> Participants and the vocabulary test used
>
> For each vocabulary test, a random sample of 67 words and 33 nonwords was selected. For each letter string, participants had to indicate whether or not they knew the stimulus. At the end of the test, participants received information about their performance, in the form of a vocabulary score based on the percentage of correctly identified words minus the percentage of nonwords identified as words. For instance, a participant who responded “yes” to 55 of the 67 words and to 2 of the 33 nonwords received feedback that they knew 55/67 – 2/33 = 76% of the English vocabulary.
Granted, there's nothing stopping a participant from responding "Yes" to a word they can't define. But I think that's more likely to happen on words that the participant knows of even if they can't define it.
Here's the test, BTW [2]. It appears to still be up.
[1] https://link.springer.com/article/10.3758/s13428-018-1077-9#...