David Coleman

Publication Date


Advisor(s) - Committee Chair

Sebastiano Fisicaro , Daniel Roenker, Doris Redfield

Degree Program

Department of Psychology

Degree Type

Master of Arts


The purpose of this thesis was twofold: to investigate the effects of categorized versus uncategorized material on selective attention and to test predictions derived from Filter Theory (Broadbent, 1958), Response Selection Theory (Deutsch and Deutsch, 1963), and Attenuation Theory (Treisman, 1969). Subjects performed a dichotic-listening task in which they shadowed a list of words presented to one ear (i.e., relevant message) while trying to ignore a simultaneously presented list of words on the other ear (i.e., irrelevant message). Lists were 16 words in length and consisted of either categorized words (C) or uncategorized words (U) presented at a rate of one word per second. Four conditions were generated by using all pairings of C lists and U lists for relevant versus irrelevant messages: U-U, U-C, C-U and C-C. Note that the left-most symbol designates the relevant message and the right-most symbol designates the irrelevant message. Subjects received two presentations of each of the four conditions. Measurements of pupil size were taken twice (9 sec and 5 sec) before the presentation of each dichotic trial (i.e., baseline measures) and at six positions (1, 4, 7, 10, 13 and 16) in the word lists (i.e., trial measures). Since each subject received eight experimental trials (two trials in each of four conditions), there were a total of four baseline measurements and twelve trial measurements for each condition. In each condition the four baseline measurements were averaged and the two trial measurements were averaged at each of the six positions. The mean baseline was subtracted from each of the six position means in each condition. These mean difference scores were used as the basis for one analysis. Shadowing errors (i.e. omissions of relevant words, mispronunciations of relevant words, or intrusions of irrelevant words) were scored by quadrants separately for each of the four conditions. The first quadrant consisted of the first through fourth words, the second quadrant consisted of the fifth through eighth words and so on. Error scores were then converted to percents and used as the bass for a second analysis.

A 4 by 6 ANOVA with repeated measures on both factors (condition and position, respectively) was used to analyze the pupil size data. The results indicated that pupil size decreased across serial position in a similar fashion for all conditions. Furthermore, pupil size did not differ significantly among the four conditions. A 4 by 4 ANOVA with repeated measures on both factors (condition and quadrant, respectively) was used to analyze the error rate data. The results indicated in an interaction between condition and quadrant. The C-U and C-C conditions resulted in a relatively constant error rate across quadrants, while the U-U and U-C conditions exhibited an increasing error rate across quadrants. The results of the two analyses are discussed in terms of their implications for Filter Theory, Response Selection Theory and Attenuation Theory.


Psychology | Social and Behavioral Sciences

Included in

Psychology Commons