The Effect of Distraction on Selective Attention Author(s): John W. Hagen Source: Child Development, Vol. 38, No. 3 (Sep.

, 1967), pp. 685-694 Published by: Blackwell Publishing on behalf of the Society for Research in Child Development Stable URL: http://www.jstor.org/stable/1127246 . Accessed: 10/10/2011 09:42
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishing and Society for Research in Child Development are collaborating with JSTOR to digitize, preserve and extend access to Child Development.

http://www.jstor.org

ON SELECTIVE THE EFFECT DISTRACTION OF ATTENTION

JOHN W. HAGEN University of Michigan The effects of a vigilance distractorwere investigated with respect to recall of task-relevantand task-incidentalinformationon a memory task. 40 children per grade in the first, third, fifth, and seventh grades were given the task; half were tested under distractionconditions.It was found that recall of task-relevantmaterial increased regularly with age, while recall of taskirrelevant(incidental) materialshowed no significantchange with age level. Distraction had a very significant effect on the task-relevanceperformance but did not affect incidental performanceexcept at the oldest grade level. No interaction was found between type of recall measure and distraction. Correlationsbetween central and incidental performancewere significantly negative at the seventh-gradelevel only. These findings are consistent with the information-processing model presented, which accounts for the development of selective attentionprocesses. That attention and memoryabilities increasewith age during childhood is well known, but the reason for the improvementis far from established. To offer an explanation,Maccoby and Hagen (1965) have formulated an information-processingmodel based on previous work of Broadbent (1958). Broadbentconceptualizesthe nervous system as a single communication channel with a limited capacity. Within this limited capacity, selective "filters" operate to separate relevant from irrelevantinformation.These selective mechanisms encode informationon the basis of some fairly gross
This study is based on a thesis submitted by the author to Stanford University as partial fulfilment for the requirements for the Doctor of Philosophy degree. The author was a predoctoral fellow of the National Institute of Mental Health at the time. The author wishes to express thanks to Earl Whitaker, Assistant Superintendent of Schools, Redwood City, California, for his cooperation, and to Linda George and Richard Ashmore for their assistance in this study. Author's address: Department of Psychology, University of Michigan, Ann Arbor, Michigan 48104.

CHILDDEVELOPMENT characteristics. Further processing then occurs only for the set-relevant materials. Although Broadbenthas not discussed the development of such selective mechanisms, it seems reasonable to assume that the channel capacity of every organism is limited, but for some the capacity is better utilized The inabilityof childrento attend through the use of more efficient"filters." A selectively may result from lack of adequately developed "filters." recent study by Maccoby and Hagen (1965) investigatedthe role of distractionon central and incidental recall at different chronologicalages. For this study, the memory task used included both a central (task-relevant) and an incidental (task-irrelevant)measure. It was found that (a) recall of central (task-relevant) materialincreased regularlywith age; (b) recall of incidental or irrelevant material did not increase between grades 1 and 5, and decreased between grades 5 and 7; (c) distraction impaired performance on the central task but did not on the incidental task; and (d) performance scores on the two tasks were independent. This last finding was disturbing in view of the information-processing model being investigated. The present study was designed, therefore, to test more definitively the hypothesis concerningthe development of selective attention processes previously formulated. Briefly, this hypothesis states that under conditions which require much informationto be processed the efficient informationprocessor excludes irrelevantor task-incidentalinformation.If older childrenhave betterdeveloped selective attending abilities, they should be better able to "give up" incidental information under overload information conditions than younger children. However, incidental informationmight very well be processed under certain conditions and be indicative of high proficiency-of the ability to process several sets of informationat one time by "dividing attention." A distractor that required a certain amount of attention was therefore included in the experimental design. The findings cited above supported the predictions in several respects: (a) Task-relevantperformance increased with age, while task-irrelevantperformancedid not; (b) The distractorlowered task-relevantperformance,but did not impair taskirrelevantperformance.However, in two respects the predictions were not supported: (a) The effect of distraction on task-relevantperformancedid not decrease with age, and (b) The negative correlation between taskrelevant and task-irrelevantperformance at the older age levels was not found. This finding is necessary if one is to accept the interpretationthat task-relevantperformance is maintained at the expense of incidental cue attending. The experimentalsituation was altered from that used in the original study (Maccoby & Hagen, 1965) in order that the hypothesis could be more definitivelytested. The task-relevantand task-irrelevant features of the stimulus cards were counterbalanced.A second advantage of utilizing a 686

JOHN W. HAGEN new task was to determine whether the earlier results could be replicated with different materials. The materials were selected so that they were of potentially equal interest to Ss at each of the ages tested, whereas the pictures with bright colors and toys used in the earlier study may have been more attractiveto the younger Ss. Finally, a new conditionwas added. Since the irrelevant pictures on the test cards could be functioning as visual distractors, a measure of performance on the central task without the presence of the additional pictures was included. Each presentationof stimuli contained only half as many pictures. Insofar as the irrelevantpicture on the card was not ignored, the informationload would be increased in the two-picture condition. If no differences were found between these two central memory measures, it could be concluded that the second pictures did not have a distracting effect. If differences between the two central task measures were to interact with age, a clear prediction is that the differences should occur at the younger age levels. If differenceswere found but no age interaction occurred, it could be concluded that the irrelevantpicturesfunctionlike the audio distractor used in the earlierstudy. METHOD Subjects The Ss were 160 middle-classchildren, 40 each in grades 1, 3, 5, and 7, enrolled in public schools in Redwood City, California.The CA range of each grade was as follows: grade 1, 6-7; grade 3, 8-9; grade 5, 10-11; and grade 7, 12-13. Only Ss who were at their appropriate grade level were included. There were 22 boys and 18 girls in grades 1 and 3, and 20 boys and 20 girls in grades 5 and 7. The three older groups on whom measures of general ability (as measured by the CaliforniaTest of Mental Maturity) were availablewere equated as much as possible on this measure: third grade--M = 118.0, SD = 15.8; fifth grade-M = 112.5, SD = 17.1; seventh grade-M = 117.3, SD = 12.4. Grade 1 Ss were selected randomlyfrom the classes. Grade 7 Ss were in junior high school, but only those were used who had attended the grade school that the younger Ss were currentlyattending. Central Task Materials White cards measuring 3 x 6 inches were used, with each card containing two black line drawings, each picture about 2 inches across. The pictures, of objects familiar to and readily labeled even by kindergartenage children, were divided into two conceptual categories, animals and household objects. An instance of each of the two categories always appeared on each card, the two pictures placed immediately above each 687

CHILDDEVELOPMENT other. The cards were arranged in series varying in length from four to six cards. A controlconditionrequiredcards containingonly one class of pictures, either animals or household objects. Each set was identical to the main stimulus cards except for the absence of the irrelevant-class pictures. Each S received only that series of control cards which contained the pictures he was to find in the main test series. DistractorTask A distractorwas used to require S to remain vigilant for a stimulusunrelated to the central task. A tape recording of a melody of high notes on a piano, which was interrupted periodically by single low-pitch notes, was used. The average occurrence of low notes was one per 11.5 seconds. The S was required to tap the table with a pen whenever he heard the low note. Subjectsin all age groups tested could performthis task while carrying out the central task. Procedure The Ss were tested individually. Each was told that the task to be performed was a memory game and that he would be shown pictures on cards and asked to remember the location of certain pictures. A practice trial was presented which used pictures never actually appearing on the test cards. The practice card was shown, and S was asked to identify the pictures. Then the practice card was covered with the blank panel, and S was asked to point to one of the pictures. He was told that he did not have to point to the exact placement of the card but just to the card that contained the picture requested. Then the actual trials began. The one-picture-per-card condition was administered first to half of the Ss; for the other half, the two-picturesper-card condition was administeredfirst. Each series consisted of 12 trials. Seven different pairs of pictures were actually used, and they appeared repeatedly in successive trials, but the position in the displays was varied randomly. After each trial, E covered the display cards and held up a 3 X 3-inch card containing the picture which S was to find on that trial. After S pointed to the card he thought was correct, the display was shown again briefly so that the child could see whether he was correct. The task was designed so that all Ss were given the same exposure to the incidental cues regardless of their score on the central task. Two central memory scores, for one picture and for two pictures per card, were recorded for each S. The score was the numbercorrectin 12 trials. After the series of two pictures per card, regardless of the order in which it was given, the measure of incidental recall was made. For this 688

JOHN W. HAGEN measure, the animal pictures were prepared on 3 x 6-inch cards, and the animal was placed in the same position it actually occupied on the test cards. The household objects were prepared as individual cutouts and covered with plastic. These cutouts were placed before the S, and he was told to try to match the animals with the household objects with which they had always appeared. Then the cards containing the animals were shown one at a time, and S was instructed to pick up the household object that he thought was correct and to place it on the blank part of the animal card. After each response, the household object was placed back in the array so that S chose from the whole array each time. Thus, S was allowed to choose the same household object more than once. All seven animals were shown, and the score for incidental recall was the number correctout of seven. For Ss assigned to the audio-distractioncondition, the following procedure was used. At the beginning of the experimentalsession, the S was told that he would be doing two things at once. The tape recorder was started, and E held a pen in view and explained that S was to tap the table with the pen whenever he heard a very low note. Then S was given the pen for several practice trials. When S responded correctly on several trials, the recorderwas turned off and the instructionsfor the central task were given. The tape recorder was operated by means of a foot pedal and was played only during the 5-second exposure periods during the testing. Only one S had to be eliminatedbecause of inability to perform correctly on the tapping task. No distractionwas used during any of the testing for incidentalrecall. Reactionsof the Ss to the Experiment All children seemed to enjoy the task, and in no case did a child refuse to be tested or leave duringthe experimentalsession. Every effort was made to hold the experimental conditions constant across age levels as well as experimental conditions. Several seventh graders spontaneously called the task a "test," so it is possible that they considered it a measure of competence, at least more so than did the younger children. However, even the first graders exhibited a high task involvement;and if any real motivational differences existed across grade levels, they were not evident in the experimentalsituation.
RESULTS

The main results are shown in Figure 1 and Figure 2. Correct performance on the central task is an increasing function of age level and of nondistractingconditions. A pseudo-four-wayanalysis of variance (Winer, 1962, p. 337) reveals that all three of the main variables for the central 689

CHILDDEVELOPMENT
10

/
/ /
5
--

/
2 PICTURES/CARD DISTRACTION NO DISTRACTION

/
I I I 3 I 5

I 7

FIG. 1.-Central task performance at grades 1, 3, 5, and 7.

-S----

DISTRACTION NO DISTRACTION

I I 7 5 GRADE 2.-Incidental task performance at grades 1, 3, 5, and 7. FIG. task yielded significant differences. Both the age effect (F = 30.50) and distraction versus nondistraction (F = 18.54) are significant well beyond the .001 level in the expected directions. The one-picture-per-card control condition was significantly easier overall than the two-pictures-per-card condition (F = 7.02, p < .01), indicating that the presence of the second "distracting" picture did indeed have a deleterious effect at all age levels. None of the interactions was significant (F < 1.01). The failure to find the expected interaction between age and auditory distraction may indicate that this distractor also was effective as a suppressor of central task performance at all age levels tested. A two-way analysis of variance was performed on incidental recall scores. For incidental recall, there was no significant difference either by 690

I I

I 3

JOHN W. HAGEN (F age level or by distraction-nondistraction < 1.00). But inspection reveals that changes in performance occurred at the seventh-grade level. The incidental recall scores under distractionwere lower than at any of the other grade levels, as predicted, but under nondistractionthere seems to be a rise. The difference between distraction and nondistraction incidental recall for seventh grade was significant (t = 2.56, df = 38, p < .02). This internalanalysiswas performedsince the previous study as well as the model predicted the difference (see Winer, 1962, p. 208). Thus there is evidence indicating that the seventh graderswere "giving up" incidental information, possibly in order to receive the necessary central information. The increase found under nondistractionconditions for seventh graders is in contrast with the earlier finding of a decline under both experimental conditions, and lends support to the hypothesis advanced earlier that the less childlike nature of the pictures used in this study was more salient for the older Ss. Another pseudo-four-way analysis of variance was performed with central two pictures per card versus incidental performance used as the repeated measure in order to look at possible interactionsbetween central and incidental performance across age and across distraction. There was a significant interaction effect, distraction-nondistraction type of recall by measure (F = 113.96, p < .001). Distraction affected central recall scores but not incidental. Also, the grade level by type of recall measure was found to be significant (F = 27.81, p < .001), indicatingthat centralrecall scores increased with age but incidental recall scores did not change with age. These findingsreplicatethose of the earlierstudy. In order to test further the hypothesis of a "giving up" of incidental or task-irrelevant information,the correlationsamong the three main measures were obtained. One would expect that, if the interaction between central and incidental informationwere such that concentrationon relevant cues resulted in an inability to concentrate on irrelevant cues, the correlation between performance on the two measures would be negative. Because this ability to attend selectively supposedly increases with age, one would expect this negative relation to increasewith increasinggrade level. Positive correlationswere obtained among the central and incidental measures for grades 1, 3, and 5. However, at the seventh-gradelevel, the overall correlation was -.35 (p < .05). The change from fifth to seventh grades was significant,both combining across distractionand nondistractionconditions (z = 2.65, p < .01) and also for each one computed separately (distraction: z = 1.79, p < .05, one-tailed;nondistraction:z = 1.76, p < .05, onetailed). That this occurredat the seventh-gradelevel, where the changes in incidental performance were suggested by the earlier presented data, is important for the theory and is discussed later. The correlationsbetween the one-picture-per-card performanceand incidental recall also showed an interesting trend. Under distractionconditions, the results were remarkably 691

CHILDDEVELOPMENT similarto those obtained with the two-pictures-per-card measure,but under nondistraction,the pattern of correlationsfell apart. It might be that, in this simplest of all conditions, with only one picture and with no audio distractor,there is not enough informationload at any age level to cause a "givingup" of any information. The correlationsbetween task performance and the California Test of Mental Maturity scores were computed. The significant correlations obtained for the central tasks are those for seventh graders under nondistracting conditions between both the one-picture (r = .77) and the twopictures-per-card (r = .66) conditions (p < .01). Also, the correlation between incidental task performance and intelligence, combining grade levels, was significantlynegative (r = -.27, p < .05). Hence only under nondistracting conditions did any of the task performancesrelate to the obtained measure of cognitive functioning, intelligence. The directions of the obtained relationshipswere as expected: central performancecorrelated positively, and incidental performance correlated negatively. The central increasedwith age undernondistracting correlations performance-intelligence conditions (one picture: grades 3 vs. 7, z = 3.12, p < .01; 5 vs. 7, z = 2.68, p < .01; two pictures: grades 3 vs. 7, z = 2.62, p < .01; 5 vs. 7, z = 1.88, p < .10). The increase in correlationwith age between central performance and intelligence perhaps indicates that ability to use taskrelevant cues is a significant component of intelligence at older age levels only. Youngerchildren'sperformancemight be less determinedby attentionfocusing ability and more by other factors,such as memoryspan. To test the interchangeabilityof the two conceptual classes of pictures condition and on the used, the performanceson the two-pictures-per-card incidental condition were analyzed separately according to which class of pictures (animals or household objects) was central and which incidental. No differenceswere found for performanceon the central task or incidental task (F < 1.00). Thus the stimulus materials utilized did overcome the possible objection to the earlier study that the figure-groundnature of the stimuli used was generatingthe obtained interactionbetween task measures and distraction.
DISCUSSION

The first questions to be asked concern the findings in light of the initial expectations. The two most important questions have to do with (a) the possible interactions between age level and effects of distraction; and (b) the possible reciprocity or "trading"of performanceon one task, the incidental, for maintenanceof performanceon the other, central or setrelevanttask. Regarding the first question, the results are remarkablysimilar to the 692

JOHN W. HAGEN results of the early study (Maccoby & Hagen, 1965). Although the auditory distractor did have a highly significant effect on central memory performance, the magnitude of the effect by no means diminished at the higher age levels. Further evidence for this lack of interactionwas obtained in this study by the results of the one-picture-per-cardcondition. Also, it was again found that, overall, performanceon the incidental memory task was not impaired by the auditory distractor.The task by distractioninteraction was strongerin this study than in the previous study. In that study, errors were increasedby 26 per cent on the central task and by 15 per cent on the incidental task. In this study, errors on the two-picture-per-card task were 20 per cent greater under distraction conditions, but no overall increase at all in errors occurred under distraction for the incidental memory task scores. Hence, further support is found for the hypothesis that two or more task sets which are made explicit interfere with one another (such as the two sets given to the Ss in this study under distractionconditions) but that processing of irrelevant information is in some way independent of taskset tasks. The question of the selectivity of attention to maintain central task performancemust still be dealt with. The original argumentwas that older children should have more facility in sorting out task-relevantinformation from task-irrelevantinformation.Under conditions of maximal information load, the older children should be able to focus attentionbetter on the tasks at hand and be less affected by distracting stimuli. It now seems more reasonable to conceive of the amount of information processing required by the distractiontask as being constant across age levels. The remaining channel processed the informationfor both central and incidental memory tasks; and this processing was considerablymore efficient for the older Ss as compared to the younger, as shown by several of the experimental findings. First, the performanceon the central task did increase significantly with age, regardlessof the distraction.Second, incidental learning did occur at all age levels. It would certainlynot be reasonableto label all incidental learning as inefficiency in attention focusing. Perhaps, for some age levels at least, the same cues which allow for "incidentallearning"also facilitate central task performance. Support is offered in the positive correlations obtained between the two response measures at the younger age levels. Wohlwill's (1962) notion of the necessity of more redundancy in the stimulus situation for younger children as compared to older is also consistent with these findings. The negative correlationsfound between central and incidental task scores at the oldest age level supportthe hypothesisthat older children are able to ignore more task-irrelevantinformationthan are younger children, perhaps in order to perform better on the central task. The decline found in incidental recall in the earlier study at the oldest age level and the decline found for the same group under distractionconditionsin this study lend furthersupport. 693

CHILDDEVELOPMENT Although we cannot specify exactly why informationprocessing was found to be more efficient with increasing age, two reasonable alternatives can be offered. With increasing age the need for stimulus redundancy decreases, and with this decrease the room left in the channel increases, allowing a higher rate of informationprocessing. Or, with increasing age the ability to focus attention on task-relevantcues increases. Older children can more quickly separate out relevant from irrelevantinformation.Perhaps both alternativeshold; they are not mutually exclusive. However, the evidence is correlative, and neither explanation can be definitively invoked. Follow-up studies are in progress to shed further light on this provoking problem. REFERENCES
Broadbent, D. E. Perception and communication. New York: Pergamon, 1958. Maccoby, Eleanor E., & Hagen, J. W. Effects of distraction upon central versus incidental recall: developmental trends. Journal of experimental child Psychology, 1965, 2, 280-289. Winer, B. J. Statistical principles in experimental design. New York: McGrawHill, 1962. Wohlwill, J. From perception to inference: a dimension of cognitive development. Monographs of the Society for Research in Child Development, 1962, 27, No. 2 (Whole No. 83), 87-107.

694