Towards Better Decodables and Better Use of Decodables

When I was becoming a teacher I learned a five-step approach for teaching decoding.

1.         Visual discrimination

2.         Auditory discrimination

3.         Auditory-visual matching

4.         Word reading practice

5.         Text reading practice

Let’s say, you were trying to teach the letter “s” in the initial position. You’d make sure the students could recognize that letter – distinguishing it from other letters and then recognizing words that included it. Then, you’d make certain students could perceive that sound in and out of words, what we now call phonemic awareness. After that we guide students to match up the /s/ sound with the letter s. That simple basic instruction would be followed by lots of decoding practice. These days I’d add encoding practice to that, having the kids trying to write words with the target element.

The last step, text reading practice, was usually addressed by having students reading some random sentences that included words that started with s. Now, we turn the kids loose on “decodable texts,” giving them an extensive amount of concentrated practice in using their new phonics knowledge to read the text.

The same 5-step routine can be used for teaching spelling pattern-pronunciation matches like CVC, CVCe, or CVVC patterns with certain vowels, and so on. But whatever the element or pattern, the sequence is the same – make sure kids can see it, hear it, connect it, and then practice decoding and encoding words, ending up with relevant text reading practice.

This still makes sense to me, but I’m concerned about the quality and appropriateness of the decodable text step. There has been surprisingly little research on the benefits of decodables in teaching reading. It is mentioned so rarely in phonics studies, that phonics meta-analyses don’t give a clue as to its effectiveness.

Direct studies of decodables have, at best, found them to confer only very fleeting benefits (Adams, 2009; Cheatham & Allor, 2012; Jenkins, Peyton, Sanders, & Vadasy, 2004; Mesmer, 2000; Mesmer, 2005). By the end of grade 1, one can’t tell if students worked with decodables or not – they seem to decode and read equally well whether or not the instruction included them (Juel & Roper-Schneider, 1985).

Concerning is the research that suggests that decodable texts can be overdone. This is where “statistical learning” – a concept from language learning – comes in.   

Statistical learning, in this case, refers to the ability to detect patterns and regularities in text (Seidenberg, 2017). Learners unconsciously compute the probabilities of which sounds or pronunciations are likely to go with spelling patterns. This allows readers to prioritize the most useful decoding choices when it comes to reading unknown words. For instance, in CVVC words, the first vowel is long in a large percentage of one syllable English words (e.g., bait, meat, coat).

However, not all CVVC words are created equal. Words with the ai combination are almost always pronounced with a long a, while a much lower percentage of ea have a long e (95% versus 75%).  

Experience with texts allows us to prioritize certain pronunciations. When one confronts a word like bread the first time, the long-e pronunciation might be worth a try, but when that doesn’t work, it helps to sense what the next best alternative pronunciation would be.   

Decodable texts aim to provide students with intensive practice with specific decoding patterns, so they skew the statistical properties of English – increasing the presence of a pattern well beyond what it typically appears in text. This skewing misinforms kids as to the consistency of a spelling-pronunciation pattern, perhaps reducing their decoding flexibility. Studies show long term repercussions from over representing a pattern in early instructional texts – with readers overly favoring one or another pronunciation long after (Venezky & Johnson, 1973).     

This problem isn’t unique to decodable texts. That research that reported unintended negative consequences from over-representing spelling patterns early on was not a study of decodables, but of controlled vocabulary readers. Those texts depend on a small handful of words that are used repeatedly, an approach focused more on memorizing than decoding.

Leading children to believe that all words are simply decodable, that each letter or spelling pattern has a single pronunciation, is disadvantageous. Effective decoders need to be flexibile if they are to come up with the right pronunciation for the ough spelling pattern (there are 6: bough, dough, rough, cough, through, furlough) or if they are to represent the long-a phoneme properly in their spelling (there are 8: cake, rain, play, baby, great, veil, they, eight).  

Decodable texts tend to overdo it. Their intensive coverage of a particular phonics element is both a strength and potential problem. Over emphasis offers students lots of concentrated practice (that’s the strength). But such over-emphasis can mislead kids into anticipating such consistency with actual text (and that’s the problem).

I asked ChatGPT to create some simple decodable texts for the CVCe pattern with the letters O and A. I didn’t provide a lot of guidance and it’s fair to say that the stories are dopey, but they would provide lots of decoding practice with the target spelling patterns. Let’s not worry about the dopiness here. That could be improved upon with little effort but that isn’t the point.

The Rose story is 61 words long and includes 26 opportunities to decode “consonant-O-consonant-silent e” words, and Jake is 79 words, with 23 long-a words of the CVCe variety.

Of the 26 CVCes in the first story, there are 14 different words – Rose is repeated 5 times, home and cone twice each, and mole comes up 6 times. Such repetition can help, though I’d have preferred more home, hope, and stone and less mole – as those words will be more useful in the long run.

In the Jake story, those 23 CVCe instances include 11 different words, so there is slightly more repetition in this one: 7 Jakes, 2 cames, gaves and rakes, and 3 lakes. Again, the repetition of some of those (Jake) are not as worthwhile as some other possibilities – though all provide the intended decoding practice.

Rose and the Mole

Rose rode home on a lone road. A mole poked its nose from a hole. Rose chose to throw a stone at the mole. The mole hid. Rose felt bad. Then the mole left his hole and came close. He gave Rose a cone. She went home and gave her mom the cone. Her mom spoke. “I hope you help mole.”

Jake at the Lake

Jake came to the lake with Dad. "Take this bag," said Dad. Jake made a plan. He put a rake in the bag. They went to a gate by the lake. Jake saw a big wave. He gave Dad the bag. Dad made a pile of wet leaves with the rake. Then Jake and Dad sat by the lake. "It was a great day," said Jake. Dad gave Jake a smile. They came home late.

Revised Rose and the Mole

Rose rode home on a lone road. A mole poked its nose from a hole. Rose chose to throw a stone at the mole. When she was done, the mole hid. Rose felt bad. She said, "Come, mole." Then the mole left his hole and came close. He gave Rose a cone. She went home and gave her mom the cone. Her mom spoke. “Come here. I hope you will help Mole.”

Revised Jake at the Lake

I worked done and come into the first story and have into the second – high frequency words that break the pattern. Including such rule breakers in decodables, keeps kids honest and gives them a better idea what decoding requires.

There used to be some spelling programs that took a similar approach to the over consistency that we see in too many decodables. Their weekly spelling list might include cart, chart, dart, mart, part, smart, start, and tart. Kids liked it because they didn’t have to study for Friday’s test – learn one word and you can get them all right. Research was not supportive of the approach.

The problem is that such consistency encourages students to not think much about the words. The benefits of phonics, spelling, and morphology instruction is that they tend to increase kids’ word awareness.

Think of the wonderful book, Words their Way (Bear, Invernizzi, Templeton, & Johnston, 2020), and its principles of word study. Two of the 11 principles are relevant here: “Compare words that ‘do’ with words that ‘don’t’” and “Don’t hide exceptions.’”

Those authors and morphology instruction experts (Peter Bowers, for instance) would go even further than just problematizing decodables in the way that I did. They’d bring kids’ attention to the discrepancies that I introduced and would add some explanation (e.g., English words can’t end with v, done is a variant of do).

In any event, beginning readers should get decoding practice. But this practice need not be so consistent that it misleads them about how to decode. Decodables should do more than provide intensive concentrated practice, they should introduce inconsistencies that will require choice rather than rote responses.   

References

Adams, M. J. (2009). Decodable text: Why, when, and how? In E. H. Hiebert & M. Sailors (Eds.), Finding the right texts: What works for beginning and struggling readers (pp. 23-46). New York: Guilford Press.

Bear, D. R., Invernizzi, M., Templeton, S., & Johnston, F. (2020). Words their way (7th ed.). New York: Pearson.

Cheatham, J., & Allor, J. H. (2012). The influence of decodability in early reading text on reading achievement: A review of the evidence. Reading and Writing: An Interdisciplinary Journal, 25(9), 2223-2246.  https://doi.org/10.1007/s11145-011-9355-2

Jenkins, J. R., Peton, J. A., Sanders, E. A., & Vadasy, P. F. (2004). Effects of reading decodable texts in supplemental first-grade tutoring. Scientific Studies of Reading, 8(1), 53-85. https://doi.org/10.1207/s1532799xssr0801_4

Juel, C., & Roper-Schneider, D. (1985). The influence of basal readers on first grade reading. Reading Research Quarterly, 20(2), 134-152.

Mesmer, H. A. E. (2000). Decodable text: A review of what we know. Reading Research and Instruction, 40(2), 121-142.  https://doi.org/10.1080/19388070109558338

Mesmer, H. A. E. (2005). Text decodability and the first-grade reader. Reading & Writing Quarterly, 21(1), 61-85. https://doi.org/10.1080/10573560590523667

Seidenberg, M. (2017). Language at the speed of sight. New York: Basic Books.

Venezky, R. L., & Johnson, D. (1973). Development of two letter-sound patterns in grades one through three. Journal of Educational Psychology, 64(1), 109–115. https://doi.org/10.1037/h0033971

 

Shanahan On Literacy Podcast