Making the grade

This month’s guest post is from Marcos Benevides, who has some suggestions on how to hit the right note with your graded reader.

The MaWSIG blog features guest posts by members – please get in touch if you would like to write for us.

Marcos 2015Making the grade: Approaches to writing graded readers

by Marcos Benevides




There are two fundamental ways to approach the writing of graded readers, which I refer to as systematic and holistic. They are not mutually exclusive, and each has advantages and disadvantages. It is good to consider both as theoretical ends of a continuum upon which any actual writing takes place.

Systematic approach

A systematic approach presumes that graded readers are primarily tools for language learning, and therefore should above all be ‘graded’ (i.e. levelled) as objectively as possible. The idea is that an objective levelling system can help teachers and learners to progress most efficiently. This is the implicit promise of the headword count which is featured on most series covers.

As with lexis, this basic principle of objective levelling is also applied to grammar, meaning that some sort of easy-to-difficult and/or frequent-to-less-frequent grammar syllabus is also used to level graded readers. At easier levels, for example, graded readers may avoid perfect tenses and the passive voice, and may restrict sentence length.

The author is aware of these specific restrictions from the start, and must then try to tell a story using only the tools available. It is a sort of Lego® approach to writing, in the sense that one is necessarily constrained by the limited set of available ‘bricks’ (i.e. the list of allowed lexis and grammar). This approach is restrictive, but can still be creative and fulfilling, as anyone who has ever played with Lego will know.

A systematic approach to writing graded readers is also sometimes described as similar to writing within a restrictive format such as haiku or sonnet. I agree, but with caution – just as it is relatively easy to follow the rules and write a passable sonnet, it is supremely difficult to write a good one.

Furthermore, objective readability scores are an illusion. Language is not in fact made up of neat blocks that fit together perfectly, not even at the lexical level. Words, so tantalisingly countable on the surface, are actually far trickier to account for than they appear. As readers, we know that the difficulty of a text depends on many factors beyond merely linguistic ones; plot, genre, setting, number of characters – even reader interest and motivation – all have a role to play.

Holistic approach

The approach I prefer starts from the whole rather than from a set of parts, from the story rather than from a word list or a grammar syllabus. It is understood that the ultimate purpose of graded readers is to fit into difficulty levels, of course, but it is also recognised that overall readability will include complex factors arising from content and form, as well as from language.

If a systematic approach is like building a story out of bricks, a holistic approach is more like sculpting it from stone. Rather than build a story up from available pieces, the author first writes a complete narrative and then begins a process of chipping away and polishing down until the imagined shape emerges with the correct proportions (i.e. at the appropriate level).

Following a holistic approach means that the demands of the story will suggest which parts, words, and forms must be kept, and which can be cut away. This will generally follow the shape of a word list or a grammar syllabus (these are, after all, based on frequency) but it will also allow for a greater degree of flexibility in weighing items on their own merits and in their proper contexts. For example, in a story about time travel, can the past perfect be included if it is made clear by the events in the story?

A holistic process makes use of computers and corpus-derived word lists, but it does so toward the end of the process rather than at the beginning. Once a complete draft of the story has been written, tools such as VocabProfile, Concordance, and Range (– along with a lot of intuitive work by the editor – can be used to measure and reduce lexical load.

This approach can be more labour intensive than a systematic one, and it can sometimes lead to a less precise finished product in terms of objective levelling criteria. However, it can also allow for work which is more consistently impactful, and for stories which can be more engaging and interesting beyond their utility for language learning. Stories which may even, dare we say it, possess some quality of ‘literariness’?

On readability factors

One danger of a systematic approach is that it is almost too easy to list lexical and grammatical items, and comparatively more difficult to list other factors. It means that too often only lexis and grammar are counted. Other factors, however, can still compound to make otherwise easy texts more difficult.

For example, formatting and style. I am sometimes shocked at how haphazardly dialogue is dealt with in some series. This can be quite confusing for the developing reader. In my easier books, I try to be consistent, almost always using the pattern:

‘Hello,’ the girl says. ‘Do you know who I am?’

In other words, the opening of the turn is always quickly followed by the identification of the character and a full stop. This is followed by any additional dialogue, and then a paragraph break for the next speaker’s turn. This consistency can be helpful to low-level readers.

Another example of consistency is when transcribing onomatopoeic sound effects. I always use capital letters and italics: BANG! FWOOSH! CRASH! Doing so clearly sets off these off-list lexical items, and suggests a stylistic rule for the reader; something like: Read all-capital words in italics as sound effects. I like to think that such meta-linguistic ‘hooks’ can help readers to develop confidence, even if they sometimes do break the rules of lexical frequency.

We haven’t touched on readability factors at the story level, but these are important as well, and include things such as genre conventions, number and types of characters, setting, voice, and plot linearity.

Consider the image below.


I like to think that headwords for graded readers are like a volume setting on a music player. We can turn the dial and make a piece ‘better’, but only by one relatively blunt measure. Yet, like pieces of music, stories are complex things. If we want to make them maximally beautiful at each level without ‘distortion’, then we must consider all appropriate factors.


Marcos Benevides is a teacher, author and editor based in Tokyo. His coursebooks and graded readers have received multiple awards over the years, including the Duke of Edinburgh English Book Award, an ELTon, and several Language Learner Literature medals. He is currently the publisher and series editor of Atama-ii Books, a series of multiple-path graded readers.

8 thoughts on “Making the grade

  1. Personally, I have always found that the holistic approach works best. The danger of the systematic approach is that the book ends up sounding like ‘EFL-ese’.
    A further question is whether to include exercises at the end of the book or not. I feel that doing so is a mistake. When was the last time you read a novel and then sat at home answering comprehension questions about it on the last page? Graded readers serve the same role as any other type of reading in education: they engage people’s interest in the subject that they are studying. Reading one should never be seen as a chore.

    1. Thank you, Alastair. I agree with your point about ‘EFL-ese’ – and not only with regards to the language. There are cultural and content ‘bricks’ which authors are often not allowed to use, i.e. the infamous PARSNIPS list, which lead to a kind of EFL-ization of content as well. These factors undermine one of the purposes behind graded readers in the first place: to provide easy yet *not infantile* stories for language learners. Otherwise, why not simply use already available books for young children?

      As for exercises in GRs, I don’t mind them too much, as long they’re at the back, out of the way, and don’t detract from the story (but these days, why not have links to downloadables instead?). Series such as Pearson Active Readers and OUP Dominos are fine and can be quite well done, but I see them more as mini-textbooks than as graded readers, personally. They don’t work as well in extensive reading, which is my focus when I’m wearing my teacher’s hat.

  2. Great to have an insight into writing a graded reader, Marcos. I have never written one, but I think I can see some parallels in the processes involved in writing interesting and authentic feeling texts for course books. And, for that, I would definitely subscribe to the holistic approach. I try and start with the content and the genre, and as much as possible write it as if I were writing that kind of genre for native speakers, and then simplify it down as needed.

    1. Hi Rachael. Yes, there are definitely parallels to writing coursebook materials – even at the structural level, if the course is something more themed and/or task-based than the norm. If one starts from an overall themed structure, it becomes relatively easy to them select appropriate tasks that fit the context very holistically, rather than starting with a pile of ‘bricks’ such as, “must include a letter-writing task; a presentation task; a short survey task; etc.”

  3. I really enjoyed your distillation of the process, Marcos. The readers I have written (with Alastair!) started as ideas and were written as proper stories, then had the level-filter applied retrospectively. Taking the systematic approach knocks all the creativity out before you even get going. And Rachael is right of course that the same principle applies to texts in coursebooks. It’s no coincidence that my best coursebook texts were original pieces that I adapted lightly or paraphrased heavily, keeping the originality and interest of the genres and afterwards grading to meet specific requirements. I love your Readability Factors illustration.
    You haven’t talked about the role of the choose-your-own-adventure format in holding students’ interest, engendering a real sense of reading for fun, as well as keeping them coming back to re-read along alternate paths – perhaps this is a future blog post?

    1. Thanks, James – and yes, the benefits of the multiple-path format for language learning are interesting to consider. One that I suspect has a strong effect is coming to a choice point and having to predict what might come next in the story. Surely this must help comprehension. And, of course, needing to take stock of previous events in order to make a good choice. That’s not even to go into all the motivational advantages, as you know.

      Actually, I’d love to hear yours and Alastair’s take on that n a future post! 😉

  4. My first reader was written using both in a way as I’d written half of it as a draft a couple of levels higher than it was eventually commissioned at so adapted it down when I got the word and grammar list (and it was about time travel kind of, and I couldn’t use the past perfect!). So I then finished the book using the more systematic approach but I didn’t find it that limiting as it was like getting into a certain “voice” and then just writing in that voice. Because I’d rewritten the first half I more or less knew what I could and couldn’t use and had internalised the limitations. It is a bit like if you were writing in child voice or an accent/dialect, that’s not necessarily limiting and there are amazing books written in those voices.
    But if you’re literally sitting there with a list of grammar to make sure to include that won’t lead anywhere good as you say. I attempted a sample reader for an ad calling for choose your own adventure where the choices were based on language points and couldn’t do it well enough to bother sending in. I’d be surprised if that project got off the ground.
    On the topic of dialogue, I was once suggested by an editor to indicate thoughts using <> which I politely declined to do since that’s not real English punctuation!

    1. Nicola, good point about finding the right voice and staying there. Actually, I don’t think there’s anything wrong per se with being systematic – the danger, I think, comes from such an approach falling too easily into the formulaic trap. As we all know, ELT writing is a business, and unfortunately graded readers don’t come from a tradition of ‘Authorial Literature’ (capitals emphasized). So it’s tempting for publishers to commission stories simply written by numbers, and for us to churn them out that way for the paycheque. But sure, any approach can be fruitful if the writer is going to be thoughtful about creating interesting stories.

Leave a Reply to Rachael Roberts Cancel reply

Your email address will not be published.