When Misspellings are Good

Not all errors mean the same thing

I think I’m supposed to want my class’s spelling to get worse.

Recently, I was reading third graders’s sentences. They wrote good sentences, but words were being misspelled. And they’re strong readers, and stronger since the year began, going from 55% of the class proficient in the accuracy component of our oral reading fluency probe to 95%, with great gains in rate as well.

And we work on spelling, daily. Their spelling of dictated words is strong. During spelling, I circulate and check and correct as we go, modifying the planned list of words to adjust to things students have done. Spelling is explicit and systematic, with irregular words mixed in. We pronounce the words correctly and then say them using our “spelling voice,” as if they were fully regular. (Danielle Colenbrander has done some valuable research in this area.)

In their own writing however, I keep noticing errors that surprised me. They’re not disastrous errors, but the words are misspelled. I correct them as I see them. So we fix it, and one student's pharoha becomes a pharaoh and another's cacoon becomes a cocoon and a third’s confedrisy becomes confederacy. None of these are particularly surprising misspellings for third graders, but I notice more spelling errors, even among the strongest writers in the class, than I did at the beginning of the year.

I have two possible explanations for this. Both are likely to be true to some extent, and both are good (I think).

First, and probably less importantly, students might be focusing less on spelling accuracy when writing. That’s not inherently good, and I certainly hope I haven’t inadvertently encouraged careless writers. Rather, I’d like to think that students are focused more on sentence structure (an area we have done a lot of work in) and using writing to express ideas and thinking about content. Again, still not inherently good. But, I would argue, this focus is good in the context of a high degree of confidence in their improvements in decoding, knowledge of phonics and morphology, and reading fluency. Their sentences are better and about more complex ideas than at the beginning of the year. Maybe they took their gains in encoding and, instead of using that to improve their spelling, used it to reallocate working memory to other aspects of writing. When the first seatbelt regulations were passed in the 1960s, car crash fatalities stayed constant. People just drove faster (less time spent in the car) with the same level of safety. There were benefits, just in other areas. Teachers might improve their students’s facility with a skill, only to see the gains appear in some other part of that domain. Maybe. It’s hard to get evidence of what they are thinking about while writing.

Fortunately, the second possible explanation holds more promise. The kids are just misspelling different words than at the beginning of the year! Writing involves students self-selecting words they want to use in their sentences. They weren’t writing about pharaohs or the life cycle of butterflies in September. Spelling accuracy, as observed in writing, is not an independent variable. It depends on the author’s vocabulary. In the first place, oral vocabulary lexicons ought to be larger than automatic sight word lexicons. The efficiency of orthographic mapping for word acquisition depends on this relationship: you can’t map the orthography of a word you don’t know orally. Furthermore, even as grown-ups, we have plenty of words we can read automatically which are still tricky to spell— guarantee and restaurant happen to be two of mine. And we have words we need to write out and look at before we are confident they are spelled correctly— panicking or picnicking being personal examples. The set of words we can automatically identify as readers and the set of words we can automatically spell as writers are not identical.

So, students are using challenging words whose meaning they know but which they can’t spell perfectly. If those were the same words as when the year began, that wouldn’t be good. But they’re not the same words. They’re newly acquired words. And so I think these new errors are a sign not of weak spelling but of an expanding vocabulary. They couldn’t make those mistakes in September because they didn’t know those words yet.

There is a separate point here to be made about writing samples being a weak measure of spelling ability because the student gets to choose which words they use. Prompt quality matters, but there is also an inherent relationship between vocabulary strength and spelling accuracy. Weak encoders will produce more accurate samples if they also have weak oral vocabularies. Students with expansive oral vocabularies may produce more errors. That’s not necessarily evidence of an encoding weakness, it could just be evidence of a relatively strong vocabulary.

And, if we’re confident about students encoding and decoding from other evidence, we should be happy when their writing features some good, strong, misspelled words.