The PIRLS Reading Result–Better than You May Realize

by Dan Willingham
December 17th, 2012

This was written by cognitive scientist Daniel Willingham, professor of psychology at the University of Virginia and author of  “When Can You Trust The Experts? How to tell good science from bad in education.” This appeared on his Science and Education blog.

The PIRLS results are better than you may realize.

Last week, the results of the 2011 Progress in International Reading Literacy Study (PIRLS) were published. This test compared reading ability in 4th grade children.

U.S. fourth-graders ranked 6th among 45 participating countries. Even better, US kids scored significantly better than the last time the test was administered in 2006.

There’s a small but decisive factor that is often forgotten in these discussions: differences in orthography across languages.

Lots of factors go into learning to read. The most obvious is learning to decode–learning the relationship between letters and (in most languages) sounds. Decode is an apt term. The correspondence of letters and sound is a code that must be cracked.

In some languages the correspondence is relatively straightforward, meaning that a given letter or combination of letters reliably corresponds to a given sound. Such languages are said to have a shallow orthography. Examples include Finnish, Italian, and Spanish.

In other languages, the correspondence is less consistent. English is one such language. Consider the letter sequence “ough.” How should that be pronounced? It depends on whether it’s part of the word “cough,” “through,” “although,” or “plough.” In these languages, there are more multi-letter sound units, more context-dependent rules and more out and out quirks.

Another factor is syllabic structure. Syllables in languages with simple structures typically (or exclusively) have the form CV (i.e., a consonant, then a vowel as in “ba”) or VC (as in “ab.”) Slightly more complex forms include CVC (“bat”) and CCV (“pla”). As the number of permissible combinations of vowels and consonants that may form a single syllable increases, so does the complexity. In English, it’s not uncommon to see forms like CCCVCC (.e.g., “splint.”)

Here’s a figure (Seymour et al., 2003) showing the relative orthographic depth of 13 languages, as well as the complexity of their syllabic structure.

From Seymour, et. al. (2003)

Orthographic depth correlates with incidence of dyslexia (e.g., Wolf et al, 1994) and with word and nonword reading in typically developing children (Seymour et al. 2003). Syllabic complexity correlates with word decoding (Seymour et al, 2003).

This highlights two points, in my mind.

First, when people trumpet the fact that Finland doesn’t begin reading instruction until age 7 we should bear in mind that the task confronting Finnish children is easier than that confronting English-speaking children. The late start might be just fine for Finnish children; it’s not obvious it would work well for English-speakers.

Of course, a shallow orthography doesn’t guarantee excellent reading performance, at least as measured by the PIRLS. Children in Greece, Italy, and Spain had mediocre scores, on average. Good instruction is obviously still important.

But good instruction is more difficult in languages with deep orthography, and that’s the second point. The conclusion from the PIRLS should not just be “Early elementary teachers in the US are doing a good job with reading.” It should be “Early elementary teachers in the US are doing a good job with reading despite teaching reading in a language that is difficult to learn.”


Seymour, P. H. K., Aro, M., & Erskine, J. M. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143-174.

Wolf, M., Pfeil, C., Lotz, R., & Biddle, K. (1994). Towarsd a more universal understanding of the developmental dyslexias: The contribution of orthographic factors. In Berninger, V. W. (Ed), The varieties of orthographic knowledge, 1: Theoretical and developmental issues.Neuropsychology and cognition, Vol. 8., (pp. 137-171). New York, NY, US: Kluwer


  1. It often seems like Dr. Dan makes such a great living by merely stating the obvious, but we shouldn’t be fooled by his approachable writing style – highly unusual for an academic, and very welcome to a wider public audience. As the CK community knows, his seemingly common sense conclusions are based on solid research and analysis.

    His statement that Finnish kids have it easier than English-speaking kids makes perfect sense based on the chart shown above. I studied Spanish for many years and reached a fairly high level of proficiency (lost over the last 30 years), and I was always struck by how much easier Spanish phonetics are than English phonetics.

    Once you master the basic Spanish sounds, you can correctly pronounce almost any Spanish word, even if you’ve never seen it or don’t understand its meaning in context. I’d guess at most 2% of Spanish words have “irregular” pronunciation or spelling, with the English percentage many times greater.

    My son’s 9th grade Humanities class is a combination of English and history topics; a small part of the work every week is a spelling list of 20 words, handed out on Mondays, that the kids are tested on every Friday. Almost all the words are tougher, irregular words that trip up even many good spellers. Here are some words for this week: shepherd, siege, sophomore, souvenir, subtle.

    All irregular words, that simply have to be memorized because no fixed English phonetic rule completely governs their pronunciation or spelling.

    No wonder Finnish kids have a big head start learning the reading aspect of their native language. By age 7, they’ve orally mastered a good-sized vocabulary, that is much more easily transferred to reading and writing than English words are.

    Comment by John Webster — December 18, 2012 @ 11:25 pm

  2. This is the news you never see on the front page. In fact I’ll bet few teachers are aware of it. You can bet if students had scored poorly it would make the news and teachers would be blamed.

    Comment by Mary S. — December 19, 2012 @ 2:41 am

  3. Dan,

    Not intended as a criticism but – this entry is somewhat of a jargon-laced, edu-babbled analysis.

    Or could we simply say, our education reform efforts appear to be working?

    Comment by Paul Hoss — December 19, 2012 @ 7:55 am

  4. I don’t think this is obvious for most people. Yet, we can still do a better job. Diane McGuinness has written about this extensively. The U.S. needs more effective synthetic phonics programs in classrooms. Lippincott Basic Reading did a nice job of teaching the letter sequence “ough.” Doubt many beginning reading programs in the U.S. touch on it today…

    Comment by Kev — December 19, 2012 @ 9:39 am

  5. Outside of decoding is there an advantage to the complexity of English? My father one time told me that English was the language of science because it was easier to expand than other languages. Recognizing that at least in terms of the life sciences, Latin is more accurate characterization, and recognizing that could just be a chauvinistic statement, is there an advantage to this complexity?

    Comment by DC Parent — December 19, 2012 @ 4:10 pm

  6. ” My father one time told me that English was the
    language of science because it was easier to expand
    than other languages.”

    In other languages it’s harder to make foreign words “fit”.

    In Polish, for example, women’s names all end in ‘a’ and decline predictably. So what do you do about an Arabic masculine name like ‘Ali’ since in Polish it means ‘Something that belongs to the woman named “Ala”‘?

    ‘Helsinki’ is a plural, and treated as such. You don’t visit ‘it’, but ‘them’. Same for the entire country of Hungary.

    English is far simpler in this regard.

    Comment by Educationally Incorrect — December 19, 2012 @ 10:21 pm

  7. Dr. Willingham’s work is excellent and I recommend his book highly on my own website, which is devoted almost entirely to helping struggling readers.

    Over the past decade and a half, I’ve come to several conclusions about reading instruction in the U.S., most of them consistent with what Dr. Willingham writes in this article.

    English is confusing, as he points out, in both orthographic depth and syllabic structure. To address the orthography issue, U.S. educators need to decide upon a single coding structure that should be taught to all children. That is, we need to decide what is, and is not, a phonogram.

    It makes no sense for three different teachers, sometimes in adjacent classrooms in the same building, to teach the “igh” in words like “right” and “light” differently. Yet one will teach the “gh” as silent, a second will teach “igh” as a reliable phonogram for the /ie/ sound (long-i), and a third might just call it a “sight word” that requires outright memorization.

    To that end, if one searches the phrase “ontrack phonogram set” you’ll presently find at least 10 articles on my website that address this issue and propose an 84 item teaching set of phonograms that I personally taught nearly 200 struggling readers in one-on-one lessons. Whether my set should be adopted, or another set, that needs to be done if U.S. students are to ever gain consistent instruction in the orthography of English.

    As for syllable structure, one of the greatest curses ever foisted on young readers was Webster’s original syllable structure. Today, nearly every multi-syllable reading program focuses on the structure he proposed. Thus we have the mysterious six-syllable types that most literate adults couldn’t recite if their lives depended upon it. Yet that’s what we base multi-syllable instruction on. Oh, that, and looking for words within the word, and learning to sight-read some prefixes and suffixes. And then we wonder why kids just guess at the longer words. As for consistency across schools, or even across classrooms, forget it. It isn’t happening.

    In my 15 years of working with this, I developed a far easier chunking approach to multi-syllable words, an approach that has a child marching, left to right, through any long unfamiliar word, following a simple rule (stop each chunk after the vowel sound) and then applying three simple exceptions that even an 8 year old could easily learn. The result, in my experience, was that I got to watch struggling readers abandon guessing as their main strategy and take up a strategy that they realized was working for them. Searching “ontrack multisyllable” presently puts the article as the top result.

    I’m convinced that if we implemented something along the lines of the OnTrack Reading Phonogram Set in first grade, we could teach all 84 phonograms during that year, and if we started teaching the OnTrack Reading Multisyllable Method in the second semester of first grade, continuing it into second grade, reading ability in the U.S. would virtually skyrocket. Whether it be the system I developed and used with struggling readers, or something quite similar, it’s what we need.

    One final point. There are probably nearly 200 potential phonograms used to construct English words. Students should also be taught a method by which to extend the original teaching set of phonograms (84 in my set) as they encounter new, less frequently used, or easily explained, phonograms in words they encounter as their reading level increases. That issue, too, is addressed by the coding system that I use. The coding system, depending entirely upon single-underlining digraphs, numbering 2nd, 3rd, and 4th options for phonograms, and double-underlining the unusual, enables a child to learn additional phonograms on a discovery basis, deciding upon them as he goes along in his reading experiences.

    Rod Everson
    OnTrack Reading

    Comment by Rod Everson — March 19, 2013 @ 10:48 am

  8. I tossed in the “final point” in my previous comment, and I’d like to elaborate a bit in this one.

    Phonograms like the “bb” in “hobby”, and “dd,” “ff,” “tt,” “ss,” “zz,” etc., are not included in the 84-item teaching set. All that is required is to underline the “ff” in “stuff” and the “ss” in “grass” as a digraph and explain that a doubled consonant usually represents the same sound as they’ve already learned for the single letter phonogram. In other words, it’s easily explained and need not be part of the set.

    The set includes the four “split vowel” phonograms, ee, ie, oe, and ue, that by themselves (bee, pie, toe, cue) represent the long vowel sound, but also can be “split” resulting in words like here, pile, tone, and cute. (the infamous silent-e, magic-e, bossy-e situation.) If the unsplit phonogram is taught first, it is simple to explain to a child that the split vowel digraph represents the same sound, and this can be done at a level of logic that he easily understands.

    Once two or three of those are explained, then words using the split a-e, like tame and gate, are easily explained by extension of the concept, even though “ae” is not in the 84-item set.

    And finally, the double-underline is the tool by which the 84-item set is extended, often by the student himself. The first time he sees a word like psoriasis, he will chunk it psor-i-a-sis and wonder about the “ps.” Is the “p” an extraneous letter, one that he just has to remember (like the “w” in two, for example), or will he find other words someday in which the “ps” is a digraph for the /s/ sound? It doesn’t matter. For now, he double-underlines the “p” as an extraneous letter that he has to memorize if he’s to read “psoriasis” and he certainly needs to memorize it for spelling, for there’s no phonetic clue that it should be there at all.

    But he’ll have primed his brain to be on the lookout for a phonogram “ps” = /s/ in the process. When he discovers the word “psych” or “psychic” later, he might again just double-underline the “p”, but now he’ll be wondering, is “ps” a digraph for the /s/ sound? Eventually, he’ll find another, and another, and decide that it is.

    But should we teach it, especially early on? I don’t believe we should, for an obvious reason. The “ps” combination usually isn’t a digraph in most of the early reading he’ll do. Instead, he’ll be reading “hopscotch” (or just “hops” for that matter) and “chopsticks” and “lopsided” and hundreds of other words where the “ps” together in a word represent two separate sounds. The last thing we want is to add a level of complexity to the decoding process before it is likely to be needed. It is better to let the advanced reader “discover” the more complex digraphs as his reading level progresses, for he will both need to know it then, and will be more likely to remember that which he has discovered on his own in a questioning mode.

    Thus, the double-underline might stay a double-underline, or might someday become just a single-underline, indicating that the child has finally decided that the construct he’s been treating as unusual actually falls into the phonogram logic as a digraph for a particular sound, or sounds.

    The original phonogram set can even be extended. For example, the phonogram “ai” is taught as simply the /ae/ sound in “train.” A young child will soon encounter “said” and double underline the “ai” and place an “e” under it indicating the /e/ sound. Eventually, the child, depending upon his instructor and his own biases, might decide that words like “again,” “against,” “captain,” and “mountain” warrant his treating the phonogram “ai” as also representing a second sound, the /e/ sound. This is fine, although, for spelling purposes I always encouraged them to think of those words as having long-a sounds even though we pronounce them with short-e sounds.

    Eventually, though, he’ll encounter the word “plaid” and wonder again whether the phonogram “ai” has an /a/ sound as well. Though his brain will now be primed for that eventuality as well, he’s not likely to ever run across another word like it, and will someday decide that he’s just got to remember that unusual spelling of the /a/ sound.

    My point is that it is possible to provide a young child with simple tools by which, as he matures, he will be able to add to the 84-item teaching set of phonograms (or however many we decide to teach) as his reading level gradually increases.

    Some will no doubt realize that Romalda Spalding already did this with her 70-item set. My problem with her work (which was excellent) is that a child needs to learn several rules because she didn’t extend her structure to most words ending in the letter “e” and because she stuck with Webster’s syllable structure, thereby necessitating the splitting of doubled consonants into two sounds in most words, although they are almost always phonograms for a single sound. My 84-item set is larger, mainly to allow for endings like ce, ge, ze, se, le, etc., most of which are as easy to learn as bb, dd, ff, ss, etc., since they represent the same sound(s) as the consonant they contain.

    Comment by Rod Everson — March 19, 2013 @ 11:35 am

RSS feed for comments on this post. TrackBack URL

Leave a comment

While the Core Knowledge Foundation wants to hear from readers of this blog, it reserves the right to not post comments online and to edit them for content and appropriateness.