It is a common assumption that language proficiency translates to reading fluency, the ability to decode and comprehend text at the same time. However, research suggests that this may be untrue. Therefore, it is important to develop reading fluency alongside language proficiency. One reason this is important is that, by improving one’s ability to decode (that is, recognize and understand familiar words) leaves more cognitive resources for higher-order comprehension skills. This idea, the basis for repeated reading (RR), is rooted in Automaticity Theory. Gorsuch and Taguchi’s (2008) research on Repeated Reading in a EFL context shows that it leads to increased comprehension.

RR has been a successful tool in L1 reading contexts, especially when RR is assisted with audio recordings. However, it is little used in L2 language instruction. Prior research on RR in a language teaching context suggests improved reading rates and comprehension, though not always significant, and not always transferable. A lack of research and somewhat flawed research designs indicates there is not enough research to make a clear conclusion. Gorsuch and Taguchi’s (2008) study sought to add to the research with an updated research design.

The Study

Gorsuch and Taguchi conducted an 11-week quasi-experimental study with 24 Vietnamese university students in Vietnam, divided into experimental groups and control groups. The control groups received no special instruction. However, the experimental group received the RR treatment twice a week based on dividing Penguin Readers into 16 texts of about 500 words each. These texts were considered the pre-intermediate level, part of the 1200-headword series level.

Students would first read the passage and time themselves, recording the time. They would read two more times, this time assisted with an audio version of the text as well. Then, they would read two more times silently and also self-timed. At the end, they would complete a report. All together, they read each text 5 times.

To measure whether this was effective, the researchers used a short answer and recall pre- and post-test that followed a similar model of five readings of the same text. It is interesting to note that the experimental group had a significantly lower reading ability than the control group.

The Results

One aspect of RR the researchers looked at was reading fluency, measured by words per minute (wpm):

• At the end of the 11 week treatment, reading fluency increased by about 55 wpm from the first reading in the first session to the first reading of the last session.
• Similarly, looking at the fifth reading of each session, reading fluency increased by about 91 wpm (significant).
• Reading fluency decreases as new stories are begun, likely due to lack of knowledge of the story.
• Within a single session, fluency by about 105 wpm.
• On the post-tests, the experimental group did not read any faster than the control group, with the control group actually reading faster by the fifth reading, though this was not statistically significant.
  • The authors noted this is likely due to careful reading to answer questions on the test, employing various metacognitive strategies after becoming aware of questions during the first reading.
• The conclusion was that fluency did not transfer to the post-tests.
  • Interestingly, there were no analyses of the comparing pre-test and post-test wpm rates for each group.

Another aspect of RR was comprehension:

• The experimental group outperformed the control group on the post-tests
• This was true for both the first reading of the post-test and the fifth reading of the post-test
  • Recall that the experimental group initially started at a lower level of reading ability
• The scores for the fifth vs first readings are always better for both groups

Discussion

The authors felt that the experimental groups’ improved comprehension can be explained by Automaticity Theory (p. 266-267):

As readers increase their word recognition skills, more attentional resources are made available for higher order comprehension processes…Clearly, participants were processing text faster by the time they took the post-tests.

They found that being able to recognize words automatically improved fluency and comprehension, and left students more cognitive resources to use for metacognitive and various other reading strategies – hence, the lower reading speeds during the fifth readings.

Implications and Suggestions

The authors feel RR has “a robust utility.” They stated that it takes little time (35 minutes, twice a week) and there are potential benefits without drawbacks. After reading this study, I wanted to imagine how one would use RR outside of an intensive experimental context. Here are some ideas:

• Couple RR with extensive reading, using graded readers that are suitable for both your students and the overalapping goals of RR and ER
  • You can use a class set of graded readers and work chapter-by-chapter or break them into 500-word segments
• If you are worried about taking up class time, ask students to do repeated readings at home, explaining the potential benefits of the technique. Make it a habit of the students to do repeated readings.
• If an audio version does not exist, record one yourself – or ask a loved one! A few minutes is all it takes.
• If you are working on intensive reading, where students have done vocabulary building, discussion, and other comprehension activities, give students clean copies of the text and new or old questions. Have them read again for fluency and comprehension. Repeat this several times over the course of a unit / term / semester.
• If you think students will be bored and just skimming by a third or fifth reading, the researchers consider this a good thing because students may be noticing things they had missed during close reading (this was supported by student comments and a reference to another study).

Reference

Gorsuch, G., & Taguchi, E. (2008). Repeated reading for developing reading fluency and reading comprehension: The case of EFL learners in Vietnam. System, 36(2), 253-278.