THE EFFECI OF AGE ON.ACHROMATIC AND CHROMATIC DISCRIMINATIVE LEARNING By Ieland.Themee gliffcrd A.THESIS Submitted to the College of Science and Arts Michigan State University'ot Agriculture and Applied Science in.partial fulfillment of the requirements for the degree of ‘MASTER OF ARTS Department of Peyehology 1956 .35 as ‘3 'ACKNOWLEDGMENTS This study was originally conceived under the direction or Dr. Allen D. Calvin, formerly of Michigan State University and now of Rollins College. Since Dr. Calvin's absence, Dr. Alfred G. Dietse has been the imediate advisor. The author is deeply indebted to both of these men for their encouragement, helpful criticism, and generous aid. All data were collected at the Marble School, East Lansing. Sincere thanks go to Mr. Donald'W. Bath, principal of Marble School,'whose cOOperation made this possible. CONTENTS I INTRODUCTION ............................... 1 II METHOD Subjects ............................... Apparatnl............................... Procedure............................... III MLTSOOOOOO..000OOOOOOOOOOOOCOOOOOOOOOOOO. «10‘er T‘bl. I coco.cocccoecoccececccocceeeccoc Iv mmSSIONOCOOOOOOOOOOOOOO0.0.0.0...00...... 8 v SWIOOOOOOOOOOOQOOOOOOOOOOOOOOOO0.00....O 13 BIBLImR-APHIOOOOOOO00.0.0000...OOOOOCOOOOOOOOOOOO m INTRODUCTION A number of investigators have studied the relative effects of different kinds of visual stimuli on discrim- inative learning. Research in this area with sub-human primates has shown, for example, that planometric'and stereonetric cues are equally effective (9), that horizon- tal and vertical cues are equally effective (9), and that problems with stimulus objects are easier than those with patterns (9, 10, ll, 12, 19). Also from several studies it has been established that color and form are more easily distinguished than size and number, although whether color is more facilitative then form or not is as yet unresolved (6, 7, ll, 12, 16, 17, 18, 19). Calvin and his associates (1;, 5) have recently offered evidence from work with children that is not in accord with what might be'expected from the work with sub-human primates. From the various stimulus dimensions investi- gated, i. e., color, fern, size, achromatic brightness, brightness within the sane hue, and pattern vs. no pattern, they consistently found learning to be most difficult she: the cue stimuli differed in color. In one of the studies (5) the color problen consisted of a blue vs. green discrimination and in the second (1;) it consisted of a green vs. red discrimination. To account for this result it was suggested (5) that the children might be utilising conc @ts which, while highly adaptive in the majority of situations, inhibit learning when the discrimination nust be on the basis of color differences. Thus in a problem involving cues differing in size, the concepts "large" and "small" would facilitate learning, but in a problal involving cues differing in color, the concept "colored" wouldact as a handicap. Even though clearly differmt, as long as the ones were conceptualised as "colored" they would remain inseparable for purposes of solving the problm. The following experiment was designed to extmd the stucv of the relative difficulty of chromatic and achro- matic discriminative learning by including children from the kindergarten through the fifth grade. MEI'IDD Subjects: The _S_s were 120 elementary school children, 21; chosen from each of five grade levels. Within each level, half of the S3 were tested on a black vs. white discrimination and half on a blue vs. green discrimination. This gave the following ten experinmtal groups: 0A, kindergarten-achromatic; OC, kindergarten-chronatic; 1A, first grade-achromatic; 10, first grade-chromatic; 2A, second grade-achromatic; 20, second grade-chromatic; 3A, third grade-achromatic; BC, third grade-chromatic; h-SA, fourth & fifth grades-achromatic; h-Ec, fourth & fifth grades-chromatic. ‘1‘th were six boys and six girls in each group and the last two groups had as easy fourth as fifth graders. Of the _S_s who were tested on the blue vs. green discrimination half in each group had the blue card positive and half had the green card positive. A corresponding balance existed in the groups tested on the black vs. white discrimination. Within the restrictions set by the experimental design the §_s were rancbnly assigned to the various groups. The problem for any S was pro-determined by E and unknown to the teachers who selected the children for testing. aratusi The apparatus consisted of a platform 8u deep -3- behind a vertical white cardboard screen. The screen 'was 22‘ wide and 28' high and could be raised and lowered by E, Two orange cups 3' in diameter at the top were placed upside down about h“ apart on the platform and at equal distances from S. A small metal charm was con- cealed under one of the cups and served as the reward. Stimulus cards were placed flat on the platform, one in front of each cup. The four cards were black, white, blue, and green and each was 2%” x3315".1 The apparatus was situated on a small table directly in front of _S_ and concealed.§_fron § except when the screen was in the raised position. Illumination from an overhead light eliminated shadows from the p1atform.area. Procedure: Upon entering the testing room _S_ was seated comfortably at the table and given the following instruc- tions: "lb are going to play a little game. Behind this screen there are two cups and under one of them.there is a toy. If you can guess which cup the toy is under, you can keep the toy.‘' The screen was then raised and the cups and stimulus 1. Approximate Munsell values for the colored cards were 7.5 PB 3/10 and 2.5 G 5/8. -14.. cards exposed to S. If they did not do so spontaneously, the §s was instructed to choose one of the cups by pointing at it. No other for. of response was accepted. then they had pointed, the chosen cup was raised. If the choice was correct, a toy was exposed and immediately handed to S. If the choice was incorrect, the raised cup revealed nothing and _S_ was shown that the reward was under the other cup. The screen was than lowu'ed and the reward and stimulus cards replaced for the next trial according to a pro-determined random order by Gellerman (8). The criterion for learning was ten consecutive correct trials. If this was not met, the test was terminated at the first incorrect choice after trial. 30. All non- solvers were arbitrarily given a score of to. After an _S_ had completed a series of trials, the following questions were asked: "How could you tell which cup the toy was under? How were the cups different? Did you see the cards? What color were they?" can’ A: '1”; -____.‘ ‘W.e-'.a.¢.i_- -_ .. 1". 'J-‘i' ' - " ~. ' .;.<-r- ' .uTJ-Iqwr. 12-? .1 r:- - -"-rg ,~ g... .- . - rah «5'- v1"- 1‘:,_{' “1.. as .3. 5‘ RESULTS Table I shows median.trials to criterion and the percentage of §s in each group who solved their problem. No significant differences between grade levels were obtained for either the achromatic or the chromatic problem. When all §s having the achromatic problem were compared with all §s having the chromatic problem in terms of the number of §s in each group who reached criterion, a Chi-square of 7.5 was obtained. This is significant beyond the .01 level. From their answers to the questions asked after testing all.§s demonstrated ability to distinguish and name the colors on the cards correctly. IABLE I MEDIAN TRIALS TO CRITERION Trials % solving Grade level A C A c o ho ho 1a.? 25 1 12 36.5 75 50 2 22 no 66.7 33.3 3 20.5 Lo 75 25 h-S 38 39.5 50 50 combined 21 no 61.7 36.7 DISCUSSION A number of _S_s in both the groups with the chromatic problem and the groups with the achromatic problem failed to reach the criterion. This suggests the possibility of factors conmon to both problems that are preventing the relatively rapid organization necessary for solution. Recent studies on both lower organisms (l, 2, 21) and children (3) indicate the importance of spatial relations in a two choice discriminative learning situation. They show that placing the differential cues close to each other may prompt a configural organization that inhibits isolation of the directionally orienting properties. For example, if the black card and the white card were organized as a single unit, each trial would be another presentation of a black-white configuration, or an achro- matic configuration, and the ready separation of more useful elements, e.g., black-on-right and black-on-left, would be more difficult. The closer the cards were together the more difficult their separation would be. Other factors common to both problems, e.g., the iden- tical forms of the cards and the continuity of the edges in their placement, may also function to inhibit learning. Another result of this study was that the chromatic problem was significantly more difficult than the achro- -23.. matic problaa. This corroborates earlier work (1;, S) and lends support to the hypothesis put forth by Calvin and Clifford. They hypothesised that a dominant concept "colored" inhibited the differentiation of the stimulus cards necessary for learning. Further evidence for such a concept is found in Koffka. He notes, ”The Sterne report of their daughter that 'at the age of three years and two months Hilda called bright and dark things w_}_1_i_t_e and M3 otherwise she named correctly only the colour 3331. But the accuracy of the word £e_d was obviously quite accidental since _a_ll__ variegated colours were likewise called £29338 As Winch has noted, it often happens that variegated colours are distinguished from neutral tones by giving then all the same name, which indicates that all variegated colours have a common characteristic in contrast to the achromatic tones, and that this comon factor must therefore be much more influential than any differences seen betwem the variegated colours themselves. ' (15, p.285). Werner (20) discusses the general kind of sensory organization proposed here and cites it as the early phase in development to a higher level of "categorical abstraction" where the individual can purposefully shift his point of view in grouping activity. At the earlier level properties of the stimulus objects themselves automatically force their organisation and at the same time inhibit the perception of relations based on other properties. 111a absolute degree to which the children were handi- capped on the blue vs. green problem was not detennined. Had more opportunity been provided the §s who failed to solve may ultimately have done so. Actual inability of some of the _S_s to perceive a difference between the colored cards is ruled out as all of the §s without exception when questioned after testing differentiated and labeled the cards correctly. 7 ‘With reference to experimental data‘werner says that the level of categorical abstraction was never reached unaided before eight years of age and usually not until eleven to twelve years of age. He points out, however, that the evidence for growth of the abstraction process may vary with the experimental conditions. The lack of variation between age levels in the present study indi- cates that the type of abstraction observed is relatively stable from the kindergarten to the fifth grade. That this result is due to less motivation at the higher levels is unlikely. In the author's opinion, the number of toys received rather than their physical characteristics -10.. afforded the most incentive and interest in "who got the most" actually seemed to increase rather than decreasewith age. The congept "colored" is, thm, probably well established before school age. Indeed, Koffka places its establishment within the first year: '... during the first three-quarters of the child's first year of life no configurations of colour arise other than a primitive chromatic-achromatic distinction. ." (15: P028)”. It is likely, than, that learning to discriminate and learn the names of colors develops at an early age from a conceptual dichotomy of achromatic-chromatic. Later in develOpment the concept of achromatic becomes further dichotomised into white-black, light-dark, light- not light, or some similar fresework that is easily placed along a simple linear continuum and readily assim- ilated by the child. The concept of chromatic or "colored", however, is not so easily broken down. Although continua are recognised as existing between say green and blue, the new colors emerge relatively isolated from the general framework of "colored". At no time is there an organiza- tion such as light-dark in the case of the concept "achro- matic” that exists stronger than the superordinate “colored". Under more optimal conditions the concept of un- a single color, e.g., "blue", might predominate, or a relatedness might be seen between two separate colors, e.g., blue and green, but generally the condition is such as to favor the dominance of the general concept "colored" to the suppression of the more atomised subor- dinate concepts. It is important to recognise that prOposing the concept "colored” immediately implies a higher order cognitive franatork apart from the primitive sensory organisation found in lower organisms. This is due to the fact that the results of both the studies by Calvin and associates ()4, 5) were contradictory to those of similar research on sub-human primates (7, 11, 12, 16, 17, 19). Uhile color cues inhibited learning in children, monkeys learned most readin when the cues differed in color. If we may assume that monkeys operate cognitively on a lower level than first grade children, the these data must be due to something in children in addition to or superseding that which is inhermt in monkeys. Other studies such as those by Hunter (13, 1b) which showad conceptual transposition behavior in children and a lack of it in rate also sea to indicate that the process of conceptual organisation in humans is qualitatively differmt from other organisms in the pwlogmetic hierarchy. -12.. SUMMARY This study was designed to examine the relative participation of a previously hypothesized concept "colored! in solving a discrimination problem at succes- sive levels of development. Children from the kindergar- ten through the fifth grade'uere tested either on a black vs. white problem or on a blue vs. green problem. Although no differences in performance were found between grade levels the chromatic problem was signif- icantly more difficult than the achromatic problem. Implications of these results along with those of previous similar studies for both ontogenetic and phylogenetic comparisons were discussed. -13- 1. 2. 3. h. 5. 6. BIBLIOGRAPHY Bitterman, M.E., Tyler, D.W. 8c Elam, C.B. Simultane- ous and successive discrimination under identical stimulating conditions. £63. J. Psychol., 1955, éé, 237-2118. Bitterman, M.E. & Wodinsky, J. Simultaneous and successive discrimination. Eychol. Rev. , 1953, ég, 371-376. Calvin, A.D. Configurational learning in children. J. educ. Psychol., 1955, 116, 117-120. Calvin, d.D., Clancy, J.J. & Fuller, J.B. A further investigation of the relative efficacy of various types of stimulus-objects in discriminative learning by children. (personal comunication) Calvin, A.D. 8: Clifford, L.T. The relative efficacy of various types of stimulus-objects in discriminative learning by children. 522;, g, Pszghol., 1956, g2, 103-106. Chow, K.L. Stimulus-characteristics and rate of learning visual discriminations by experimentally naive monkeys. 52195. J. Psychol., 1953, éé, 278-282. Cole, J. The relative importance of color and form in discrimination learning in monkeys. J. comp. physiol. Pszghol., 1953, g9, 16-18. - 1h - 8. 9. 10. Gellerman, L.W. Chance orders of alternating stimuli in visual discrimination experiments. 1. gang: Psychol” 1933, 324 206-208. I Harlow, H.F. Studies in discrimination learning by monkeys : III. Factors influencing the facility of solution of discrimination problems by rhesus monkeys. J. g_e_r_l_. Psychol., 1916, §_2_, 213-227. Harlow, H.F. Studies in discrimination learning by monkeys : IV. Relative difficulty of discriminations between stimulus-objects and between couparable patterns with homogeneous and with heterogeneous grounds. g, 533. 332921., 191:5, :3, 317-321. Harlow, H.F. Studies in discrimination learning by monkeys: V. Initial performance by experimentally naive monkeys on stimulus-object and pattern discrim- inations. g. gag. Psychol., 1915, 3_3_, 3-10. Harlow, H.F. Studies in discrimination learning by monkeys: VI. Discrimination between stimuli differing in both color and form, only in color, and only in form. J. gag. Psychol., 1915, .3_3_, 225-235. Hunter, I.M.L. An experimental investigation of the absolute and relative theories of transposition behavior in children. £131.. :1. Psychol., 1952, _h_3_, 113-128. -15.. 1h. 15. 16. 18. 19e 20. 21. Hunter, I.H.L. The absolute and relative theories of transposition behavior in rats. J. comp. physiol. Psychol., 1953, 563 1193-1197. Koffka, K. The Growth of the Hi__n_d. New York: Humanities Press, 1928. Warren, J .H. Additivity of cues in visual pattern discrimination by monkeys. 3. camp. physiol. szchol” 1953, gg, hBh-h86. Warren, J .H. Perceptual dominance in discrimination learning by monkeys. g. 9332. pfigiol. Psychol., 1951;, 91, 290-292. Warren, J.H. Some stimulus variables affecting the discrimination of objects by monkeys. g. M. Psychol” 1956, _8_8_, 77-80. Warren, J.H. a Harlow, H.F. Learned discrimination performance by monkeys after prolonged post-operative" ‘ ' recovery from large cortical lesions. _J. Egg: M: Psychol., 1952, .115, 119-126. 1; \ K.\ ‘ \ Werner, H. Comparative Psychology o_f_ @4ng- 92239 Chicago: Follett, 19148. 8‘ Wodinsky, J., Varley, Mai. & Bitterman, H.E. Situa- tional determinants of the relative difficulty of simultaneous and successive discrimination. J. 32152. physio . Psychol., 1951;, {41, 337-340. -16.. 1" ROOM USE (WW M '3‘ Jul 31 ,x'! fl‘r' ’7 _ SE ./ £383! 17% i885 Demco-293 It 03046 5466 I" L“ H H I S“ III! N I. I l 1293 IIHIWIHHIIIHI , 3