Does my child have to master rhyme before starting letter sounds?

No. In one study, children who scored no better than chance on rhyme and sound-matching tasks were still using what they knew about letter names to work out letter sounds. Rhyme is useful. It is not a gate you have to pass first.

Is synthetic phonics better than analytic phonics?

The argument is much smaller than the public debate suggests. Pooling the trials, synthetic approaches came out at 0.45, larger-unit approaches at 0.34, and mixed approaches at 0.27, and those three figures did not differ to a statistically meaningful degree. All three beat teaching phonics unsystematically or not at all. Having a plan matters far more than which plan.

Is there a proven correct order for digraphs, blends and vowel teams?

Not one settled by a trial that we could find. The broad shape is agreed on across teaching schemes, and reviews have found the main families of systematic approach performing about the same as each other, which is a reason not to worry much about the fine detail.

When does the balance shift from sounding out to understanding?

In the first school years, how well a child works out the words explains most of the difference in what they understand. Later, language understanding takes over. For typically developing readers in English that shift lands somewhere around third or fourth grade, roughly ages eight to nine.

Is it too late if my child is already six or seven?

No, but earlier is easier. Systematic phonics scored 0.55 when it began in kindergarten or first grade and 0.27 when it began after that. Smaller, still real, still worth doing.

Try a lesson

Research, in plain words

The order children learn to read in

Ears before eyes, sounds before letters, letters before books. Here is the sequence, where it bends, and which bits of it nobody has actually tested.

The sequence

Read this as a rough running order, not a set of locked doors. Almost every child is working on two or three of these at once.

1. Hearing sounds, with no letters in sight
It starts big and gets small. First a child hears that a sentence is made of separate words. Then that “butterfly” has three beats in it. Then that “sun” and “sock” begin the same way. Last, and hardest, that “cat” is three separate sounds: /k/ /a/ /t/.
That final step is the one that matters most for reading, and it is genuinely difficult, because nobody speaks in separate sounds. We say “cat” as one smear of noise. Pulling it apart is a trick that has to be taught.
2. Letters: shapes, names, sounds
Now the letters arrive, and children learn both what a letter is called and what sound it stands for. The two support each other rather than competing, and for a good number of letters the name is a clue to the sound [2]. There is a whole page on this, because it turns out some letters are much harder than others.
3. Blending: pushing the sounds back together
Knowing that c says /k/ is not reading. Reading starts when a child can run /k/ /a/ /t/ together fast enough to hear “cat” come out. Short words with a simple consonant-vowel-consonant shape come first, because there is less to hold in the head at once.
This is the step where sounding out a word the child has never seen before becomes possible, which is the entire payoff of everything before it. Pooled across trials, teaching phonics systematically had its strongest effects exactly here: 0.67 on reading regularly spelled words and 0.60 on reading invented words that could only be worked out, never remembered[1].
4. Letter teams and longer words
Then English starts being English. Two letters making one sound (sh, ch, th). Clusters where both sounds survive (st, bl, str). The silent e that reaches back and changes the vowel before it. Vowel pairs like ai and ee. Words with more than one beat.
5. Actual sentences, actual books
Words in a row, then sentences, then something with a story in it. Early books are usually written so that almost every word can be sounded out with what the child has been taught so far, which is why they read a little oddly to an adult.
From here the job quietly changes. In the first school years, how well a child works out the words explains most of the difference in what they understand. Later, language understanding takes over, and for typically developing readers in English that handover lands somewhere around third or fourth grade [3].

Where common words fit in

Words like “the”, “said” and “was” do not sit at one point in the sequence. They are threaded through it, because they are too common to avoid and too irregular to sound out cleanly.

Worth knowing: the modern approach is not to have children learn them as whole shapes on flashcards. Even “said” is mostly regular. The s and the d behave normally, and only the middle is odd. Teaching it that way gives a child something to hang the word on rather than one more picture to hold in memory.

These are not locked gates

The neatest version of this sequence, where each step is a door a child has to unlock before the next one opens, is wrong, and it causes real anxiety in parents whose child is still shaky on rhyme.

Here is a concrete piece of evidence for that. In one study, researchers looked at children who scored no better than random guessing on rhyme judgment and sound matching, which are the standard tests of sound awareness. Those children were still using what they knew about letter names to work out letter sounds, in exactly the same pattern as everyone else[2]. The authors' conclusion was blunt: the phonological skill that standard tests measure is not required in order to start learning what letters say.

Which is not a reason to skip sound awareness work. It is a reason not to hold a child at step one for months waiting for a rhyme test to come good.

The synthetic versus analytic argument

If you have read anything about reading policy, you have seen this fight. Synthetic phonics teaches letter sounds first and blends them into words. Analytic and larger-unit approaches start from whole words and work down to the pieces, often through chunks like “-ight” rather than single sounds. In the United Kingdom this was a genuine political argument with government reviews attached to it.

The research is calmer than the argument. When the trials were pooled and split by type of approach, synthetic came out at 0.45, larger-unit at 0.34 and mixed approaches at 0.27, and the three did not differ to a statistically meaningful degree[1]. All three beat unsystematic phonics or none.

So the finding that survives is about having a plan, not about which plan. A child taught in a deliberate order does better than a child taught letters as they come up in a story. Which deliberate order is a much smaller question than the volume of the debate implies.

The fine detail is convention, not evidence

Should sh and ch come before st and bl, or after? Should the silent e arrive before vowel pairs like ai and ee? Teaching schemes have firm answers to these, and their answers mostly agree with each other.

We could not find a trial that settles any of them. What exists is broad agreement between schemes, plus the pooled finding that the main families of systematic approach perform about the same as each other [1], which is a reason to believe the fine detail is not carrying much weight either way.

Anyone telling you their particular running order is the research-backed one is going further than the research goes. Including us. We have an order, we think it is a good one, and we are not going to pretend a trial proved it.

Questions parents ask

Does my child have to master rhyme before starting letter sounds?: No. In one study, children who scored no better than chance on rhyme and sound-matching tasks were still using what they knew about letter names to work out letter sounds. Rhyme is useful. It is not a gate you have to pass first. [2]
Is synthetic phonics better than analytic phonics?: The argument is much smaller than the public debate suggests. Pooling the trials, synthetic approaches came out at 0.45, larger-unit approaches at 0.34, and mixed approaches at 0.27, and those three figures did not differ to a statistically meaningful degree. All three beat teaching phonics unsystematically or not at all. Having a plan matters far more than which plan. [1]
Is there a proven correct order for digraphs, blends and vowel teams?: Not one settled by a trial that we could find. The broad shape is agreed on across teaching schemes, and reviews have found the main families of systematic approach performing about the same as each other, which is a reason not to worry much about the fine detail. [1]
When does the balance shift from sounding out to understanding?: In the first school years, how well a child works out the words explains most of the difference in what they understand. Later, language understanding takes over. For typically developing readers in English that shift lands somewhere around third or fourth grade, roughly ages eight to nine. [3]
Is it too late if my child is already six or seven?: No, but earlier is easier. Systematic phonics scored 0.55 when it began in kindergarten or first grade and 0.27 when it began after that. Smaller, still real, still worth doing. [1]

Keep reading

What the science of reading actually means, and the one idea worth taking from it.
Teaching letter sounds, and why b and d get mixed up for so long.

Sources

Every source here was read before it was cited. Follow any number in the text to land on its entry.

[1]Ehri, L. C., Nunes, S. R., Stahl, S. A., & Willows, D. M. (2001). Systematic phonics instruction helps students learn to read: evidence from the National Reading Panel's meta-analysis. Review of Educational Research, 71(3), 393-447. doi:10.3102/00346543071003393
[2]Treiman, R., Pennington, B. F., Shriberg, L. D., & Boada, R. (2008). Which children benefit from letter names in learning letter sounds? Cognition, 106(3), 1322-1338. doi:10.1016/j.cognition.2007.06.006
[3]Catts, H. W. (2018). The simple view of reading: advancements and false impressions. Remedial and Special Education, 39(5), 317-323. doi:10.1177/0741932518767563