Why NativeLingo works the way it does
Every rule in the product comes from a study or from simple arithmetic about your time. Where the evidence is thin, we say so.
Recall beats rereading
A week later, people who kept testing themselves remembered about 80% of the words. People who only reread them kept 33–36%.
Spacing fits the goal
The best gap between reviews is 10–20% of how long you need to remember. The FSRS scheduler works it out for each word.
Reading builds recognition
After ten or more meetings in a book, readers recognised 84% of new words but could recall the meaning of only 55%.
One word in a hundred
To read without a dictionary you need to know about 98% of the words on a page. Below 95% the text stops making sense.
So NativeLingo aims at 1% of the words on a page. It never goes above 2% on a page or 5% in one paragraph. You keep reading, and the new words stay rare enough to notice.
Eight to ten meetings per word
Clear gains start at around ten meetings with a word. NativeLingo shows each word at most once per page and twice a day.
After ten meetings, or once you remember it for three weeks, the word leaves the pages. If you miss it in a review, it comes back.
You type the answer
Looking at an answer and pressing "I knew it" is rereading, not recall. So the review asks you to type the word.
Typing also beats picking from options when spelling matters, and you read these words with your eyes.
When you get a word wrong, the right form stays on screen until you close it. Feedback after a mistake raised one-week retention by 494% in one study. After a right answer it changed almost nothing.
What it costs per day
Reviews grow slowly with time. After a year, each new word a day means about ten reviews a day. That is why choosing a pace means choosing minutes.
| New words a day | Reviews a day after a year | Minutes a day |
|---|---|---|
| 5 | ~50 | 7–10 |
| 10 | ~100 | 13–20 |
| 20 | ~200 | 27–40 |
What we do not know yet
No study tells us how lenient to be with a missing accent or a one-letter typo. Today a near miss counts as correct and shows you the exact form. We treat that as a setting to check, not a fact.
Nobody has measured whether meeting a word on a page changes how well you recall it later. We check these choices against review results inside NativeLingo.
Sources
- Karpicke, Roediger (2008). The critical importance of retrieval for learning. Science 319: 966–968.
- Cepeda, Vul, Rohrer, Wixted, Pashler (2008). Spacing effects in learning. Psychological Science.
- Pellicer-Sánchez, Schmitt (2010). Incidental vocabulary acquisition from an authentic novel. Reading in a Foreign Language.
- Hu, Nation (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language. Nation (2006), Canadian Modern Language Review.
- Webb (2007). The effects of repetition on vocabulary knowledge. Applied Linguistics 28: 46–65.
- Pashler, Cepeda, Wixted, Rohrer (2005). When does feedback facilitate learning of words? JEP: LMC.
- Nakata (2016). Effects of retrieval formats on second language vocabulary learning. IRAL 54(3).