210 likes | 309 Views
What kind of vocabulary is in course books and graded readers?. Rob Waring Notre Dame Seishin University JALT Vocab SiG Symposium June 29, 2013. Steps. Decide a scale to use (ERF Scale) Make a base wordlist based on the scale Scan in the texts and remove proper nouns
E N D
What kind of vocabulary is in course books and graded readers? Rob Waring Notre Dame Seishin University JALT Vocab SiG Symposium June 29, 2013
Steps Decide a scale to use (ERF Scale) Make a base wordlist based on the scale Scan in the texts and remove proper nouns Run the analysis in AntWord Count the running words in each text at each of the wordlist levels Identify a typical average frequency profile (by baseword level) at each reading level for the GRs and course books Decide the number of average texts to be ‘read’ (30) Decide how many times a word has to be met before it’s learnt (20)
Percentage of words at each ERF Reading level by Wordlist level Wordlist levelERF Reading level
% of families at each level which occur more than 20 times (minus proper nouns) Wordlist levelERF Reading level
Average book length without proper nouns % of each book in level
How many words do you meet if you read 30 books at each level?
Accumulated coverage for 30 books per level to 95% coverage of the families at 20 meetings for each type at each level Accumulated reading amount
How many words are you likely to ‘know’ (20 meetings) after reading all that?
Summary 450 books = 2894 ‘known’ words (20 meetings) Many words at each level won’t be met enough times to ‘learn’ them even after having read 30 titles at each level
Course books 6 Japanese Junior High texts 21 Japanese High school texts 18 Korean Middle School texts 15 Korean High School texts 5 Mexican Middle and Senior High texts
How many words will a learner meet on average in these texts in a middle or high school?
words metvs number of words probably learnt (>20 meetings) in various course books
Likely uptake (words met more than 20 times from reading 30 texts at each level)
Summary Course books only leads to low gains most words forgotten Course books plus reading doubles vocabulary BUT these data underestimate learning because the data • do not include partially known words (probably double that), collocations, colligations, multi-word phrases etc. • are unfair to the Mexico group who were restricted to low level reading (so we could compare)
It’s a work in progress …. Some levels in my wordlist need redoing • level 3 has lots of past forms and irregular verbs -> bump in data • level 6, 8, 15 & 16 are short of families Some levels short of texts • level 12 and level 15 Next I’ll … • add higher level texts 17-20 when they become available • replicate Paul’s study on how many words you need to meet to learn X,000 words with this corpus of SL texts • analyze which GR series best represents their stated levels • find out how many texts are needed before learners have covered say 05% of the words at a set level • re-do the stats for 12, 30 meetings
Phew! Yes Paul, I’ll publish it!