Which works better for SRS vocabulary drill and general language acquisition: standardized vocab lists, vocab lists from lessons, or generated vocab lists from immersion materials?

Standardized frequency lists (the General Service List, Academic Word List, or BNC/COCA-derived lists) are built so that a relatively small number of words covers a very large share of the words you'll actually encounter — the first 2,000-3,000 most frequent word families typically cover 80-95% of running text in general and even academic materials.^6,9,10,14

When researchers compared teacher- or textbook-chosen vocabulary lists to what students themselves picked as unfamiliar or useful from the same texts, the teacher-chosen words were concentrated in the most common 3,000 words, while student-chosen words spread much wider, including many mid- and low-frequency words the teacher had not flagged.⁴

Word frequency alone is a moderate but incomplete guide to which words are worth teaching; adding instructor ratings of a word's usefulness and difficulty, or machine-learning ranking that factors in context and collocation, produces word lists that experienced teachers judge more pedagogically useful than frequency-only lists.^1,2

Standardized general-purpose lists leave large gaps for specific domains: medical, accounting, or engineering texts require tens of thousands of additional specialized words beyond the general 3,000-word core, and general classroom vocabulary lists commonly cover only about half of the words actually appearing in local textbooks and syllabi.^12,14

Learners do pick up new words simply from reading novels or listening to authentic audio without deliberate study, but the gains are partial and depend heavily on how often a word appears — real, noticeable learning typically requires roughly 6 to 20+ exposures, and even after many exposures, learners are much more likely to recognize a word's meaning than to be able to recall or spell it.^3,5,8

In a direct comparison, listening to an audiobook version of a novel produced significantly larger vocabulary gains than reading the same text, with gains still measurable three weeks later, suggesting immersion-derived vocabulary lists built from listening material may be especially valuable.¹⁵

Extensive reading and listening are not equally good at teaching all types of words — nouns tend to be learned well for their meaning even without much repetition, while verbs need many repeated exposures before learners can actively use their correct form, meaning immersion-generated lists may under-represent verbs unless the source material repeats them often.^7,13

Immersion or reading materials only support vocabulary growth if they are pitched near the learner's current level; texts written for children or general native audiences require knowing around 10,000 words for good comprehension, while texts purpose-built for language learners (graded readers) need only about 3,000 word families, making them far more efficient sources for generating a learner's vocabulary list at lower levels.⁹

Whatever the source of the vocabulary list, actively drilling the words — using spaced repetition or flashcards, rather than only encountering them passively in a lesson or immersion material — produces measurably larger and faster vocabulary gains.^3,11,16

How many times do I need to see a new word before it actually sticks?

Across several incidental vocabulary studies, meaningful learning gains started appearing around 5-8 exposures, with a much bigger jump at 10-17 exposures; going beyond that (28+ exposures) helped further but with diminishing returns. Even at high exposure counts, learners were more likely to recognize a word's meaning than to correctly recall or spell it, so seeing a word often is necessary but not sufficient — some form of active review still helps solidify it.