Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)




The purpose of this pair of studies was to investigate three hypotheses generated by the frequency theory of verbal discrimination learning. The frequency principle suggests that a number of sequential events occur in verbal discrimination learning and that each event in the sequence leads eventually to a greater number of frequency units being associated with the correct alternative in each verbal discrimination unit.

It was hypothesized that an increase in study time would be positively related to acquisition of a verbal discrimination list. Second, it was predicted that discrimination pairs constructed of high frequency stimuli would be more difficult to discriminate than comparable discrimination pair constructed of low frequency items. Third, it was predicted that correct first trial guessing would enhance second trial discrimination.

The subjects were 80 female students, selected from the Introductory Psychology course at the University of florth Dakota. Forty subjects were used in each study and, within each study, were randomly assigned to one of two experimental conditions. The basic design in each study was a 2x2 factorial type with two study interval durations and two levels of subjective frequency for the discrimination pairs. The effects of correct first trial guessing were analyzed by means of the t statistic and Sign test.

Stimulus materials for Study I were selected from two syllable English adjectives with a frequency of occurance of one per 4.5 million. In Study II stimulus materials were selected from CVC trigrams with assocation values of 50-70 percent. All stimulus presentations were made by a projector and timed with a variable tape programmer.

Differential frequencies of the discrimination pairs were created through the use of a free recall list prior to verbal discrimination learning. The discrimination lists were of a mixed-list design, containing both high and low frequency units.

Four presentations of the discrimination list were given, in which the experimenter provided verbal reinforcement of "right" or "wrong". First trial responses were based on the subjects' guess as to the correct member in each pair. The dependent variable consisted of the number of errors during the final three learning trials.

The effect of varying the length of the study interval was not supported in either study. Although the order of the means in Study II was predicted, a similar ordering was not found in Study I.

The position that high frequency pairs should be more difficult to discriminate was strongly supported in each study. This result was viewed as strong evidence that an extra frequency unit added to a member of a high frequency discrimination pair, does not provide as much cue value as a frequency unit added to a member of a low frequency discrimination pair.

Enhancement of second trial discrimination by correct first trial guessing was given only very weak support in either study. The direction of differences among means supported the hypothesis in both studies. In neither study however were the magnitudes of these differences significant and only in Study I was a directional hypothesis supported statistically. These results were parallel to earlier work in which the guessing effect was found to be small, and its appearance / dependent upon specific reinforcement methodologies within the verbal discrimination task.

The writer concluded that the frequency hypothesis may be usefully employed as an explanation in learning a single verbal discrimination list. Further research was suggested which could lead to clarification of predictions generated by the frequency hypothesis.