How to Teach Spelling: Why Weekly Lists Fail (and What Works)

Here is a scene every homeschool parent recognizes. Your child studies ten words all week, aces the Friday test — 10 out of 10 — and then, in Monday's journal, spells half of them wrong. The word they wrote perfectly three days ago is suddenly unrecognizable. It feels like the studying evaporated over the weekend, because in a sense it did. This is the central problem with the way most of us were taught to teach spelling, and it's fixable once you understand what's actually going on.

The usual diagnosis is "she just needs more practice" or "he's a bad speller." Both are almost always wrong. The real issue is the format: an arbitrary list of words, memorized for a test, and then never reinforced. That format fights against how spelling is actually learned. To be clear up front, this isn't an argument that word lists are evil or that you should throw out your curriculum. It's an argument about what goes on the list and what you do with it.

Let's start with why the Friday-test version fails, and then get concrete about what to do on Monday morning instead.

Spelling is a code skill, not a memory skill

The reason memorize-and-test spelling doesn't stick is that spelling isn't fundamentally a feat of memory — it's a feat of decoding the code in reverse. When a child reads, they translate letters into sounds. When they spell, they translate sounds back into letters. Reading researcher Linnea Ehri calls the underlying process orthographic mapping: the brain bonds a word's sounds, letters, and meaning together into a single durable connection. Words that are mapped this way are remembered almost effortlessly. Words that are merely rehearsed for a test are held in short-term memory and then released — which is exactly what you're seeing on Monday.

This is the same engine that drives reading, which is why a structured approach to teaching phonics and a structured approach to spelling are really two sides of one coin. The sound-spelling patterns a child learns to read are the same ones they need to spell. Teach them once, well, and they pay off in both directions.

One nuance worth knowing, because it relieves a lot of parental worry: spelling a word is genuinely harder than reading it. Reading only asks a child to recognize a word — the letters are right there in front of them. Spelling asks them to produce every letter from memory, in order, with nothing to look at. Recognition almost always develops before production. So a child who reads fluently but spells shakily isn't broken or lazy; they're at a completely normal stage. The gap closes as the underlying patterns get mapped.

English is more predictable than it looks

Parents often assume English spelling is just chaos that has to be brute-forced word by word. It isn't. The foundational study here is Hanna and colleagues' 1966 computer analysis for the U.S. Office of Education, and it's still the number people quote sixty years later. They found that about half of English words can be spelled correctly from sound-letter rules alone — which is the part that gets quoted to argue English is hopeless.

But that's only half the finding. When you add in the next layers — letter-position rules, common patterns, word origins (etymology), and word parts (morphology) — predictability climbs to roughly 84 to 87 percent. In other words, the vast majority of English spelling is rule-governed once you know which rules to look at. English isn't 50% phonetic and 50% random; it's a layered system. That single fact is the case for teaching spelling by pattern instead of by arbitrary list, because patterns generalize to thousands of words a child has never studied.

What actually builds spellers, by stage

Structured-literacy programs — the approach backed by the research on how children's brains learn print and codified in the International Dyslexia Association's standards for teachers — sequence spelling the same way good phonics is sequenced: from simple, reliable sound-spellings toward more complex patterns and word parts. The exact words change with your child's stage, but the method is consistent. Here's how it looks across the K–3 years.

Three techniques do most of the work, and none of them requires special materials.

1. Segment, then spell (K–1). Before a young child writes a word, have them say it slowly and tap or count the sounds: /s/ /u/ /n/, three sounds, three letters. This is the single most important habit you can build, because it teaches the child to spell from the sounds rather than guessing at a visual shape. A few minutes a day of segment-and-spell with short-vowel words builds the foundation everything else sits on.

2. Word sorts (Grade 1 and up). Instead of a mixed list, give your child a small pile of words built on two or three patterns and have them sort the words into columns by what they hear and see. For example: sort cake, ship, bike, chin, lake, that into a silent-e column and a digraph column. The sorting is the learning — the child is actively noticing the pattern, not just copying. A set of phonics worksheets grouped by sound pattern gives you ready-made word banks for this without having to invent them.

3. Dictation in context (all stages). This is the step that builds transfer to real writing, and it replaces the Friday test. The routine is short and the same every time:

• You say a target word, use it in a simple spoken sentence, then say the word again. ("Ship. The ship sailed away. Ship.")
• Your child writes the word — and, a few times a week, writes the whole sentence.
• If they get stuck, you guide them back to the sounds ("what's the first sound? what spells that sound?") instead of just giving the letters.

Dictation works because it asks the child to retrieve the spelling and immediately use it the way they'll use it in writing — which is the whole point that a spelling test misses.

The words that really do need memorizing

Now the honest exception, because pretending it away is how these articles lose credibility. Some words can't be fully built from patterns — the truly irregular high-frequency words like said, was, have, of, and they. These show up constantly in writing, so they can't wait. But the move still isn't blind memorization. Map the regular parts and flag only the tricky bit: in said, the s and d behave exactly as expected; it's only the ai spelling the short-e sound that has to be learned by heart. Teaching a child to spell the predictable letters normally and circle the one "tricky part" — some programs call these heart words, because that one part has to be learned "by heart" — is far more durable than treating the whole word as a random string.

Homophones are the other place where a little memory work is legitimate: to, too, and two sound identical, so the child has to attach each spelling to its meaning, and a bit of visual memory genuinely helps here. The point isn't that memorization never happens — it's that it should be reserved for the small set of words and word-parts that truly need it, not applied to every word as the default.

How to know it's working without a Friday test

If you drop the weekly test, how do you know your child is progressing? You look at the place that actually matters: their everyday writing. Pull a journal entry or a story they wrote and read it for pattern use, not perfection. Is the child applying silent-e correctly even on words they never studied? Are short vowels solid? That transfer — correct spelling in unstudied words during real writing — is the only assessment that proves the pattern was mapped rather than crammed. Quick daily check-ins during dictation tell you the same thing in miniature, and cumulative review (folding last month's patterns back into this week's practice) keeps mapped words from fading.

For practice that reinforces patterns without feeling like a test, screen-free word work helps: building words from sound tiles, or pattern-based word search puzzles that put a child face-to-face with a family of related spellings. And because spelling and reading grow together, time spent reading decodable books that match the patterns your child is learning reinforces the same mappings from the reading side.

Frequently asked questions

Should I ever give a spelling test at all? A test is fine as a periodic check on whether patterns have transferred — but make the test words ones the child hasn't drilled, so you're measuring the pattern, not the cram. A test of memorized list words tells you only that short-term memory worked on Friday, which is the thing that fails by Monday.

My curriculum comes with weekly lists. Do I have to ditch them? No. Look at the list first. If the words already share a pattern (most good programs group them this way), keep the list and just swap the memorize-and-test routine for segment, sort, and dictation. If the list is a random or theme-based jumble, keep the words but teach them through the patterns they happen to contain. You rarely have to throw the curriculum out — you change what you do with it.

My second grader reads beautifully but spells terribly. Should I worry? Almost certainly not. Reading (recognition) develops ahead of spelling (production) for nearly every child, so a strong reader who is a weak speller is the normal pattern, not a red flag. Keep building spelling through sound patterns and it will catch up. Persistent, severe difficulty across both reading and spelling is the combination worth asking a specialist about.

When should we start spelling, and how many words a week? Start informally as soon as a child knows letter sounds and can blend simple words — usually kindergarten — with sound-by-sound spelling of short words, not a list. Word counts matter far less than consistency; a handful of pattern words practiced daily and cumulatively will outperform a long list crammed weekly at any age.

If you change one thing this week, change the format: take this week's words, find the pattern hiding inside them, and practice with a daily five-minute sort and a short dictation instead of a Friday test. Then check next month's writing samples. The words that used to vanish over the weekend will start to stay.