Encoding In STM: Baddeley (1966)
Baddeley gave his particpants 4 sets of words. Set A was "acoustically similar" (e.g.man, cap, cab, mad, max, cat), set B was "acoustically dissimilar" (pit, few, cow. pen, sup, bay, day), set C was "Semantically similar" (great, large, big, huge, broad, long, tall, fat, wide, high) and set D was "semantically dissimilar" (good, huge, hot, safe, thin, deep, strong, foul).
When participants were asked to recall lists of words immediately (using STM) they did better on the acoustically dissimilar words than the acoustically similar words, suggesting STM memory codes are mainly acoustic. When participants were asked to recall word lists after an interval (using LTM) they performed just as well on list A as they did on List B, but differently on the lists with words that were semantically different. This suggests that LTM is encoded semantically.