New Site for Vocabulary Experiments
The new site is at:
While eventually there will be a range of activity types and some spaced repetition practice, there is just a single activity type at the moment, based on work by vocabulary acquisition expert Paul Nation in the 1980s and 1990s.
It is a receptive vocabulary test, which means it focuses on whether you can understand a word when you come across it in text rather than whether you can produce the word in the right context. Each step of the activity asks you to select a word that best matches a given gloss, taken over a list of word-gloss pairs with a range of different frequencies.
Nation’s original tests (for English as a Foreign Language learners) used word lists split into frequency bands like the top 2000, top 3000, top 5000, and so on.
I took the common nouns in the Greek New Testament and similarly broke them in to frequency bands. Rather than have identically-sized buckets, I went by frequency cut offs:
- bucket 1 : 32 or more times
- bucket 2 : 16 to 31 times
- bucket 3 : 4 to 15 times
- bucket 4 : 2 or 3 times
- bucket 5 : 1 time
(Whether these are appropriate buckets will be assessed as part of this work.)
From each bucket, 36 word-gloss pairs were randomly chosen (the glosses coming from Dodson’s public domain glosses of NT lexemes). Of those 36, only 18 are tested, the 18 untested words used for distractors. This follows Nation’s approach.
So each activity of this type involves 90 items. I’ve so far generated two activities but it’s easy for me to generate more over time. I’ll also expand the items to other parts of speech and a larger Greek corpus (including Classical). As long as I have frequency information and glosses, I can easily generate activities.
I also have some other types of activities I’d like to implement, based on the research literature. I’d like to roll out a new activity once every couple of weeks or so.
There are some fairly basic, fundamental questions that I’ll be able to start to answer once I get more people trying the initial activities:
- how reliable is a test like this at estimating one’s NT Greek vocabulary size?
- how much is frequency a factor in how likely a student is to know a word?
- what other factors contribute to likelihood a student knows a word?
Future activities will be able to explore some of this in more detail such as the impact of English cognates or relatedness between lexemes due to derivation, etc.
Ultimately this is all input into producing better learning tools. It will feed directly into the adaptive online reading environment I’m currently working on.
Thank you to everyone who has tried the activities so far and PLEASE continue to do more activities as I roll them out and help spread the word. The more people of varying ability I get doing these activities, the richer and more insightful the data will be.
I’ll share those insights on this blog as things progress.