W. R. Garner

The Stimulus in Information Processing

It is argued that the nature of the stimulus must be understood before we can ask meaningful questions about human information processing since it is the stimulus that determines the nature of the information to be processed. As one example, it is shown that redundant stimulus dimensions provide an improvement in information processing only if the dimensions combine to produce in effect a new dimension; and such dimensions are termed integral. And as a second example, inadequate performance may occur because stimuli are process-limited in that inadequate differentiation of stimuli occurs, or they may be state-limited in that inadequate energy is available for the stimulus to be represented in the organism.

Stimules configuration in selective attention tasks

The possibility that perceptual configuration of stimulus elements impairs the ability to attend selectively to individual elements was tested with two-element stimuli, constructed by placing two curved lines in close proximity. Ss rapidly classified series of these stimuli which could differ on both elements or on only one with the second held constant. It was hypothesized that if the two elements formed a configuration, then Ss should have difficulty attending selectively to the element relevant for classification while filtering information from the other element. This result was obtained in one experiment with both stimulus elements oriented vertically, and it is concluded that these stimuli were perceived as unanalyzed, nominally related shapes. In another experiment, with one stimulus element oriented horizontally, selective attention to the relevant element was possible.

The effects on speeded classification of implicit and explicit instructions regarding redundant dimensions

Two experiments were run in which speed of sorting decks of stimulus cards was measured. Stimuli were constructed from two dichotomous dimensions, used either alone or perfectly correlated. The lack of evidence for a selective serial processing (SSP) strategy (in which S sorts by his most preferred dimension whenever the dimensions are correlated) in a similar study by Garner and Felfoldy (1970) was thought to be due to Ss’ failure to notice that the dimensions were correlated, and thus that SSP would be effective. Ss in the present experiment therefore received either implicit instructions concerning the existence of the correlated deck (by seeing only decks in which there were just two different stimuli), or explicit instructions that on some trials the two dimensions would be correlated and S could sort by any means he preferred. When dimensions of size of circle and angle of diameter were of unequal discriminability, implicit instructions produced partial use of the SSP strategy, while explicit instructions produced nearly total use of SSP by all Ss. When the Munsell dimensions of value and chroma were varied in two separate color chips and were equally discriminable on the average, evidence for a small amount of SSP was found in both the implicit and explicit conditions. With neither pair of dimensions did implicit or explicit instructions regarding the correlated task produce integration of information.

Filtering and condensation tasks with integral and separable dimensions

Speed of classification was measured with a filtering task (only one dimension relevant) and a condensation task (two dimensions relevant) with integral dimensions of Munsell value and chroma and separable dimensions of value and size. The filtering task was easier with separable than with integral dimensions, but the condensation task was easier with integral than with separable dimensions. The results emphasize the importance of dimensional structure with separable dimensions and similarity or distance structure with integral dimensions.

Facilitation and interference with a separable redundant dimension in stimulus comparison

A comparison experiment was carried out with four letter stimuli such that some of the letter pairs provided hard and others easy discrimination when the pairs were different. A control condition confirmed the differences in difficulty with black letters. Two other conditions used two letters in each of two colors. When letter discrimination was hard, this redundant dimension produced facilitation of reaction time when the colors differed and were thus compatible with the “different” response required by the letters, and interference when the colors were the same, thus being incompatible with the “different” response. With easy letter pairs, only interference was found with incompatible color. Two additional experiments using classification procedures found color to have very little effect. Previous experiments had also shown different results concerning the separability or integrality of stimulus dimensions with the classification and comparison procedures. The interpretation for the difference suggested here is that response compatibility is an inherent aspect of the comparison procedure, since any dimension can have levels that are the same or different. The compatibility relation is available for only a few dimensional pairs with the classification procedure. When compatibility exists with either procedure, both interference and facilitation can occur, depending on whether the response interactions are compatible or incompatible.

Effects of focusing strategy on speeded classification with grouping, filtering, and condensation tasks

Speeded classification was studied with three tasks varying in the way in which extra stimuli are provided for a single class:grouping, with additional levels on a single relevant dimension;filtering, with additional levels on an irrelevant dimension; andcondensation, with additional levels on a second relevant dimension. (Both relevant dimensions must be processed for correct classification.) For the dimensions used (color and form), filtering was easiest, followed by grouping and condensation. In the latter two cases, asymmetric classifications, in which one class had a single member, eliminated the difficulty of classification. The presumed mechanism is focusing, in which the single stimulus is seen as a positive set and all other stimuli as negative.

Visual texture segregation based on orientation and hue

We investigated visual texture segregation, using a task for which reaction time to locate a discrepant quadrant in an array of 36 elements was the dependent measure. Two dimensions of segregation were used: line orientation (horizontal vs. vertical, horizontal vs. left diagonal, and left vs. right diagonal) and hue (9 vs. 7 Munsell color steps). Levels on these two dimensions were varied singly to produce control arrays, in orthogonal combination to measure possible interference effects, and in redundant combination to measure possible facilitation effects. Segregation for control arrays was much more difficult when the two diagonal lines produced regions of texture than when arrays used lines involving one or two main axes, and 7-step hue-difference arrays were more difficult to segregate than 9-step arrays, as expected. Orientation and hue were equally effective in producing fast texture segregation when used with optimal levels. The pattern of interference and facilitation for orthogonal and redundant combinations depended on the particular pair of line orientations used to produce texture: With horizontal and vertical lines, there was symmetric interference of each dimension on the other, as well as a redundancy gain. With horizontal and left diagonal lines, there was asymmetric interference of hue variation on orientation segregation, and no redundancy gain. With the two diagonals, there were symmetric interference and strong redundancy gains. These results clarify why different line orientations provide different effectiveness in texture segregation, and suggest that both interstimulus confusability (as with the two diagonals) and intrastimulus stability (as with horizontal and vertical lines) properties determine the ability of a dimension to interfere with, or to be interfered with by, another dimension in texture segregation.

Perceptual organization of nine-element auditory temporal patterns

All possible basic patterns formed from repeating sequences of nine dichotomous elements were presented auditorily, and Ss were required to describe the sequence after it became heard as an organized pattern. These descriptions show that whenever a particular pattern is a preferred organization, so also is its temporal reversal, whether that temporal reversal occurs in the same or in another basic pattern. Thus, it is argued that the pattern organizations are wholistic. The major organizational priniciples are: preference for (a) patterns that are balanced in time with long runs at the ends, and (b) patterns that have a directional simplicity with run lengths either increasing or decreasing in a regular order. Variability of pattern organization is related to the number of runs in the sequence, the range of run lengths represented, and the availability of a preferred organizing principle. It is concluded that the organism selects a pattern or organization from a set of alternatives presented by the structural features mentioned, and that this organizational selection is made on the basis of the whole pattern, not on parts of it built up sequentially.

The effect of goodness onencoding time in visual pattern discrimination

Previous results have implied that pattern goodness, as defined by rotation/reflection equivalence set size, is a determinant of encoding time. A direct test of this implication was made via a discrete RT task using good and poor dot patterns in a two-stimulus discrimination. The following discrimination conditions were used: (1) good vs good; (2) poor vs poor; and (3) good vs poor. Overall RTs for Conditions 1 and 3 were equal, but each was significantly faster than Condition 2. In the first two conditions, there was no difference in RT within pairs and no consistent transitive ordering of the patterns. In Condition 3, however, the good pattern consistently produced the smaller RT. These results indicate that patterns of equal goodness are equally encodable, and the better the pattern, the faster it is encoded. In addition, there was evidence that pattern similarity affects RT.

Speed of discrimination with redundant stimulus attributes

Speed of sorting decks of 32 cards with two alternative stimuli was measured. The stimuli were pairs of dots with attributes of distance between dots, angle of orientation of the pair, and position of the pair to the right or left of center. The attributes varied in relative discriminability and were used in all possible pair- and triple-correlated (redundant) combinations. Results showed an increase in speed of sorting with either an increase in discriminability or an increase in number of redundant attributes. It is argued that the increased speed with addition of redundant attributes is due both to a selective serial processing of attributes (requiring attribute separability) and to an increased discriminability with combined attributes (requiring attribute integrality).