Olly English
Olly is preparing your space...
Loading page content...
Olly is preparing your space...
Editorial standards
Every grammar unit on Olly English follows the same five-part structure, and every part of it exists for a stated reason. This page sets out that structure, the research it draws on, and — just as important — the things that research does not prove. If you teach, study, or are deciding whether to trust our material, this is the page to read.
Applies to Test English practice across all CEFR levels, A1 to C1.
A unit moves from understanding a form to producing it, removing support at each step. The order is fixed. The exercise type used to fill each slot is not — it is chosen per topic, by the test described further down.
You meet the form in context and connect it to a meaning — a time, a reference, an attitude. The question is what the sentence means, not how the verb conjugates.
The same target with fewer cues. Meaning stays clear; the support drops. You retrieve the form rather than pick it out of a line-up.
You separate the target from a near-miss that real learners actually produce, and the explanation names the boundary between them.
The form appears across a connected passage or conversation, where the surrounding text is what makes one answer necessary.
You produce the form yourself, with a model to work from. This is the block that asks you to write rather than choose.
Multiple-choice questions are quick to answer and quick to mark, and a practice site built entirely from them looks productive. The evidence says that is a trap. In a meta-analysis of 35 research projects, Shintani, Li and Ellis found that comprehension-based and production-based practice were about equally good at building receptive knowledge — but only production practice produced significant gains when learners were tested on producing the language.
So each unit carries a minimum share of items you have to type rather than select, and that minimum rises with level. It is checked automatically before a unit can be published: a unit that drifts back into recognition-only practice fails the check and does not ship.
Exercise variety is easy to fake and easy to get wrong. A different-looking task that can be solved without ever thinking about the grammar point teaches nothing. We use Loschky and Bley-Vroman's three-way test on every block before writing a single question.
| Relation | What it means | Verdict |
|---|---|---|
| Essential | The task cannot be completed without the structure. | We use it |
| Useful | Easier with the structure, but completable without it. | Early blocks only |
| Natural | The structure may occur; nothing requires it. | We reject it |
The practical move is what we call the defeat test: could a learner reach the right answer while ignoring the grammar point entirely? Word-ordering an article exercise fails it — you can sort a / apple / want / I from instinct and never once choose between a and an. Rewriting the sentence from singular to plural cannot be done without producing the plural, so that is the task we use.
More questions is the easiest thing to advertise and the weakest thing to rely on. Practice follows a power function: each additional question inside one session returns less than the one before it. In the study closest to how our units are used, learners practised twelve sentences per session, and what predicted durable knowledge was returning across three or four separate days — not doing more in one sitting. We size units for useful coverage of a point, then stop.
Citing research is only useful if the limits come with it. None of the sources below support the following, and we will not imply otherwise:
Deeper dives into individual grammar points, written with the same methodology outlined on this page.
Yes. Every unit is drafted by the Olly English content team and reviewed against a written methodology before it is published. The same five-block structure is applied to every grammar point, from A1 to C1.
We draw on second-language acquisition research, including Shintani, Li & Ellis (2013) on production practice, Loschky & Bley-Vroman (1993) on task design, and DeKeyser (1997) on practice effects. Full references are listed on this page.
Research shows comprehension and production practice build receptive knowledge equally, but only production practice improves spoken and written output. Every unit therefore contains a minimum share of typed answers, and the minimum rises with level.
Enough to cover the point without padding. We size units for useful coverage, because extra questions in one sitting return less and less learning per item.
Not yet. Durable knowledge comes from spaced return across separate days. A single unit gives you focused practice; mastery needs repeated exposure over time.
Primary sources for everything above. Where a paper is open access or has a permanent identifier, the citation links to it.
Meta-analysis of 35 research projects. Both practice modes helped receptive knowledge; only production practice moved production outcomes. This is why our units have a minimum share of written answers.
The essential / useful / natural distinction we use to accept or reject an exercise type.
Practice follows a power function — the reason we cap unit length instead of padding it.
Durable knowledge came from relearning sessions on separate days, not from more items in one sitting.
The case for focused, explicit instruction on a single point at a time.
Noticing — why our explanations name the cue and the most plausible wrong answer.
Form–meaning mapping, which is the job of the first two blocks.
Why every answer gets an explanation rather than a right/wrong mark.
Swain, M. (1995). Three Functions of Output in Second Language Learning. In Cook & Seidlhofer (Eds.), Principle and Practice in Applied Linguistics.
The output hypothesis behind the final block of every unit.
Council of Europe (2020). Common European Framework of Reference for Languages: Companion Volume.
The source of our A1–C1 level descriptions.