On the locus of the stimulus familiarization effect in young children

Twenty kindergarten Ss received 50 reaction time trials. They withdrew their finger from a button as soon as 1 of 2 colored lights came on. 1 color occurred on the familiarization trials 1–20 and a 50:50 random sequence of that familiar color and a novel one was given on the last 30. Faster speeds to novel than to familiar indicate that the stimulus familiarization effect is not attributable to competing attending responses learned during the familiarization which then interfere with responding to the familiar stimulus during test trials, as has been suggested. Thirty-two kindergarten children observed 1 color for 3 seconds on each of 20 trials and then performed in a discriminative reaction time task in which the familiarized color and a novel one each serve as the signal to initiate (immediately withdraw the finger from a starting button) and execute (press that 1 of 2 buttons having the same color as the signal light) a discriminative response. One-half the Ss had sequences of novel and familiar stimuli with a tendency to alternate (TA) 75% of the time; for the other one-half the stimuli repeated (TR) 75% of the time. Initiation, but not execution, speeds were greater to the novel than to the familiar stimulus. Execution, but not initiation, speeds showed the change effect (CE) (found by Cantor and Cantor in two studies) with the TA sequences and a reverse CE with the TR sequences. The results show that the CE is an artifact and also that the hypothesis of surprisingness of the novel stimulus is probably false.