Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The emergence of class concept formation in preschool children Fryer, Margo 1974-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
UBC_1974_A8 F79.pdf [ 2.27MB ]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0076802.json
JSON-LD: 1.0076802+ld.json
RDF/XML (Pretty): 1.0076802.xml
RDF/JSON: 1.0076802+rdf.json
Turtle: 1.0076802+rdf-turtle.txt
N-Triples: 1.0076802+rdf-ntriples.txt
Original Record: 1.0076802 +original-record.json
Full Text
1.0076802.txt
Citation
1.0076802.ris

Full Text

THE EMERGENCE OF CLASS CONCEPT FORMATION IN PRESCHOOL CHILDREN BY MARGARET L. FRYER B.A., University of British Columbia, 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of Psychology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1974 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Psychology The University of British Columbia Vancouver 8, Canada Date September 6, 1974 Abstract The ability to classify complex visual forms was studied in three, four, and five year old children. Each subject performed two tasks based on two classes of computer-generated stimuli. The oddity task required the identification of the odd form in a set of three eight-sided polygons. The sequential task required the assignment of each sequentially presented single polygon to one of two classes. No feedback was given. The results revealed a marked developmental change in classification ability occurring between about 4 1/2 and 5 1/2 years of age. The oddity task appeared to be a more sensitive test of class concept formation. Signature of Supervisor TABLE OF CONTENTS Page 1. INTRODUCTION 1 2. METHOD 4 3. RESULTS 10 4. DISCUSSION 25. REFERENCES ...... 27 6. APPENDICES ...... 28 ii List of Tables page Table I. Summary of analysis of variance of number of correct 12 responses, both tasks included. Table II. Mean number of oddity problems correctly solved (per 16 block) by age and by difficulty level. Table III. Proportion of correct responses and most frequently 17 chosen pattern for nine escpecially difficult problems. Table IV. Summary of analysis of variance of number of correct 19 responses on the oddity task. Table V. Mean number of sequential problems correctly solved 21 (per block) by age and by difficulty level. Table VI. Summary of analysis of variance of number of correct 22 responses on the sequential task. Table VII. Proportion of correct responses by variability level 24 and by age (on the oddity task). iii List of Figures page Figure 1. Prototypes and sample problems at three difficulty 6 levels. iv Acknowledgement I would like to express my appreciation and gratitude to Tannis M. Williams for the assistance, support, and encouragement given to me at every point from the preparation to the completion of this thesis. V The ability to classify complex visual patterns was studied develop-mentally by Aiken and Williams (1973). Using a task in which subjects were required to identify one odd pattern in sets of three, they found that children in Grade 5 and adults, who did not differ, were both more accurate than children in Grades 1 and 3, who also did not differ. How ever, while the two younger age groups were statistically significantly less accurate than the two older age groups, the absolute differences were small and, the youngest group, Grade 1, still performed signifi cantly above chance levels. These results indicated to Aiken and Williams (1973) that development of the ability to perform their oddity classi fication task occurs sometime prior to age six. In order to assess developmental change of strategy in approach to their classification task, Aiken and Williams (1973) conducted several psychophysical analyses. They found that subjects at all age levels tended to make the same errors and used the same physical features for judging pattern class membership. The data indicated that the two younger age groups were merely slightly less proficient at using the same strategies used by the older groups. The stimuli used by Aiken and Williams (1973) were eight sided polygons generated from each of two prototypes. Psychophysical analyses of their data revealed that subjects at all age levels used both the general similarity between patterns and their prototypes and two parti cular physical pattern features to judge pattern class membership. In other words, neither prototypes nor distinctive features were sufficient to account for performance. This finding contradicts the view of 2 E. J. Gibson (1969) who has argued that the learning of distinctive features is the most important process in perceptual learning and development. To support her argument, she cites work done by Pick (1965) on visual and tactual form discrimination. Pick found that the discovery of distinctive features facilitated transfer of learning more than did the formation of prototypes. She trained kindergarten children to discriminate between standard forms and specified trans formations of them and then made three tests for transfer of learning. The experimental group which was given new standards and transformations which had the same dimensions of difference as those in the training session made fewest errors in the transfer task, indicating that learning depended on discovering the dimensions by which transformations and their standards differed. The group which had the same standards (i.e., prototypes) but new transformations made fewer errors than the group which had both new standards and new transformations. Thus, Gibson (1969) argued that while prototype learning may play a role when reten tion over time is required, distinctive feature learning is the more important process in perceptual learning and development. In discussing the roles played by distinctive features and pro totypes in perceptual development, it is important to note the type of perceptual task under consideration. When the task requires the de tection of what is different among stimuli, as in discrimination, it is reasonable to assume that distinctive features will be more useful. On the other hand, when the task requires the detection of what Is common among stimuli, as in classification, it seems reasonable to 3 assume that prototypes will be more useful. Pick (1965) used a dis crimination matching task; Aiken and Williams (1973) used a classification task. It is therefore not surprising that the results of the former study indicated that subjects used distinctive features more than they used prototypes. While the subjects in the Aiken and Williams (1973) study did use distinctive features, they selected pattern-features which were unrelated to class memberships. Indeed, reliance on these features misled subjects at all age levels on certain problems. Thus, classification accuracy was primarily due to the use of prototype information rather than to the use of distinctive pattern features. The present study was designed to answer some of the questions left unanswered by Aiken and Williams (1973). Of primary interest was the question of when in development (prior to age six) the ability to classify complex visual patterns first occurs. Accordingly, the oddity task employed by Aiken and Williams (1973) was given to children 3, 4, and 5 years of age. In addition, the forms used in the oddity task were presented sequentially to the same subjects. While the oddity task is based on the assumption that accurate performance requires the assignment of each form to one class or the other, it is possible that subjects need only discriminate the odd form in any group of three. The sequential task was included in the present study because it provides a more stringent test of the prototype use in the classification of complex visual forms. Direct comparison of the stimuli is impossible and therefore accurate performance must reflect pattern class learning. A sequential classification task using the same pattern classes 4 employed in the oddity task was used successfully with adults by Aiken and Brown (1971). Method Subjects Subjects for the study were three, four and five year old children attending eight day care centres in metropolitan Vancouver, British Columbia. Half of the twelve children in each age group were female and half were male. Subjects were selected so that at least four months had passed since their last birthday in order to make the age groups more homogeneous and therefore avoid obscuring developmental age-related changes. The mean chronological ages of the three age groups were 3 years, 8 months (S.D. 2.22 months), 4 years, 8 months (S.D. 2.27 months), and 5 years, 7 months (S.D. 3.11 months), respectively. All subjects were tested during June and July, with all but 8 of the 72 testing ses sions occurring during the morning. Twenty potential subjects were eliminated for a variety of reasons. Five subjects were absent for their second testing session (one 3 year old, two 4 year olds, and two 5 year olds). One 4 year old had to be eliminated because of pro jector malfunction. Fourteen subjects were eliminated for inattention or inappropriate responses to the sequential task. Of these, five were three year olds, four were 4 year olds and five were 5 year olds. Be cause the number of subjects at each age level eliminated from the study due to their inability to perform the task was comparable across age levels, age-related results were not likely to be due to subject 5 selection. Stimuli The stimuli were eight-sided polygons computer generated from two prototypes to form two classes of patterns. Within each class, patterns were generated at three levels of similarity to their prototypes and thus three levels of classification difficulty: low, moderate and high. Patterns most similar to their own prototypes are easiest to distinguish from patterns of the other class. The procedure for gene rating the patterns has been described in detail by Aiken and Brown (1971). The overall principle is one of producing random changes in each of the prototype vertices. Oddity task problems consisted of three patterns placed horizon tally on a 4" x 6" card. Examples of problems are shown in Figure 1. On each problem, two patterns were from one prototype class and one was from the other class, but all were of the same degree of simi larity to their prototype and thus were all of low, moderate, or high difficulty. There were 36 problems in total, selected from the 63 problems used by Aiken and Williams (1973) so as to include 12 at each difficulty level. The correct pattern occurred equally often in each position and equally often from each prototype class. No more than two problems of the same difficulty level occurred in sequence. The correct pattern occurred in the same position and was from the same class no more than three times in a sequence. The problems were arranged in three blocks of 12 trials each with four problems of each difficulty level in each block. The blocks were presented in the 6 PROTOTYPES PROBLEMS low Difficulty 4 4 ^ MOOEXATi tuffICUITY HIGH DlffiCUUY V w 4 Figure 1. Prototypes and sample problems at three difficulty levels. [On each problem the subject's task is to choose the pattern that is odd or different (e.g., Patterns 3, 3, and 2, respectively). Subjects were never shown the prototypes.] 7 three orders of a balanced Latin Square. In each age group, two males and two females were assigned to each block order. In each case, the first block was presented again at the end of the task as a measure of improvement over time. Aiken and Williams (1973) had found that 15 (approximately 25%) of their 63 problems were especially difficult; less than a chance proportion of subjects at one or more age levels got each of these 15 problems correct. In order to maintain a com parable proportion of difficult problems, 9 of the 15 problems found by Aiken and Williams (1973) to be especially difficult were included in the present study, with three occurring in each block of 12 patterns. Patterns for the sequential tasks were eight sided polygons from the stimulus samples used in the oddity task. In the sequential task, each problem consisted of a single pattern on a transparent slide. As in the oddity task, there were 36 problems arranged in 3 blocks of 12 problems each. Each block contained four problems from each dif ficulty level and no more than two problems of the same difficulty level occurred in sequence. Within any block, variations on each pro totype occurred equally but no more than three times in a row. As in the oddity task, the blocks were presented in the three orders of a ft balanced Latin Square and the first block was repeated after the third, making a total of 48 trials. Each child received the same block order for both the sequential and the oddity task. Task presentation order was counterbalanced, with half the females and half the males at each age level receiving each task first. The tasks were presented in separate sessions for each 8 child. Time between sessions ranged from three to eight days, with a mean of 6.10 days. Procedure All children were tested by the same woman in a quiet room away from the rest of the day care centre. Oddity Task. The child was seated opposite the experimenter and given the following instructions. I am going to show you some cards. There are three shapes on each of these cards. Two of these shapes belong to one family and one of the shapes belongs to another, different family. When I show you a card, I want you to look very carefully at the shapes. You will see that two of the shapes go together and one does not go with the others. I want you to point to the one shape that does not go with the others, the one that does not belong. Now, let's look at some cards for practice; (the experimenter showed the child the card — two squares and a circle). Point to the one that does not belong with the others. That's right. These two are the same, they belong together. (Experimenter points to two squares) and this one (points to circle) is different. This is the one that does not belong. The second, third, and fourth training problems consisted of two similar but not identical shapes, such as a square and a rectangle, and one different shape, such as a circle. The fifth training problem con sisted of three polygons similar to the experimental stimuli. After the child had responded to each of the training problems, the experimenter 9 verbalized the solution while pointing to the appropriate shapes. Sequential Task. The sequential task stimuli were back-projected onto a screen, placed on the table about two feet in front of the child. About 6" directly in front of the child was a 6" x 4" x 1/2" black wooden panel. Two plastic circles 1 1/4" diameter were nailed flat on the black panel. The circle on the left was red; the one on the right was blue. The experimenter, who operated the projector with a remote control, gave the following instructions: I am going to show you some shapes. There are two different kinds of shapes. If you see one that looks like this, (experimenter shows first slide, low variability example of one class) I want you to point to this circle here, the red one. Now you point to it. This is what one kind of shape looks like. This (experimenter shows second slide, low variability example of other class) is what the other shapes look like. If you see one that looks like this, you point to this circle, the blue one. Now you point to it. All the shapes you will see look either like this one (second slide) or like this one (first slide). O.K. Let's look at some new ones. (Then, two examples of first family, also low variability, are shown; then two examples of second family, also low variability, are shown.) When each slide is on experimenter says, "Which circle do you think you should point to now?" Mistakes made by children on training trials were corrected. The experimenter says, "No, for this one you should point to the red circle, because this shape looks like the other one where you pointed to the red circle." Each experimental session lasted from 10-20 minutes depending on 10 the speed with which the child made choices. Results Two sets of analyses of variance were performed, one on the data from only the first three blocks (Times 1-3) and one on the data from all four blocks (Times 1-4, including the repeated first block). In each case three analyses were conducted: one on the data for both tasks, one for only the oddity data and one for only the sequential data. Only the results of the analyses of the data from all four blocks (Times 1-4) will be discussed because the differences between the two sets of analyses were not substantial. The strengthening of significant effects that occurred in the Time 1-4 analyses can be attributed to the increased numbers of task items. The results of the analyses of the data from only the first three blocks (Times 1-3) are given in the Appendix. The results of the three types of analyses are presented in order, with the data from the combined analyses being discussed first. Combined Analyses The probability of being correct by chance on any one trial was .33 for the oddity task and .5 for the sequential task. In order to make the data from the two tasks comparable, the number of correct re sponses per block per subject was divided by the chance number correct per block for that task. The resulting values were analyzed in an analysis of variance of Order (oddity first, sequential second, or vice-versa) by Sex by Age (3, 4, 5, years) by Task (oddity, sequential) 11 by Variability Level (low, moderate, high) by Time (trials 1-12, 13-24, 25-36, 37-48). Conservative degrees of freedom were used for all tests of significance involving repeated measures. The results of the analysis are summarized in Table 1. The Age main effect was significant (F£ 24 = 6.61, p_ < .01), with a Newman Keuls analysis revealing that 5-year-olds were significantly more accurate than both 4 and 3 year olds (p_ < .01 in both cases), who did not differ. The significant Task main effect (F1 ^ = 25.60, p_ < .01) reflected greater overall accuracy on the oddity than on the sequential task. Both the Age and Task main effects must be interpreted in conjunc tion with the significant Age by Task interaction (F^ ^ = 5.54, £ < .05). A simple effects analysis revealed that there were significant age dif ferences in performance on the oddity task (F£ 24 = H-24, p_ < .01) but not on the sequential task (F < 1). A Newman Keuls analysis further revealed that on the oddity task, 5 year olds were significantly more accurate than both 4 year olds and 3 year olds (p_ < .01 in both cases), who did not differ. When task differences were examined at the various age levels, simple effects analyses revealed task differences only for the 5 year olds (F^ ^ ~ 31.71, p_ < .01), who were more accurate on the oddity than on the sequential task. The variability main effect was significant (F.. „ = 69.28, 1, iz p_ < .01) with all difficulty levels being differentiated in the expected direction (p_ < .01 for all comparisons in a Newman Keuls analysis). The Age by Variability interaction was also significant (F„ = 12 Table I. Summary of analysis of variance of number of correct responses on both oddity and sequential tasks. Source df ms F Between subjects A (order) 1 .17 < 1 B (sex) 1 5.94 3.28 C (age) 2 11.95 6.61** AB 1 .02 < 1 AC 2 2.13 1.18 BC 2 1.96 1.08 ABC 2 1.81 1.00 Ss with grps. 24 1.81 Within subjects D (Task) 1 11.48 25.60** AD 1 .26 < 1 BD 1 .40 < 1 CD 2 2.48 5.54* ABD 1 .32 < 1 ACD 2 2.05 4.57 BCD 2 .18 < 1 ABCD 2 .88 1.96 D x Ss 24 .45 F (variability) 2 20.55 69.28** AF 2 .15 < 1 BF 2 .45 1.52 CF 4 1.26 4.25* ABF 2 .74 2.50 ACF 4 .34 1.16 BCF 4 .91 3.08 ABCF 4 .03 < 1 F x Ss 48 .30 G (time) 3 .07 < 1 AG 3 .01 < 1 BG 3 .25 < 1 CG 6 .38 1.01 ABG 3 .90 2.43 ACG 6 .42 1.13 BCG 6 .40 1.07 ABCG 6 .29 < 1 G x Ss 72 .37 13 Table I (Cont'd) df ms F DF 2 7.20 23.53* ADF 2 .46 1.51 BDF 2 ...17 < 1. CDF 4 .74 2.45 ABDF 2 .55 1.79 ACDF 4 .14 < 1 BCDF 4 .41 1.34 ABCDF 4 .27 < 1 DF x Ss 48 .31 DG 3 .61 1.83 ADG 3 .58 1.76 BDG 3 .19 < 1 CDG 6 .16 < 1 ABDG 3 .45 1.35 ACDG 6 .28 < 1 BCDG 6 .06 < 1 ABCDG 6 .05 < 1 DG x Ss 72 .33 FG 6 .21 < 1 AFG 6 .13 < 1 BFG 6 .37 1.01 CFG 12 .36 < 1 ABFG 6 .12 < 1 ACFG 12 .23 < 1 BCFG 12 .30 < 1 ABCFG 12 .30 < 1 FG x Ss 144 .37 DFG 6 .14 < 1 ADFG 6 .34 1.25 BDFG 6 .13 < 1 CDFG 12 .28 1.01 ABDFG 6 .30 1.10 ACDFG 12 .31 1.14 BCDFG 12 .25 .90 ABCDFG 12 .14 .51 DFG x Ss 144 .27 Note: Conservative degrees of freedom were used for tests of all effects involving repeated measures. * indicates p_ < .05; ** indicates p <.01. 14 4.25, p_ < .05). A simple effects analysis revealed that there were age differences in accuracy on low (F^ 24 = H-24, £ < .01) and moderate (F2 24 = 6.05, £ < .01) but not on high difficulty problems. Newman Keuls analyses indicated that on both low and moderate variability problems, 5 year olds were significantly more accurate than both 4 and 3 year olds (p_ < .01 in both cases), who did not differ. Simple effects analyses also revealed that there were significant differences in ac curacy due to pattern variability at all age levels (for 5 year olds, F2 48 = 50'90» £• < •01> for 4 year olds» F2 48 = 16-30» P_ < -01; for 3 year olds, F2 ^g = 10.66, £ < .01). At all age levels, accuracy on low variability items was significantly greater than on both high (£ < .01 for all age groups) and moderate variability items (£ < .01 for 5 and 4 year olds and £ < .05 for 3 year olds). Furthermore, at all age levels performance on moderate items was significantly more accurate than on high variability items (£ < .01 for 5 and 4 year olds and £ < .05 for 3 year olds). Thus, at all age levels, all variability levels were significantly differentiated. The Task by Variability interaction was significant (F^ 24 = 23.53, £ < .01), with a simple effects analysis revealing that performance across pattern variability levels differed significantly on both tasks (for the oddity task, F2 ^g = 85.80, £ < .01; for the sequential task, ^2 48 = °.46, P_ < .01). Newman Keuls analyses indicated that on the oddity task, performance on all variability levels was significantly differentiated in the expected directions (£ < .01 for all comparisons). On the sequential task, accuracy was greater on low than on higher 15 variability problems (p < .01) and also greater on moderate than oh high difficulty problems (p_< .05). When performance on the two tasks was examined at each variability level, greater accuracy on the oddity than on the sequential task was found for low (F^ ^ = 66.66, ja < .01) and moderate variability problems (F^ ^ = 6.02, p_ < .05), with no task differences on the high variability problems. None of the effects involving Order, Sex, or Time were significant. Oddity Task Analyses The mean number of oddity problems (per block) correctly solved by age and by difficulty level is shown in Table 2. The performance of 5 year olds was significantly above chance on low (p < .01) and on mode rate (p_ < .05) variability problems but not on high variability problems. Four year olds performed above chance levels on only low variability problems (p_ < .05) and the performance of 3 year olds did not exceed chance at any variability level. The proportion of subjects in each age group correctly answering each of the nine problems found to be especially difficult in the Aiken and Williams (1973) study, and the most popular pattern choice, are given in Table 3 along with the data for the older age groups in the Aiken and Williams (1973) study. Problems that were especially dif ficult for older subjects were apparently also difficult for pre-school children, who tended to make the same wrong choices. Number of correct oddity task responses was examined in an analysis of variance of Order by Sex by Age by Variability by Time. Conservative degrees of freedom were used for all tests of significance involving 16 Table II. Mean number of oddity problems correctly solved (per block) by age and by difficulty level. Variability Level Age Low Moderate High Overall 3 years 2.08 1.72 1.25 1.69 4 years 2.31* 1.78 1.44 1.84 5 years 3.31** 2.22* 1.69 2.41 X 2.57 1.91 1.46 Note: In relation to the chance probability of 1.33 correct per block, * = _p_ < -05; ** = p_ < .01. Table III. Proportion of correct choices and most frequently chosen pattern for nine especially difficult problems. (Data for grades 1, 3, 5, and adult groups are from Aiken and Williams, 1973.) Problem Proportion Correct Age Most Popular Choice Age Number 3; yrs 4 yrs 5 yrs Gr 1 Gr 3 Gr 5 Adult 3 yrs 4 yrs 5 yrs Gr 1 Gr 3 Gr 5 Adult Low Difficulty 1 .25 .50 .58 .44 .48 ' .31 .19* 3 1 1 1 1 3 3 Moderate Difficulty 2 .25 .25 .50 .22* .25 .31 .36 3 3 l,3a 3 3 3 3 3 .17* .25 .25 .15* .02* .11* .38 2 3 3 3 3 3 1 4 .25 .25 .42 .39 .17* .22* .24 2 1 3 2 2 2 2 High Difficulty 5 .42 .17* .17* .10* .08* .18* .17* 2 2 2 2 2 2 2 6 .17* .25 .17* .15* .10* .31 .21* 2 1 1 1 1 1 1 7 .25 .17* .33 .15* .04* .07* .05* 3 1 1 3 3 3 3 8 .17* .08* .17* .05* .06* .13* .24 2 2 2 2 2 2 3 9 .17* .33 .50 .39 .23 .20* .26 2 2 l,2a 2 2 2 2 Note: * indicates choice made by significantly fewer than chance number of subjects (p_ < .05). cl two patterns chosen by equal number of subjects. i— 18 repeated measures. The results of the analysis are summarized in Table 4, and serve to confirm the results reported above for the combined analysis. The Age main effect was significant (F^ 24 = 10.5, p_ < .01), with a Newman Keuls analysis indicating that 5 year olds were signifi cantly more accurate than both 4 year olds (p_ < .01) and 3 year olds (p_ < .01), who did not differ. The Variability main effect was also significant (F^ ^ = 69.55, p_ < .01) with all difficulty levels significantly differentiated from one another in the expected direction (p_ < .01 in all comparisons). The only significant interaction was Age by Variability (F£ 24 = 3.78, p_ < .05). A simple effects analysis revealed that the age groups differed on low (F2 24 = 17.72, p_ < .01) and moderate (F2 24 = ^.12, p_ < .05) but not on high difficulty problems. Newman Keuls analyses revealed that in both low and moderate difficulty problems, 5 year olds were significantly more accurate than both 4 and 3 year olds (p_ < .01 in all cases), who did not differ. When the effects of variability were examined at each age level, simple effects analyses revealed that variability affected performance at all age levels (for 5 year olds, F2 48 = 47'83> P. < *015 for 4 year °lds> F2 48 = 18«31» £ < '01'> and for 3 year olds, F2 = 10.98, p_ < .01). Newman Keuls analyses indicated that the accuracy of both the 5 and 4 year age groups was significantly greater on low than on both moderate and high variability problems (p_ < .01 for all comparisons). The 3 year age group was more accurate on both low and moderate than on high difficulty problems 19 Table IV. Summary of analysis of variance of number of correct responses on the oddity task. Source df ms F Between subjects A (order) 1 .01 < 1 B (task) 1 8.33 3.93 C (age) 2 22.40 10.57** AB 1 .15 < 1 AC 2 3.72 1.76 BC 2 .92 < 1 ABC 2 1.09 < 1 Ss within groups 24 2.12 Within subjects F (variability) 2 45.65 69.56** AF 2 .56 < 1 BF 2 .13 < 1 CF 4 2.48 3.78* ABF 2 2.24 3.41 ACF 4 .77 1.18 BCF 4 1.58 2.40 ABCF 4 .30 < 1 F x Ss 48 .66 G (time) 3 .84 1.05 AG 3 .52 < 1 BG 3 .38 < 1 CG 6 .50 < 1 ABG 3 1.50 1.86 ACG 6 1.06 1.32 BCG 6 .62 < 1 ABCG 6 .24 < 1 G x Ss 72 .81 FG 6 .49 < 1 AFG 6 .56 < 1 BFG 6 .80 < 1 CFG 12 .82 1.02 ABFG 6 .40 < 1 ACFG 12 .67 < 1 BCFG 12 .72 < 1 ABCFG 12 .44 < 1 FG x Ss 144 .80 Note: Conservative degrees of freedom were used for tests of all effects involving repeated measures. * indicates p_ < .05; ** indicates p < .01. 20 (p_ < .01 in both cases). Sequential Task Analyses The mean number of sequential problems solved by age and by dif ficulty level is shown in Table 5. The performance of the 5 year olds was significantly above chance on low (p_ < .05) and on moderate (p_ < .05) but not on high variability items. The performance of the 4 and 3 year age groups did not exceed chance at any variability level. The number of correct sequential task classifications was examined in an analysis of variance of Order by Sex by Age by Variability by Time. The results of the analysis are summarized in Table 6. As with the oddity task analysis, the results serve to confirm the findings described above for the combined analysis. Only the Variability main effect was significant (F „, = 8.40, -L , i. H p_ < .01). A Newman Keuls analysis revealed that performance on both low and moderate variability problems was significantly better than perfor mance on high variability problems (p_ < .01 for both comparisons). Discussion The results of the present study provide convincing evidence that the ability to classify complex visual patterns does indeed develop prior to six years of age. A marked developmental change in accuracy takes place between the ages of about 4 1/2 and 5 1/2 years. Evidence for this developmental shift can be found in the significant age main effects, the age-related interactions and the data concerned with per formance relative to chance levels. The performance of the 3 year age 21 Table V. Mean number of sequential problems (per block) correctly solved by age and by difficulty level (maximum possible = four). Variability Level Age Low Moderate High Overall 3 2.52 2.21 2.23 2.32 4 2.52 2.44 2.23 2.40 5 2.96* 3.04* 2.21 2.74 X 2.67 2.56 2.22 Note: In relation to the chance probability of 2.0 correct per block, ** indicates p < .01; * indicates p < .05. Table VI. Summary of analysis of variance of number of correct responses on the sequential task. Source df. ms F Between subjects A (order) 1 1.69 < 1 B (sex) 1 6.50 1.53 C (age) 2 7.09 1.67 AB 1 1.02 < 1 AC 2 8.31 1.96 BC 2 6.47 1.53 ABC 2 8.31 1.96 S_s within groups 24 4.24 Within subjects F (variability) 2 7.78 8.40** AF 2 1.17 1.27 BF 2 2.20 2.37 CF 4 2.44 2.64 ABF 2 .09 < 1 ACF 4 .17 < 1 BCF 4 1.73 1.87 ABCF 4 .51 < 1 F x Ss 48 .93 G (time) 3 .82 < 1 AG 3 1.19 1.21 BG 3 .90 < 1 CG 6 1.02 1.03 ABG 3 2.01 2.03 ACG 6 .42 < 1 BCG 6 .42 < 1 ABCG 6 .81 < 1 G x Ss 72 .99 FG 6 .26 < 1 AFG 6 .61 < 1 BFG 6 .29 < 1 CFG 12 .69 < 1 ABFG 6- .80 1.03 ACFG 12 .67 < 1 BCFG 12 .55 < 1 ABCFG 12 .74 < 1 FG x Ss 144 .77 Note: Conservative degrees of freedom were used for tests of all effects involving repeated measures. * indicates p < .05; ** indicates p < .01. 23 group was not above chance on either task, the 4 year age group performed above chance on only low variability items in the oddity task, and the five year age group performed significantly above chance levels on both low and moderate variability items on both tasks. This developmental change in classification ability can also be clearly seen in Table 7, which gives the proportion of correct responses on the oddity task at each variability level achieved by 3, 4, and 5 year olds in the present study along with the proportion of correct responses achieved by adults and 6, 8, and 10 year olds in the Aiken and Williams (1973) study. The sequential task was included in the present study because it provides a more stringent test of classification skills and the use of prototypes in classification than does the oddity task. The subject sees the prototypes themselves in neither task, but accurate performance in the oddity task might be possible merely through discrimination of the odd pattern in each set of three, rather than through prototype learning. In the sequential task, the subject must classify each pat tern merely on the basis of his past experience with instances of each class. That performance relative to chance levels was comparable on the two tasks in terms of the skills being tapped lies in the absence of significant effects involving order of task presentation. If prototype learning occurred differentially in the two tasks, then performance would be likely to vary with order of task presentation, since the prototypes on which the classes were based were the same in both cases. It is not surprising that the sequential task was in general more difficult than the oddity task. Indeed, it is more surprising that 5 year olds were 24 Table VII. Proportion of correct responses on the oddity task by variability level and by age. Variability Age Low Moderate High 3 yr .5104 .4323 .3177 4 yr .5729 .4427 .3229 5 yr .8177 .5729 .4167 6 yr .8240 .5643 .4476 8 yr .8264 .5644 .4541 10 yr .8960 .6041 .4930 adult .9035 .6398 .5038 Note: Data for ages 6, 8, 10 yrs and adults are from Aiken and Williams, 1973. These proportions are based on sixty-three problems. Proportions from the present study are based on forty-eight problems. 25 able to perform at above chance levels on the sequential task, which requires pattern comparisons from memory. In summary, the results pro vide strong evidence that the oddity and sequential tasks assess the same underlying perceptual and/or cognitive processes. While the absence of significant effects involving order and sex was expected, the absence of significant effects involving time is more thought-provoking. While it is possible that with a larger sample of problems (e.g., as used by Aiken and Williams, 1973) performance might improve significantly over time, the results suggest that learning the nature of the classes, including whatever prototype learning took place, occurred relatively early in both tasks. It is also likely, however, that the very young subjects tired as the trials progressed, and that improvement in classification ability and fatigue effects tended to cancel one another. Although the present study did not include psychophysical analyses relating prototype and distinctive feature measures to classification accuracy, the finding that 5 year olds in the present study were so similar in accuracy to the 6 and 8 year olds in the Aiken and Williams (1973) study suggests that they did indeed use the same bases of judg ment. Further evidence that subjects in the present study used strategies similar to those used by older subjects exists in their comparably poor performance in the nine particularly difficult problems, and the fact that they tended to make the same wrong choices (see Table 7). These results clearly indicate that mastery of the ability to classify complex visual patterns occurs between 4 1/2 and 5 1/2 years 26 of age. The question to be answered now is why 3 and 4 year olds are unable to perform the task successfully. It is tempting to hypothesize that this change in classification ability is related to advancement from the preoperatory to the concrete operatory stage of development, but confirmation would have to come in a study demonstrating that subjects who are more accurate on the classification task are more ad vanced in terms of Piaget's developmental stages. 27 References Aiken, L. S. and Brown, D. R. A feature utilization analysis of the perception of pattern class structure. Perception and Psychophysics, 1971, 9, 279-283. Aiken, L. S. and Williams, T. M. A developmental study of schematic concept formation. Developmental Psychology, 1973, 8_, 162-167. Gibson, E. J. Principles of perceptual learning and development. New York: Appleton-Century-Crofts, 1969. Pick, Anne D. Improvement of visual and tactual form discrimination. Journal of Experimental Psychology, 1965, J59_, 331-339. 28 APPENDIX I: Raw Data , Number of Correct choices (out of a possible 4) Oddity Task Sequential Task Subject Task Sex Age Low Mod High Low Mod High Number Order (in Time Time (l=odd. 1st) years) Time Time Time Time 2=seq. 1st) 1234 1234 1234 1234 1234 1234 01 1 M 3 2122 2111 1113 1233 2231 2222 02 1 M 3 3344 3322 2232 4444 4442 2423 03 1 M 3 2130 1222 2122 2204 1132 2222 04 1 F 3 3213 2212 2212 3122 1012 3200 05 1 F 3 3243 2113 1210 3221 3112 2322 06 1 F 3 2130 2112 1010 2312 0112 2121 07 2 M 3 1423 2311 3101 2323 0203 2332 08 2 M 3 3200 2211 1010 4241 3442 3211 09 2 M 3 3323 1221 1302 2112 3433 1234 10 2 F 3 2222 1223 2112 2344 0242 3242 11 2 F 3 2010 1121 1012 2123 3111 2322 12 2 F 3 2203 2322 2100 4444 4444 4324 13 1 M 4 4410 2113 2222 4444 4434 2432 14 1 M 4 2213 2222 0220 0331 2242 3212 15 1 M 4 1123 2212 0111 3333 2131 1122 16 1 F 4 2222 0231 2101 1222 0420 3221 17 1 F 4 2211 2011 2320 2222 2132 1233 18 1 F 4 1322 2122 0211 3202 4122 3112 19 2 M 4 4444 3234 1130 4344 4234 3332 20 2 M 4 3104 2201 0220 3114 2222 3142 21 2 M 4 4444 3331 2212 4344 3244 3224 22 2 F 4 1432 2221 2131 2231 1332 1131 23 2 F 4 2121 2021 1222 3212 3242 1232 24 2 F 4 1331 2112 0110 2231 2123 4341 25 1 M 5 3434 3243 3232 4434 4444 2243 26 1 M 5 2443 2233 1222 2323 4242 0213 27 1 M 5 4233 2213 2211 2223 3313 2232 28 1 F 5 2443 2222 1231 4433 3432 1222 29 1 F 5 4442 2231 1201 4444 3333 4123 30 1 F ' 5 3434 3213 3113 4444 4424 3333 31 2 M 5 3414 4321 2221 1234 2334 2222 32 2 M 5 4334 2232 1321 4444 3444 2332 33 2 M 5 4422 2224 2112 1112 4214 1122 34 2 F 5 4433 2213 1122 2332 4231 2222 35 2 F 5 4243 1122 3121 1232 2333 1312 36 2 F 5 3333 4323 1202 3244 3224 2234 APPENDIX 2: Summary of analysis of variance of number of correct choices, both tasks included, for time 1-3 (problems 1-12; 13-24; 25-36) only. Source df ms F Between subjects A (order) 1 .11 < 1 B (sex)  3.24 2.17 C (age) 2 7.46 4.98* AB 1 .00 < 1 AC  1.88 1.26 BC 2 1.08 < 1 ABC  .83 < 1 Ss within groups 24 1.50 Within subjects D (task) 1 10.75 27.88** AD 1 .07 < 1 BD  ° .55 1.42 CD 2 2.16 5.60* ABD 1 .80 2.08 ACD  1.65 4.27 BCD 2 .22 < 1 ; ABCD  .60 1.57 D x Ss 4 .39 F (variability) 2 14.00 40.84** AF 2 .14 < 1 BF  .57 1.66 CF 4 1.16 3.39 ABF 2 .35 1.01 ACF  .37 1.07 BCF 4 .73 2.13 ABCF  .11 < 1 F x Ss 8 .34 G (time) 2 .11 < 1 AG 2 .01 < 1 BG  .18 < 1 CG 4 .27 < 1 ABG 2 1.32 3.49 ACG  .60 1.58 BCG 4 .35 < 1 ABCG  .19 < 1 G x Ss 8 .38 30 APPENDIX 2 (Cont'd) Source df ms F DF 2 5.92 17.79** ADF  .40 1.21 BDF  .07 < 1 CDF 4 .66 1.98 ABDF 2 .19 < 1 ACDF  .12 < 1 BCDF 4 .29 < 1 ABCDF  .11 < 1 DF x Ss 48 .33 DG 2 .67 2.00 ADG  .81 2.41 BDG  .21 < 1 CDG 4 .21 < 1 ABDG 2 .34 1.01 ACDG  .42 1.24 BCDG 4 .04 < 1 ABCDG  .05 < 1 DF x S_s 48 .34 FG 4 .20 < 1 AFG  .16 < 1 BFG  .31 < 1 CFG 8 .46 1.33 ABFG 4 .07 < 1 ACFG  .27 < 1 BCFG 8 .16 < 1 ABCFG  .13 < 1 FG x Ss 96 .34 DFG 4 .06 < 1 ADFG  .29 1.24 BDFG  .14 < 1 CDFG 8 .32 1.37 ABDFG 4 .24 1.03 ACDFG  .43 1.84 BCDFG 8 .33 1.41 ABCDFG  .13 < 1 DFG x Ss 96 .23 Note: Conservative degrees of freedom were used for all tests involving repeated measures. * indicates p_ < .05; ** indicates p_ < .01. 31 APPENDIX 3: Summary of analysis of variance of number of correct choices on the oddity task only, for time 1-3 only. Source Between subjects A (order) B (sex) C (age) AB AC BC ABC S_s within groups Within subjects F (variability) AF BF CF ABF ACF BCF ABCF F x Ss G (time) AG BG CG ABG ACG BCG ABCG G x Ss FG AFG BFG CFG ABFG ACFG BCFG ABCFG FG x Ss df 1 1 2 1 2 2 2 24 2 2 2 4 2 4 4 4 48 2 2 2 4 2 4 4 4 48 4 4 4 8 4 8 8 8 96 ms .00 5.71 15.58 . .69 3.11 .30 .45 1.72 33.19 .60 .23 2.09 .73 .65 .96 .27 .70 1.00 .74 .54 .51 1.75 1.55 .49 .26 .85 .40 .56 .73 .99 .20 .95 .67 .39 .66 < < 1 3.31 9.04** 1 1.81 1 1 < < < < < 47.47** 1 1 2.98 1.05 1 1.37 1 1.18 1 1 1 2.05 1.82 1 1 1 1 1. 1. 1 1. 10 49 44 1.02 1 Note: Conservative degrees of freedom were used for all tests involving repeated measures. * indicates p_ < .05; ** indicates p_ < .01. 32 APPENDIX 4: Summary of analysis of variance of number of correct choices on the sequential task only, for time 1-3 only. Source df ms F Between subjects A (order) 1 .69 < 1 B (sex) 1 2.25 < 1 C (age) 2 3.26 < 1 AB 1 1.63 < 1 AC 2 7.06 1.94 BC 2 4.51 1.24 ABC 2 4.71 1.29 Ss within groups 24 3.64 Within subjects F (variability) 2 4.63 4.13 AF 2 .84 < 1 BF 2 2.06 1.84 CF 4 2.57 2.29 ABF 2 .48 < 1 ACF 4 .49 < 1 BCF 4 1.91 1.70 ABCF 4 .28 < 1 F x Ss 48 1.12 G (time) 2 .85 < 1 AG 2 1.62 1.72 BG 2 .34 < 1 CG 4 .76 < 1 ABG 2 2.69 2.86 ACG 4 .57 < 1 BCG 4 .46 < 1 ABCG 4 .37 < 1 G x Ss 48 .94 FG 4 .12 < 1 AFG 4 .53 < 1 BFG 4 .14 < 1 CFG 8 .89 1.09 ABFG 4 .80 < 1 ACFG 8 .64 < 1 BCFG 8 .43 < 1 ABCFG 8 .17 < 1 FG x Ss 96 .82 Note: Conservative degrees of freedom were used for , involving repeated measures. ** indicates p_ < .01 * indicates .05; 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Country Views Downloads
United States 11 5
China 6 0
Canada 2 0
Russia 1 0
City Views Downloads
Ashburn 8 0
Shenzhen 3 0
Beijing 3 0
Seattle 2 0
Longueuil 2 0
Saint Petersburg 1 0
Redwood City 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0076802/manifest

Comment

Related Items