Stuck at the Same DET Writing Score? You're Retelling, Not Analyzing

If you’ve taken the Duolingo English Test more than once and your writing score keeps landing in the same place — within five points of your last attempt — this post is about you.

The short answer: the most common cause we diagnose at LU English is retelling — writing answers built entirely from information already in the prompt. Retelling proves comprehension, not production, so it caps your writing score no matter how much vocabulary or grammar you add. Below, you’ll see it in real students’ own words, what the research calls it, and a 60-second test that tells you whether it’s yours.

You’ve probably already ruled out the obvious suspects. You studied more vocabulary. You reviewed grammar. Maybe you bought a word list or watched a channel full of “DET writing tips.” And the score didn’t move. (If your whole score is stuck, not just writing, start with the five thinking mistakes that keep DET scores from moving — this post goes deep on one writing-specific gap that guide doesn’t cover.)

Here’s a conversation we have almost every week at LU English. This is a real exchange from one of our lessons this month, with a student preparing for PhD admission — an engineering student, strong vocabulary, comfortable grammar. Only the identifying details have been removed.

Student: I just want to ask you — just want to confirm that I don’t have any problem in vocabulary, right?

Teacher: No, your vocabulary is good.

He was right to check, because everything he’d been told about improving his English score pointed at vocabulary. And everything he’d been told was wrong about him. His ceiling was somewhere else entirely — somewhere no word list can reach.

The symptom: your answer only contains the prompt

Here’s what our teacher told him next, word for word:

Teacher: The issue is you’re telling me exactly what the story says, and I already know it.

That’s the whole diagnosis in one sentence. When this student answered a DET-style writing task about a reading passage, every sentence in his answer was information from the passage, rearranged. Different word order, some synonyms swapped in, connectors like because and therefore used correctly — and not one idea that wasn’t already on the page.

Teacher: In your answer, you never really told me something that’s not in the text. You said, “here’s what the text said, I’m going to put it in my answer.”

He’s not unusual. In the past two weeks alone, three of our students — three different first languages, three different target scores, all preparing for university admission — hit this exact wall. The pattern is so consistent that we can usually spot it in the first piece of writing a new student sends us, before the first lesson has even started. And it catches strong students, which is exactly why they spend months fixing the wrong thing.

Researchers have a name for what you’re doing

This isn’t a personal failing, and it isn’t laziness. Second-language writing researchers have studied this behavior for decades under the name patchwriting — building your answer out of the source text with light modifications. Studies of learners writing from sources have found that second-language writers copy and make small word-level changes far more often than native writers do, typically by swapping in synonyms or deleting a few words from a borrowed sentence while keeping its structure intact. Read that description again, then look at your last practice answer.

Crucially, the research treats this as a training gap, not cheating — a stage writers pass through when producing original analysis in a second language is still out of reach, so they lean on the language the prompt already gave them. That’s the honest framing, and it’s also the frustrating one: it means the habit is normal, invisible to you, and stable. It does not go away on its own with more test attempts. It goes away when someone trains you out of it.

You can’t see it from the inside

Here’s the part that makes self-study so dangerous for this particular problem: students with this ceiling consistently misjudge their own writing — not by a little. Another real exchange, from a lesson this month reviewing a student’s practice essay:

Teacher: What would you grade this if you were grading this on the IELTS? What band score?

Student: I would put maybe like 6.5.

Teacher: No, no, this one’s much lower because it’s only 75 words. It’s too short, first of all. So consider all of it. What would you give it?

Student: Um, 4 to 5.

Two full bands between what she felt and what the essay was. She isn’t careless — she’s fluent enough to hold this whole conversation in English. But retelling feels like writing. The sentences are grammatical, the vocabulary is real, the connectors are in the right places. From the inside, it looks finished. That’s why students with this ceiling retake the test expecting a different result: the answer they wrote genuinely felt like a good one.

And when we ask these students to go beyond the prompt, the confusion is immediate and genuine. From the engineering student’s lesson, when the discussion turned to what his answer was missing:

Student: But where is the step here? … What is the conclusion from that part? Like, what is the lesson?

He isn’t being difficult. He honestly cannot yet see what a developed answer contains that his doesn’t — which is exactly why “just add more detail” and “develop your ideas,” the feedback plateaued students hear constantly, changes nothing. You can’t add what you can’t see.

Retelling is not the only writing gap — check which one you have

We diagnose three distinct writing failure modes at LU English, and they need different repairs, so it’s worth knowing which is yours before you train anything.

Running dry is a volume problem: your answer stalls around 50 words because the ideas stop coming. If that’s your pattern, the issue is structure, and it’s covered in Why You Run Out of Words on the DET Writing Sample.

The explicitness gap is a development problem: you have ideas, but you skip the small steps between them, assuming the reader fills them in. There’s a five-minute self-diagnostic for that one — the Sandwich Test.

Retelling — this post — is a source problem: your answer is full-length and connected, but everything in it came from the prompt. You can fail exactly one of these, or more than one. The tell for retelling is that your writing looks healthiest of the three: good length, good flow, good vocabulary, flat score.

Why it caps your score

Test scorers — human and automated — have a name for answers built from the prompt: recall. Recall proves comprehension. It cannot prove production, because every idea in the answer was supplied by the test itself. Duolingo’s own Official Guide describes the open-response writing tasks as evaluated by a grading engine trained to judge proficiency the way expert human raters do — and no expert rater, on any test, awards high marks for restating the prompt.

We grade IELTS writing professionally and teach the DET every day, and the line between the score bands is the same on both tests: at a certain point, restating stops earning anything. An answer built entirely from the prompt has a hard ceiling no matter how advanced the vocabulary inside it is — which is exactly why studying more words doesn’t move the number. You’re polishing the part of your answer that was never the problem. (For what the test expects at each band instead, see what the DET is actually looking for at 110, 120, and 130, and Duolingo’s official scores page.)

When our teacher finally showed the engineering student his ceiling directly, here’s what he said:

Student: Yeah — that point, my mind is stuck.

And here’s the answer he got, because it’s the honest version of what every plateaued student needs to hear:

Teacher: It doesn’t matter how many more lessons we do together — you’re not going to get the score you need if you don’t start thinking, “well, what does this really mean?”

The Highlighter Test: 60 seconds to your answer

Take the last practice answer you wrote. Any DET writing task with a reading passage or picture works.

Highlight every sentence that contains something which is not in the prompt — a consequence, a comparison to the real world, a conclusion the passage never states.

Now look at your answer. If almost nothing is highlighted, you’ve found your ceiling. It was never vocabulary. It was never grammar. You’re retelling, and the test already knows what the passage says.

Two more signs worth checking while you’re there. First, your last two official attempts landed within five points of each other despite real study in between — stable scores with growing vocabulary is the signature of a ceiling that vocabulary can’t touch. Second, the feedback you’ve received before says some version of “develop your ideas more,” and you honestly don’t know what that means in practice. Both are the same diagnosis wearing different clothes.

The real cost of not knowing

Every retake costs money, and for most of our students the money is the small part. People take the DET with something hanging on the result — an admission window, a scholarship condition, a visa decision with a date attached. Retaking on hope, with the same invisible ceiling in place, doesn’t just repeat the fee; it burns weeks you may not have and returns the same number. The most expensive thing about this problem is how long students spend not knowing they have it.

Why a blog post can’t fix this — and we won’t pretend it can

If retelling were fixable with a tip, the internet would have fixed it by now — there is no shortage of free DET advice. It isn’t a tip-sized problem. The research above calls it a developmental stage for a reason: it’s a thinking pattern, and in our experience it takes several lessons of live, back-and-forth correction to rebuild — a teacher watches you get stuck in real time, in your own writing, and interrupts the pattern at the exact sentence where it happens. That’s not something a listicle can do, and we’d be misleading you if we packaged three bullet points here and called it the solution. We’ve written before about the gap between knowing something and being able to use it under test conditions — why knowing the rule doesn’t raise your score — and this is that gap at its widest.

What we can do in a post is what we’ve done: show you the ceiling, prove it with real students’ real words and the published research, and give you a test that tells you whether it’s yours.

If the Highlighter Test just described you

Then you now know something most plateaued test-takers never find out: exactly why the score isn’t moving. That’s the diagnosis. The repair is what we do.

Book a diagnostic → — bring your last DET writing attempt (or write one fresh), and in the session we’ll show you, sentence by sentence, where your answers stop analyzing and what your realistic score timeline looks like. We’ll tell you the truth about that timeline even when it’s not what you want to hear; ask any of our students.

Not ready to book? Try the free DET practice and run the Highlighter Test on your own answers first.