 10). This shows that the eye could see faster and the processing of the words could occur faster than speech. But it does not show that inner speech should be eliminated. Inner speech—subvocalization—does not operate at the same rate as vocalization. We can think words faster than we can speak them.

Rayner (1998), in a review of literature, reported that for oral reading, “mean fixation durations are longer than in silent reading, and the eyes tend to get ahead of the voice; consequently, there are many fixations in which the eyes appear to be holding in place so as to not get ahead of the voice” (p.375). Clearly, adults can read faster silently than they can read aloud; speaking the words aloud takes more time than thinking the words. Since the average reading rate for adults is faster than speech (Carver, 1990), vocalizing the words slows down the reading process, but this does not mean that subvocalizing, or silent speech does. In fact, subvocalizing is a necessary part of silent reading, and many studies show that rates may decrease when it is circumvented, and comprehension is almost always decreased (Carver, 1990; (Baddeley, Eldridge, and Lewis, 1981).

Correlation of Rate and Comprehension

Carver (1983) showed that reading rate is constant for individuals and changes only when readers move from rauding to some other reading process. Comprehension for each process is negatively correlated with the rate for using that process. One gets the most comprehension with memorizing, which is the slowest process, and almost no comprehension for material that is scanned, the fastest process.

That readers who read faster comprehend more than readers who read slower may have started with studies that looked at children’s reading. When looking at general reading for third grade readers, Joshi & Aaron (2000) found that reading rate—speed of processing—was positively correlated with reading comprehension. However, since early readers are still having to decode many words, this would affect their reading rates as well as their comprehension. Pausing to decode a word interrupts the processing of sentences into thoughts. So students who are still decoding to read, or adults who must resort to decoding because of limited vocabulary, will read slower and with less comprehension than students who can read smoothly and fluently.

Carter (1990) reported findings for a 1931 study by Greene, a 1958 study by Poulton, a 1966 study by Jester and Travers, a 1973 study by Keesey and many others, including studies he conducted himself. All the correlations between rate and comprehension for adults were high and negative. “Given the large amount of empirical support for the above theoretical relationships, it is wrong to suggest that individuals can improve their comprehension when they read if they will simply increase their rate….If you want to increase your comprehension, A, then decrease your rate, R” (p349).
speed” (p. 172).

Conclusion: How Can Reading Rate Be Increased?

Carver (1990) says that four things affect how fast adults read: apping, rauding rate, cognitive speed, and practice. Apping stands for “automatic pilot for prose.” This is the “rhythmic and habitual eye movements during ordinary reading” (p. 173). Most adults read at their rauding rate (Rr) when they are reading ordinary prose, reading normally. “However, some individuals, maybe 20% of college students for example, do not typically read at their Rr. Instead of operating their rauding process at their rauding rate, as about 80% do, they typically read at the rate of a learning process…. They have an automatic pilot for prose that is slower than their rauding rate” (p. 173). Frequent practice, reading easy material for enjoyment, “helps maintain the apping rate at the rauding rate” (p. 441).

SØvik, Arntzen, and Samuelstuen (2000) studied oral and silent reading performance of twelve year olds and determined that reading speed could be predicted by recognition span, fixation duration, and number of regressions. Since fixation durations are longer when the word is not familiar (Sereno & Rayner, 2003), it would seem logical that increasing the number of familiar words, or sight words—improving vocabulary—would be one way to improve reading rate. Regressions occur when students are intentionally learning or memorizing, or when readers move to learning or studying instead of rauding because the level of the material is more difficult than the student’s reading level. Building a larger reading vocabulary and increasing background knowledge should help to reduce the number of regressions. But the suggestion given in some college textbooks (Wassman & Rinsky, 2000) to use a card to cover the words as they are read, in order to prevent regressions,will most likely decrease comprehension and frustrate the reader.
Mary Dillard (2003) in her textbook gave good advice to students: “The more prior knowledge, experience, and interest you bring to a reading selection, the more rapidly you will read. So work on developing interest in a variety of topics, read more than you have ever read before, and use every opportunity you have to learn something new if you seriously want to read faster. No quick fix exists. Reading faster by reading down the center of the page or reading only the important words are not linguistically sound approaches” (p. 2).

In summary, to help students increase reading rates, encourage wide and frequent reading of easy material, and help students build background knowledge, vocabulary, and sight words. Do not encourage students to expand their word perception spans; do not force students to avoid regressions; and do not try to make them avoid silent speech when reading.


Baddeley, A., Eldridge, M., & Lewis, V. (1981). The role of subvocalisation in reading. Quarterly Journal of Experimental Psychology, 33A, 439-454.

Barker, T.A., Torgesen, J.K., & Wagner, R. K. (1992). The role of orthographic processing skills on five different reading tasks. Reading Research Quarterly, 27 (4), 334-345.

Brozo, W. G, Johan, J. L. (1986). A content and critical analysis of forty speed-reading books. College Reading and Learning Assistance Technical Report. 86-04.

Brysbaert, M., Drieghe, D. & Vitu, F. (2005) Word skipping: Implications for theories of eye movement control in reading. Oxford: Oxford University Press.

Buswell, G. T. (1951). The relationship between rate of thinking and rate of reading. The School Review, 59 (6), 339-346.

Carver, R.P. (1977-8). Toward a theory of reading comprehension and rauding. Reading Research Quarterly, 13 (1), 8-63.

Carver, R.P. (1983). Is reading rate constant or flexible? Reading Research Quarterly, 18 (2), 190-215.

Carver. R.P. (1990). Reading Rate: A Review of Research and Theory. San Diego, CA: Harcourt Brace, Jovanovich.

Demb, J.B., Boynton, G.M., & Heeger, D.J. (1998). Functional magnetic resonance imaging of early visual pathways in dyslexia. The Journal of Neuroscience, 18 (17), 6939-6951.

Dillard, M.L. (2003). Reading in the Academic Environment. Upper Saddle River, New Jersey: Prentice Hall.

Dyson, M.C. & Haselgrove, M. (2000). The effects of reading speed and reading patterns on the understanding of text read from screen. Journal of Research in Reading, 23 (2), 210-223.

Feifer. S.G. (1998). Neuropsychological features of dyslexia. Paper presented at the 30th Annual National Convention of the National Association of School Psychologists, Orlando, FL, April 14-18, 1998. ERIC document: ED 421 810.

Gould, J.D., Alfaro, L., Finn, R., Haupt, B., & Minuto, A. (1987). Why reading was slower from CRT displays than from paper. ACM SIGCHI Bulletin, 18 (4), 7-11.

Hancock, O. H. (2004). Reading Skills for College Students, Sixth Edition. Upper Saddle River, NJ: Pearson Prentice Hall.

Joshi, R.M. & Aaron, P.G. (2000). The component model of reading: Simple view of reading made a little more complex. Reading Psychology, 21, 85-97.

Legge, G.R., Pelli, D. G., Rubin, G.S., Schleske, M.M. (1985). Psychophysics of reading: I. Normal vision. Vision Research, 25 (2), 239-252.

Miller, W.M. & deOrozco S.S. (2001). Reading Faster and Understanding More Book Two. New York, NY: Longman.

Muter, P. & Maurutto, P. (1991). Reading and skimming from computer screens and books: The paperless office revisited?

Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124 (3), 372-422.

Rayner, K. & Well, A.D. (1996). Effects of contextual constraint on eye movements in reading: A further examination. Psychonomic Bulletin & Review, 3 (4), 504-509.

Robb, M.P., Maclagan, M.A, & Chen, Y. (2004). Speaking rates of American and New Zealand varieties of English. Clinical Linguistics & Phonetics:

Rozakis, L. (1995). Power Reading: Increase Your Reading Speed in Just 15 Minutes a Day. New York, NY: Macmillan.

Sereno, S. and Rayner, K. (2003). Measuring word recognition in reading: eye movements and event-related potentials. Trends in Cognitive Sciences, 7 (11).

Smith, B. D. (2005). Breaking Through College Reading, Seventh Edition. New York, NY: Pearson Longman.

SØvik,N., Arntzen, O., & Samuelstuen, M. (2000). Eye-movement parameters and reading speed: A study of oral and silent reading performances of twelve-year old children. Reading and Writing: An Interdisciplinary Journal, 13, 237-255.

Taylor, E.E. (1965). Eye movements while reading: Facts and fallacies. American Educational Research Journal, 2, 187-202.

Tiu, R.D.Jr., Lee, A. T., Lewis, B.A. (2003). The role of IQ in a component model of reading. Journal of Learning Disabilities, 36 (5), 424-436.

Troyka, L.Q. & Thweatt, J. W. (2003). Structured Reading, Sixth Edition. Upper Saddle River, New Jersey: Prentice Hall.

Van Boven, R.W.; Hamilton, R. H.; Kauffman, T.; Keenan, J.P; & Pascual–Leone, A. (2000). Tactile spatial resolution in blind readers. Neurology, 54, 2230-2236.

Wassman, R. & Rinsky, L.R. (2000). Effective Reading in a Changing World, Third Edition. Upper Saddle River, NJ: Prentice Hall.

Whittaker, S.G., & Lovie-Kitchin, J. (1993). Visual requirements for reading. Optometry and Vision Science, 70(1), 54-65.

Questions or comments? Contact the author at

More about the author

Printable Version


Back to top

Home:: Past Articles :: Conferences :: Citation Information :: Feedback :: About the Authors :: Subscription Information :: Submission Guidelines

Site Last Updated October 31, 2007.
Sponsored By AccuTrack and NCLCA