add create_norms() and get_norm_scores() functions?

kachergis commented 2 months ago

Background:

lots of people ask for norms, and we've duplicated effort (and scripts) across projects
it's a standard procedure (although we don't have demographics for many languages: just EN and SP):
look up by age and sex, look up in manual tables and interpolate between 5 percentiles, etc..

Potential solutions: 1) offer 'standard' script (on GH / email with proper warnings) -> or just a wordbankr function? (with good documentation and a vignette example) 2) scoring tool (in wordbankr, shiny, or both?): import your CSV (id, age, sex, sumscore) - too many people used it incorrectly... (need to clearly document expected mismatch between published tables and wordbank %iles) 3) add 'wordbank_percentile' column to https://wordbank.stanford.edu/data/?name=admin_data

@vmarchman , @alvinwmtan and I thought adding functionality to wordbankr makes the most sense, as it offers (sophisticated) users the option of using our SotA that should match published norms where we have them

kachergis commented 2 months ago

@mcfrank your thoughts welcome!

mcfrank commented 2 months ago

I think providing a column seems easy, but ideally a standard script that we use to provide percentiles via wordbankr would be great. I guess the issue is that our particular percentiles involve weighting etc...

On Thu, Jun 13, 2024 at 1:20 PM George Kachergis @.***> wrote:

@mcfrank https://github.com/mcfrank your thoughts welcome!

— Reply to this email directly, view it on GitHub https://github.com/langcog/wordbankr/issues/40#issuecomment-2166694937, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAI25F5JDUPC2ZU2CQTDHG3ZHH5I3AVCNFSM6AAAAABJJC3GQOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRWGY4TIOJTG4 . You are receiving this because you were mentioned.Message ID: @.***>

langcog / wordbankr

add create_norms() and get_norm_scores() functions? #40