Phonics scope and sequence: Peabody's 16 steps on one page

Peabody (peabody.kids) teaches reading through a hand-written phonics sequence of sixteen steps, from s a t p i n to the suffixes -ing, -ed, -er and -est. Every word in a step can be sounded out with the letters of that step and the steps before it, no step uses an irregular word, and the irregular words a beginner needs (the Dolch list) are taught separately as heart words right after the step they fit. A child hears a word and taps it; the word is never printed above the boxes. Peabody is a free AI-driven adaptive learning platform for kids, free through the end of 2026, and this cheat sheet is the whole sequence as the code holds it. AI writes the stories; code decides the rest.

How to use this: find the step your child is on (the one where they can read most but not all of the example words), practice that step's tip for a few minutes, and let the step before it stay easy. A step opens in Peabody when 80% of the step before is solid, so the sequence always runs one step ahead of what is known.

The sixteen steps

Steps 1 to 7 contain no consonant blends at all: no spin, no ant. A plural -s is allowed early. The letters j, v and w arrive at step 5 so that jet, van and web have a home. ea is only the long e of leaf, ow is only the long o of snow, and y is a consonant at step 7 (yes, yak) and a vowel at step 13 (fly). The blend steps list their blends as graphemes so a child can see them; they are letter pairs, not new sounds.

The heart words taught after each step

A heart word is a word a beginner needs that cannot be fully sounded out with the steps so far. Each one is taught right after the step it fits, with its tricky part marked as the part to learn by heart. Steps 10 and 12 to 16 add none, because by then a child can decode almost everything on the list.

The first step's list is long because the first books need function words before a child can decode them: I see a cat is a level 1 sentence and only cat is on the phonics path. In Peabody's sight-word strand the Dolch lists come first in their own order, then any heart word not already in them.

Why the steps avoid blends until step 8

A word like sat has three sounds and three letters, one each. A word like spin has four sounds in four letters, but two of them, s and p, sit together with no vowel between, and a beginner tends to hear them as one lump ("sp") or drop one ("sin"). Blends are not new sounds, which is why they are cheap to teach later, but they double the amount a child has to hold in mind while blending, which is why they are expensive to teach early. Keeping steps 1 to 7 to one sound per letter (plus the two-letter digraphs sh, ch, th, ck, ng, nk, which are one sound each) means a child has about a hundred fully regular words to practise the core skill on before the words get crowded. The two blend steps then add about ninety words made only of sounds the child already has.

This is a design choice, not a law. Other sequences put a few blends earlier. The reason Peabody's does not is that every game and every first book uses the steps as written, and a word that leans on a pattern the child has not met turns a decoding question into a guessing one.

Why Dolch words are taught as heart words, separately

The Dolch list is the classic set of about 220 service words (the, said, was, of) that make up a large share of any early text. Many of them break the rules a beginner knows: said does not say "sayed", was does not rhyme with has. If they were mixed into the phonics steps, the steps would stop being decodable, and a child would learn that sounding out is unreliable.

So Peabody keeps the two lists apart. A phonics word is always decodable with the steps so far; a heart word is taught with its regular part sounded out and its irregular part learned "by heart". This follows Ehri's account of how words become sight words: not by memorising shapes, but by bonding each spelling to its pronunciation and meaning, letter by sound, with irregular words still mostly regular (Ehri's phases of word reading). Said is s and d as usual with an odd ai in the middle; that is one thing to learn by heart, not four.

In the practice brain, phonics (phon) and sight words (sight) are two separate strands on the reading ladder, which climbs first sounds → rhymes → phonics → sight words → spelling → word meanings. Spelling takes only the phonics and sight words a child already reads.

Orthographic mapping, in plain words

Orthographic mapping is the process by which a word a child has sounded out several times becomes a word they recognise instantly. It is not photographic memory. Each time the child matches the letters to the sounds and lands on the right word, the spelling gets bonded to the pronunciation and the meaning a little more firmly, until the word is read as a whole in a fraction of a second. The sounding out is what builds the bond; a word only looked at does not get one.

The research behind this:

• Ehri (2014) describes the phases a reader passes through: pre-alphabetic (guessing from pictures and shapes), partial alphabetic (first letter and a guess), full alphabetic (every letter, slowly), consolidated (chunks and whole words). The phase tells you what an error means: a first-letter guess at horse for house is partial alphabetic; a slow letter-by-letter h-o-u-s-e is full. Peabody's "which word did you hear" questions are built to need the full phase, because the three boxes differ by a vowel or an ending, not a first letter (Ehri's phases).
• Share (1995, Cognition) proposed the self-teaching hypothesis: each successful decode of a new word is itself the learning event for that word, so a child who sounds out shed correctly has taught themselves shed (pubmed). The practical consequence for software is to maximise successful decodes of new words, not displays of them.
• How many times. Reitsma (1983) found four to six exposures before a new letter string was recognised on sight; Ehri and Saltmarsh (1995) reported about 4.4 exposures for strong readers and 9.3 for weak ones (both as summarised in a 2020 review); Nation, Angell and Castles (2007) found learning rose with one, two and four exposures, with some words learned in one (Oxford). The numbers are small, which is encouraging, and they are for successful reads, not glances.
• Phonics itself. Ehri, Nunes, Stahl and Willows (2001, Review of Educational Research) pooled 66 comparisons from 38 experiments and found systematic phonics instruction helped with an effect of d = 0.41 overall, 0.55 when begun in kindergarten or first grade and 0.27 later (journal). Camilli and colleagues (2003) reanalysed and got 0.24 with equal weighting, so the size is argued over; the direction is not.
• Decodable text. Cheatham and Allor (2012, Reading and Writing) found decodability increases decoding-strategy use and accuracy, and argued for multiple criteria rather than decodability alone (doi:10.1007/s11145-011-9355-2). Peabody's levels 0 to 2 are checked for decodability at the child's step plus sight words, repetition and sentence length.

Castles, Rastle and Nation (2018) are the caution: phonics is necessary and well-evidenced for cracking the code, and not sufficient, because comprehension depends on language, vocabulary and knowledge that phonics does not supply (Ending the reading wars).

How to hear a step's words in a game: heard, not shown

Every reading question in a Peabody game follows one rule: when the word is the answer, Peabody says it, and it is never printed above the boxes that carry it. A "which word did you hear" question plays ship and shows three boxes, ship, shop and chip. A "starts with" question shows a picture of a fish and three letters, never the word fish. The distractors are near misses from the same step, so a child cannot answer from the first letter.

Why this matters: if the word ship were printed above three boxes reading ship, shop, chip, a child could match the shapes without reading anything, and the practice brain would record a right answer that taught nothing. Letter matching looks like reading on a tablet and is not reading. The research report behind this guide notes that with three boxes a blind guess is right one time in three, so a single right answer is weak evidence; Peabody retires a single question after one right but passes a phonics step only when most of its words are solid, and marks a word quick only when the answer came in about 2.5 seconds for a heard word or 3 seconds for a decoded one.

Where to find the steps in play:

• Reading mode in any game (every game has Mix, Numbers, Reading and Just play) draws on the reading ladder, so a child on step 6 hears step 6 words at the crates, jars, rocks and stations.
• Deep Dig at /games/dig is the phonics sequence made visible: its mining layers are the sixteen steps in order.
• Word Rescue at /games/word-rescue and Cell Break at /games/escape are word games (in Cell Break, letters finish the words chalked on the rocks).
• Every word on a first-book page can be tapped to hear it, and level 0 and 1 pages carry a picture for the word.

New words come at most two a round and four a day per domain, each taught once and rehearsed a few questions later; a missed word comes back three questions on; a known word returns for review after 1, 3, 7 and then 21-day gaps. The pass bar for the phonics rung is 30 words seen with 80% solid, and a step opens when 80% of the step before it is at "right first try" or better.

What this sequence doesn't tell you

• It is one good order, not the only one. Published programmes differ on where ck, qu and the blends go. What matters most is that the order is systematic and the text matches it, and that is what the tests on Peabody's bank check.
• It is not a complete phonics curriculum. There is no handwriting, no teacher modelling the blending, no oral segmenting drill. Peabody is practice on a path, best used beside a school programme or a parent on the sofa.
• Peabody cannot hear your child. It never records, so it cannot tell whether a child said ship or sip before tapping. "Quick and right" on a tablet is evidence, not a reading assessment.
• It does not diagnose. A child stuck on a step for weeks is a reason to talk to a teacher, not a result Peabody can interpret.
• Peabody has no published efficacy study. The sequence follows the evidence above; whether practising it in Peabody helps a particular child has not been tested in a trial.

The best two minutes a parent can add is the one Sénéchal and Young (2008) found to matter most: listening to your child read, and teaching the one specific thing they stumbled on (meta-analysis summary). Open a first book in the free library, let your child read a page, and when they stop on a word, cover its ending and ask for the first sound.

The free phonics games page lists the games that practise these steps, and reading level 1 shows what a first-level story looks like.

Questions parents ask

My child knows the letter names but not the sounds. Where do they start?

At step 1, and in Peabody's reading ladder a rung before it: first sounds, where a child hears a picture word and picks the letter it starts with. Letter names are a different skill from sounds, and the sounds are the ones that unlock words.

Why does Peabody teach said and the separately from phonics?

Because they do not follow the patterns a beginner has met, and mixing them into a decodable step teaches a child that sounding out is unreliable. They are taught as heart words right after the step they fit, with the regular part sounded out and the odd part learned by heart.

Is sh a blend?

No. sh is a digraph: two letters making one new sound, taught at step 6. A blend is two sounds side by side, each still said, like fr in frog, and those start at step 8. Peabody lists the blends as letter pairs so a child can see them, but there is nothing new to learn except saying both.

How long should a child spend on a step?

Until about 80% of its words are read right first try, which is the rule that opens the next step in Peabody. For most children that is days rather than weeks on the early steps. A child stuck on one step for a long time should have more words from the step read aloud with a grown-up, not more taps.

Does Peabody show the word while it says it?

Only when the word is not the answer. When the question is "which word did you hear", the word is heard and never shown above the boxes, because printing it would turn reading into shape matching.