Priming is consistent but different across languages: The data release and results from the Semantic Priming Across Many Languages Project

Erin M. Buchanan & The Psychological Science Accelerator

Harrisburg University

Outline

  • The SPAML Dataset
  • Official Results
  • The semanticprimeR package

Many Thanks

Semantic Priming

FASTER

SLOWER

SPAML Stimuli

  • Selected by language word frequency and similarity using Open Subtitles
  • 1000 cue-target pairs
  • Matched across 31 languages/dialects

SPAML Procedure

Participants: Numbers

  • 35904 participants that opened the experiment
  • 30798 completed at least 100 trials
  • 26973 also performed at 80% accuracy

Participants: Languages/Geopolitical Regions

Participants: Gender

Participants: Education

Items

  • 101397 unique word forms
    • 47144 words
    • 54509 non-words
  • Words with 50+ answered trials
  • 75008 unique word forms
    • 35157 words
    • 40002 non-words

Priming: What Primed?

  • Total Pairs:
    • 19000 unique pairs with at least 50 answers on both the related and unrelated condition
    • 14170 priming pairs “work”
  • Matched Pairs:
    • 32 “work” in every language
    • 947 “work” in half of the languages
    • All “work” in at least 4/19 languages (dna-fingerprint, dvd-video)

Priming: What Primed?

Priming: What Primed?

Cue Target
capacity ability
washington boston
girl boy
dude bro
seller buyer
dog cat
november december
west east
internal external
near far
male female
two four
third fourth
apple fruit
kilo gram
sandwich hamburger
cocaine heroin
carbon hydrogen
romeo juliet
june july
ketchup mayonnaise
garlic onion
vagina penis
maybe probably
stand sit
daughter son
saturday sunday
what that
here there
hurricane tornado
apply use
normally usually

Priming: Consistent

Priming: Consistent

Priming: Some Variability

Data Creation

library(semanticprimeR)
  • Import word frequencies and/or word vectors import_subs
  • Calculate word similarity calculate_similarity and find the most related words top_n
  • Create non-words (simple) fake_simple
  • Create non-words (Wuggy) fake_Wuggy

Data Access

library(semanticprimeR)
  • View the data available data("primeData")
  • Import specific SPAML datasets import_prime
  • View related data available data("labData")
  • Import specific related datasets import_lab

Thank You

  • Much priming, big wow!
  • Very similar priming across languages, but some variability
  • Please use the data, stimuli, and more