Stimulus familiarization (n) in nonverbal selective learning

- Stimulus familiarization (n) in nonverbal selective learning' CLYDE E. NOBLE, JAMES P. ALLISON AND THOMAS A. JONES UNIVERSITY OF GEORGIA AND UNIVERSITY OF MONTANA One-hundred sixty Ss practiced lO-choice 4-unit tasks on the Selective Mathometer for 40 noncorrection trials under one of 8 different experimental conditions. The stimuli were paralogs of low meaningfulness and familiarity on Noble's !!l.. and Lscales, presented in a paired-associate fashion. The 8 matched groups of 20 Ss each were formed by 6 combinations of 3 levels of relevant stimulus familiarization and 2 kinds of criterion-task stimulus treatment plus 2 high-frequency irrelevant st·imulus familiarization and articulation control conditions. There were large, significant practice effects but no main effects or interactions due to the other experimental conditions. The null results pose difficulties for theories of stimulus pre differentiation. It is known that experimentally-controlled familiarization <!!> increases familiarity, strengthens free recall, and facilitates serial and paired-associate verballearning proficiency, the amount and direction of the effects depending upon frequency, temporal, and articulation factors (Baker & Noble, 1965; Noble, 1963). Comparable experiments for selective-learning tasks have not been performed, although a number of studies of "verbal pretraining" and "stimulus predifferentiation I I share common features with verbal familiarization research. The pertinent literature has recently been reviewed by Ellis (1965), Hall (1966), and Noble (1966). In this "transition" experiment a variant presentation method is used. It requires S to attend carefully to successive stimuli in order to select the correct reaction keys with promptness and accuracy. Although the pretraining is eminently verbal, S's verbalizations do not indicate the correct manual reaction to be associated later with each stimulus; articulation during familiarization is logically unrelated to responses demanded by the criterion task in the transfer phase. On the other hand, stimulus pronunCiation during transfer is not incompatible with key pressing, so the problem of verbalizations interfering with criterion responses is avoided. According to predifferentiation theorists, stimuli of low meaningfulness (!!!l and familiarity (!.l are initially not very distinctive to control Ss. Therefore, if the familiarization treatment is relevant (assuming task comparability). then experimental Ss should profit from increasing amounts of such pretraining, while Ss who pronounce the stimulus term during the transfer phase should outperform Ss who do not. Negative results would imply additional restrictions on the range of application of predifferentiation theories. Method The Ss were 160 undergraduates at the Universities of Montana and Georgia. There were 77 men and 83 Psychon. Sci., 1966, Vol. 5 (12) women. whose ages ranged from 15 to 58 yr. (mean = 22.7 yr.). All were naive to verbal and selective learning. The apparatus used for both matching and criterion tasks was the Selective Mathometer (MK II), described elsewhere in detail (Noble. 1966; Noble. Fuchs. & Thompson. 1963). An S's problem is to discover which one of 19 pushbuttons is to be selected in response to each slide presented on a projection screen. The reinforCing stimulus for correct responses (R+) is the onset of a 3-w green light centered above S's keyboard. All events are recorded by counters and polygraph. The time constants were: stimulus duration = 2 sec •• interstimulus interval = 2 sec., intertrial interval = 8 sec.. reward delay = 0 sec .• reward duration = 3/4 sec. For the 6choice 6-unit matching task 6 keys were unmasked (no. 3. 5. 9. 11.15.17). and all were relevant. For the 10choice 4-unit criterion task 10 keys were available (no. 1. 3. 5. 7. 9. 11. 13. 15.17.19). but only 4 were relevant (no. 1. 7. 13. 19). Stimuli for the matching task were 6 distinctive electrical circuit symbols; stimuli for the criterion task were 4 disyllabic paralogs of low meaningfulness (mean !!!=1.26) and low familiarity (mean.! = 0.28) on the !!! and.! scales (cf .• Noble. 1963). All Ss practiced by the E-paced noncorrection procedure. receiving 10 trials of an invariant S-R sequence on the matching task and 40 trials of a variant S-R sequence in 8 different orders on the criterion task (or 5 successive perfect trials). stimuli for the familiarization task were prepared in 2 lists of 8 paralogs each (mean !!! '" 1.28; mean.! = 0.33) andpresentedbyaStoeltingmemory drum at a 2-sec. rate in 6 different orders. The Ss were instructed to pronounce each paralog aloud as it appeared and to maintain consistency of pronunciation. The experiment was divided into 3 phases: matching task. familiarization task. criterion task. After the first ,hase. 20 Ss of comparable ability were assigned to each of 8 treatment conditions to be administered in the second and third phases. There were three levels of stimulus frequency <!!> during the familiarization phase. denoted as O-F. 20-F. and 40-F. Each of these conditions was followed by one of two kinds of criteriontask stimulus treatment. denoted as F-P (pronunciation) and F-NP (no pronunciation). The above 6 combinations gave 4 groups of Ss receiving relevant familiarization experience (20-F-P. 20-F-NP. 40-F-P. 40-F-NP) and 2 groups with none (O-F-P. O-F-NP). Two additional control groups received high-frequency (!!=40) irrelevant familiarization experience in the second phase and one of the two articulation treatments in the third phase; they are denoted as 40-I-P and 40-I-NP. Four 453 1.00 a. a:: >- .80 ~ ::i iIi .60 P ___ NP } 40-1 0 - - . NP « CD a:: a... ~P } 40-F .40 L>--6P } _NP 20-F W (/) Z 0 a... ~P } .20 _NP O-F (/) w a:: .00 0 3 8 13 18 23 28 33 38 MIDPOINTS OF 5-TRIAL BLOCKS (Nl Fig. 1. Rp curves for the 8 groups as a function of N. Each curve contains 20 Ss. of the 8 paralogs presented during the second phase were critical items for Ss in the 20- F and 40- F groups; the remaining 4 stimuli were fillers. No Ss in the O-F groups received familiarization; Ss in the 40-1 groups were familiarized with 8 stimuli not encountered in the third phase. All details of programming the stimulus and response permutations during the third phase were based on a previous experiment (Farese & Noble,1960). Results Acquisition curves for all 8 groups on the criterion task are shown in Fig. 1 plotted in terms of response probability (lip), or relative frequency of R+, as a function of successive 5-trial blocks (N). Each curve has the familiar skewed, sigmoidal form characteristic of selective-learning tasks (Noble, 1966), and is approaching mastery by the sixth block of practice. A 6 by 3 by 2 mixed-factorial Trials by Familiarization by Pronunciation analysis-of-variance test was performed on the R+ data of the relevant-familiarization experimental and control groups over Trials 1-30. The Trials main (...truncated)