Perceptual equivalence between visual and tactual. pattern perception: An anchoring study 1 A' ~ A 2 A 3 ~ A\ 3.

Transcription

1 Perceptal eqialence between isal and tactal. pattern perception: An anchoring stdy 1 STANLY H. BRUMAGHIM, TH BOING OMPANY D. R. BROWN:2 PURDU UNIVRSITY An anchoring design was sed to inestigate correspondence between isal and tactal scaling of complexity when two samples of random polygons sered as stimls forms. The reslts spport the conclsion that within an adstment for differential acity between ision and actie toch inpt is referred to a common perceptal mechanism. A nmber of experimental approaches hae been directed toward demonstrating that inpt from the arios sensory modalities is referred to a common perceptal mechanism. These hae inclded demonstrations of intermodal transfer of learning (e.g., Borkman, Garill, & Molander, 1965), cross-modal matching (Steens, Mack, & Steens, 196), intermodal recognition (ainess & Gibson, 1964; Lobb, 1965), and intermodal anchoring (Behar & Bean, 1961; Goldstone, Boardman, & Lhamon, 1959; Goldstone & Goldfarb, 1963). In earlier stdies (Brown & Brmaghim, 1968; Owen & Brown, 1968), it was emphasized that the stdy of perceptal eqialence raises the qestion of stimls eqialence. If eqialent responses occr to stimlation presented to different modalities, the qestion of what stimls attribtes accont for the eqialence arises. This latter qestion remains almost totally nanswered in iew of the difficlties associated with specifying a perceptallyreleant physics for complex stimls sitations. In iew of progress in deeloping the psychophysical stdy of form perception (Attneae & Arnolt, 1956; Behrman & Brown, 1968; Brown & Andrews, 1968; Brown & Owen, 1967), it is possible, howeer, to systematically stdy sch qestions for modalities in which form perception is meaningfl. In pattern perception positie transfer of learning has been demonstrated between isal and tactal conditions of presentation (e.g., Gaydos, 1956; Semmes, Weinstein, Ghent, & Teber, 1954), intermodal matching has been demonstrated (ainess & Gibson, 1964; Lobb, 1965) and some sccess in relating isal and tactal complexity dgments to a common set of pattern attribtes has been reported (Brown & Brmaghirn, 1968; Owen & Brown, 1968). This stdy extends preios work by demonstrating that cross-modal anchoring effects occr in isal and tactal pattern perception. MTHOD Sbects ighty niersity stdents, 4 males and 4 females, sered as Ss. Of these, 17 sered to satisfy partial reqirements of an introdctory psychology corse and each of 63 was paid $ for participation. Patterns Two samples of 1 random polygons, designated Set A and Set B, were selected from preios research (Owen & Brown, 1968) to be approximately eqidistant on a nidimensional isal complexity continm. Patterns were generated in a 1 by 1 nit matrix according to a modified Method I of Attneae and Arnolt (1956) with a single modification of these rles to assre a specified (4, 8, 12, 16, or 2) nmber of independent sides. All forms were eqated for area at 125 sq nits and were geometrically centered on 1 x 1 mm white posterboard sqares on which they were presented. Forms were ct from black, finely-graded sandpaper (Belir-Manning 36A) and gled to the posterboard sqares. These forms, shown in Fig. I with their complexity scale ales, were sed in a preios stdy (Brown & Brmaghim, 1968) and a detailed discssion of the stimls domain from which they were sampled has been presented elsewhere (Brown & Owen, 1967). Apparats and Procedre The apparats, which has.been described elsewhere (Owen & Brown, 1968), consisted essentially of a Plexiglas display srface for presenting forms tactally and isally, circitry for measring latency, and peripheral eqipment for shielding forms from sight dring tactal presentation and between trials. The experiment was carried ot in two sessionsfor all Ss: in the first session, forms were nidimensionally scaled for complexity and in the second session forms were rescaled in the presence of anchors. In the first session, the 8 Ss were diided into grops (1 males and 1 females) by randomly assigning them to for conditions which were defined by the factorial combination of two leels of form complexity (Low and High) and two modes of presentation (Visal and Tactal). Ten low-complexity forms were obtained by selecting the fie forms from Set A (A}, A 2, A 3 ' A 4, A 6 ) andset B (B}, B 2, B 3, B 4, B 6 ) with the lowest complexity scale ales obtained in a preios stdy (Owen & Brown, 1968). Similarly, the 1 high-complexity forms were selected (A 6-A} and B 6-B} ) from the remainder of the scale with form A 6 and B 6 shared by both high and low-complexity forms. For both isal and tactal presentation, each of the 1 forms was presented three times so that each S receied 3 trials with the order of forms randomized within blocks of 1 trials. For both isal and tactal ratings, Ss were asked to assig., nmbers from "I" (lowest complexity) to "9" (highest complexity) to each form and 1 practice trials with the appropriate forms preceded experimental testing. For grops receiing isal presentations, S was seated across a table from and was told that his rating latency wold be measred bt that was interested in his normal reaction time. On each trial, a crtain was drawn in front of the apparats. When presented a ready light, the crtain was withdrawn, S released a depressed panel to start timing the trial, and the rating spoken into a oice key terminated the timed trial. For tactal presentations, the same procedre was followed except that Ss explored forms behind the crtain which remained drawn. OMPLXITY OMPLXITY SHAP SAL VALUS SHAP StAL VALUS A' A A l8 A A\ 3. A 5 ' ' A 6., It A !I 4.98 A A A lo Fig. 1. The two sets of shapes and their isal complexity scale ales. Perception & Psychophysics, 1968, Vol. 4 (3) opyright1968, PsychonomicJornals, Santa Barbara, alif. 175

2 They were free to explore the forms in any manner which did not reqire moing them from their position in the apparats. In the second session, Ss again rated the same forms they had rated in the flrst session (standard forms), bt in the presence of anchoring forms. ight grops of 1 Ss each (fie male and fie female) were formed by i:liiding the for grops from Session I on the basis of the modality in which anchoring forms were introdced. Both intramodal and intermodal anchors were sed and the standard forms were presented in the same modality as they had been in Session I for each S. These eight grops were defined, then, by the factorial combination of High and Low complexity forms, isal and tactal presentation of standard forms, and isal and tactal presentation ofanchoring forms. The low-complexity forms within each.set were anchored by the fie high-complexity forms in the same set (A or B); the fie high-complexity forms within each set were anchored by the fie low-complexity forms in the same set. With this restriction, 25 trials were defined by combining all possible standards and anchors within Set A and 25 trials were defined by these combinations for Set B. Ss in au grops receied the 5 trials with the order of pair presentations determined randomly. The procedre for Session 2 was similar to that sed in the first session except that dgments were made in the presence of anchors. On trials in which high-complexity forms anchored low-complexity standards, Ss were instrcted to assign "9" to the anchoring form; low-complexity anchors for high-complexity standards were assigned "I" as a scale ale.ss were gien fie practice trials, were told to make their ratings as independently as possible across trials, and no attempt was made to inflence the distribtion of ratings. Standard forms were presented to the left of anchor forms nder all conditions. Under tactal-tactal conditions, Ss explored the left anchor form and then the right standard form with their preferred hand and were free to examine both forms repeatedly. Under tactal anchor-isal standard, isal anchor-tactal standard, and isal anchor-isal standard ao. 7. D > 6 ' " 5. >- -i 4., g V). s 4 A, Low Fig. la. Relation between the isal complexity scale and mean isal ratings for Set A formsscaled isally. A 6 "AAA High ompl.xity )( Gl o 8. 5 Visal A 2 A) A. Fig. 2b. Relation between the isal complexity scale lidli mean isal fttings for Set A formsscaled tactauy. AI, AI> A, As High A 9 A IO Scale Vales conditions, exploration of both the standard and anchor forms occrred simltaneosly. RSULTS The scaling data from Session I were examined in terms of their reliability and in terms of their agreement with preiosly-obtained reslts. Intra-class correlations were sed to estimate reliabilities within the for experimental grops and were fond to be.85,.86,.85, and.54 for low-complexity isal and tactal ratings and high-complexity isal and tactal ratings respectiely. The relatiely low reliability of tactal ratings of complex forms will be discssed below. Mean ratings oer Ss and trials were sed for scale ales and are smmarized for the for conditions in Fig. 2a-2d. In those figres, the abscissa ales are those which were sed from an earlier stdy (Owen & Brown, 1968) to select the shapes. The ordinate ales are those obtained herein. In reading these graphs, it is well to keep in mind that different grops of Ss scaled the low- and high-complexity shapes. learly, both the tactal and isal data are in close linear agreement with the preiosly-obtained reslts. It is also clear that tactal and isal scale ales are ery highly related. Finally, since the shapes are ordered in pairs on the basis of the nmber of independent sides, both tactal and isal complexity ratings are linearly predictable from the nmber of independent sides. Analyses of the transformed latency scores also spported preios reslts. There was a large modality main effect. The isal mean latencies were 1.52 sec and 1.85 sec for low- and high-complexity forms respectiely. The corresponding tactal mean latencies were 14.5 sec and sec. Moreoer, the tactal latencies were linear with a positie slope oer the complexity scale while the isal latencies were qadratic oer the scale with a peak in the middle of the scale and shorter latencies at both scale extremes. 176 Perception & Psychophysics, 1968, Vol. 4 (3)

3 8., D ::;) > D A) ) >-'( D ::;) > B B 3 B 4 low,!j High Visal Scale Vales Fig. 2c. Relation between the isal complexity scale and mean isal ratings for Set B forms scaled isauy. Since the reslts from the first session were reliable and highly similar to preios reslts with these forms, the data from the second session were examined for inter- and intramodal anchoring effects. These data are smmarized in Fig. 3a-3d where each figre smmarizes the data for two of the eight experimental conditions. Anchoring effects are typically described as "assimilation" effects when ratings are shifted toward the ale of anchoring stimli and "contrast" effects when ratings are shifted away from anchors. Fig. 3a-3d make it apparent that both types of effects occrred. In those figres, the shift for an entire set of fie standard forms is shown for each anchor form sed in that condition. The ordinate ales were compted as a per cent of the possible shift from the median of the Session I ratings. Ths, if the median from Session I were 4.5, the anchored median cold shift from 4.5 to I or from 4.5 to 9. The degree of anchoring and its direction were compted as a per cent of this total possible shift nder each condition. When attention is restricted to the effects of high-eomplexity anchors pon the ratings of low-complexity forms (Fig. 3a-3b) an interesting type of asymmetry of effects is apparent. In Fig: 3a, the effects of isal anchors on both isal and tactal ratings and of tactal anchors on tactal ratings are ery similar. When the anchor was similar to the series being rated (low complexity) assimilation occrred, bt as the standard and anchor series became less similar contrast effects occrred. Howeer, when tactal anchors were combined with isal standards assimilation effects occrred. Similarly, for Set B (Fig. 3b) the fnctions (no effects changing to contrast effects) were similar nder all conditions except that tactal anchors again created assimilation effects with isal ratings. The reslts when low-complexity anchors were presented with standards of high complexity (Fig. 3c-3d) were highly consistent nder all conditions. Minor contrast occrred with the most deiant anchor, assimilation effects were obtained otherwise, and the effects of tactal and isal anchors were ery similar. DISUSSION The Session I data are ery clear and spport the general conclsion that complexity dgments are ery similar for isal and tactal conditions of exploration. The difference in shape of the fnctions relating latency to the complexity scale for the two modalities has been discssed (Owen & Brown, 1968). It was sggested that tactal latencies largely reflect information gathering which increases linearly. as complexity increases. The isal latency fnction, in contrast, largely reflects information processing where the redction in latency at both extremes may indicate that stimli need be compared to fewer "ideals." The anchoring data reqire differential interpretation for lowand for high-eomplexity anchors. When anchors were of low complexity, the anchoring effects were highly similar both within and across modalities. When anchors were of high complexity, howeer, the tactal anchoring of isal ratings created assimilation effects where consistent mixed effects occrred nder all other conditions of anchoring. This asymmetry, copled with the low reliability of tactal ratings of complex shapes, increased latency of tactal ratings with increased complexity, and the fact that tactal ratings at the high end of the scale were lower than isal ratings nder comparable conditions, all sggest that this anchoring difference merely reflects an intermodal acity difference. Stated simply, if the tactal forms were adsted for the size of detail to be comparable to isal acity, these differences wold disappear and the reslts wold be consistent with the rest of the data which proide ery strong spport for a single perceier B, B 2 B) B 4 low omplexity shape. on Ql > 6 A)!!.s I/) 5 ')( Ql ' z B 6 B 6 B 7 Be B 9 BID HIgh Visal Scale Vales shape. Fig. 2d. Relation between the isal complexity scale and mean isal ratings for Set B forms scaled tactauy. Perception & Psychophysics, 1968, Vol. 4 (3) 177

4 c.2 6 "6 5 : til 4 til « cf 1 -til 2! ' 3 V 4 ' 5 6 V. If TactaI ". Visa I Visal _ Visal Tactal.... Tactal Visal _ Tactal SSSION I 1--.""----'yL--",.,--'\o:----'rnr-MDIAN Anchoring Form c 6 ''5 5. '.iii Ụ 1 cf >. LO 2 V 5. U. 7. If T.\ A, A) 4 Anchoring TactalllDDl Visal Visal _ Visal Tactal Tactal Visa I '< Tactal A,, Form SSSION ' MDIAN Fig. la. Anchor effects when Set A, high complexity shapes were IeI nder for conditions of sc:aiing. Fig. 3<:. AdJor effects when Set A, low complexity sbapes were UIed nder for conditions of seating. c: ''5 5. ". i 2 V., -til -c: U ' 5. ) Ụ & 7J Tactal."., Visal Visal-Visal Tactal Tactal Visal Tactal - SSSION I MDIAN c' 7..2 "6 ' 5. 'g «' V. & til. 2 i: V i: 5 6 Ṿ a B. """ Anchor Standar Tactal. Visal Visal _ Visal Tactal.. Tactal Visal >< Tactal Anchoring Form Anchoring Form B B.: B,. l Fig. 3b. Anchor effects when Set B, high complexity shapes were UIed nderfor conditions of scaling. Fig. 3d. Anchor effects when Set B, low complexity shapes were UIed nderfor conditions of scaling. 178 Perception & Psychophysics, 1968, Vol. 4 (3)

5 RFRNS ATINAV, F., & ARNOULT, M. D. The qantitatie stdy of shape and pattern perception. Psychol. Bll, 1956,53, BHAR,. J., &: BVAN, W. The perceied dration of aditory and isal interals: ross-modal comparison and interaction. Amer. J. Psychol., 1961,74, BHRMAN, B. W., &: BROWN, D. R. Mltidimensional scaling of form. Percept. &:Psychophys., 1968, in press. BJORKMAN,M., GARVILL, J., & MOLANDR,B. ross-modal transfer as a fnction of preparatory set and distinctieness of stimls aspects. Reports from the Psychological Laboratories, Uniersity of Stockholm, 1965, No BROWN, D. R., & ANDRWS, M. H. A mltidimensional analysis of isal shape discrimination. Percept. &: Psychophys., 1968, in press. BROWN, D. R., &: BRUMAGIDM, S. H. Perceptal eqialence, pattern perception, and mltidimensional psychophysics. Sbmitted to Percept. &: Psychophys BROWN, D. R., &: OWN, D. H. The metrics of isal form: Methodological dyspepsia. Psychol. Bll. 1967,68, AVINSS, J. A., &: GIBSON, J. J. The eqialence of isal and tactal stimlation for the perception of solid forms. Paper presented at astern Psychological Association, Atlantic ity, April, GAYDOS, H. F. Intersensory transfer in the discrimination of form. Amer. J. Psychol; 1956,69, GOLDSTON, S., BOARDMAN, W. K., &: LHAMON, W. T. Inter-sensory comparisons of temporal dgments. J. expo Psychol. 1959,57, GOLDSTON, S., &: GOLDFARB, J. 1. Jdgment of filled and nfilled drations: Intersensory factors. Percept. mot. SkiOs, 1963,17, LOBB, H. Vision erss toch in form discrimination. Ganad. J. Psychoi; 1965,19, OWN, D. H., &: BROWN, D. R. Visal and tactal form complexity: A psychophysical approach to perceptal eqialence. Sbmitted to J. expo Psychol; SMMS,J., WINSTIN, S., GHNT, L., &: TUBR, H. Performance on complex tactal tasks after brain inry in man. Amer. J. Psychol: 1954, 67, STVNS, J.., MAK, J. D., &:STVNS, S. S. Growth of sensation on seen contina as measred by force of handgrip. J. expo Psychol., 196, 59,6-67. NOTS 1. This research was spported by Research Grant HJ>.OO99 from the National Institte of hild Health and Hman Deelopment. 2. Address: Department of Psychology, Prde Uniersity, Lafayette, Indiana (Accepted for pblication May ) Perception & Psychophysics, 1968, Vol. 4 (3) 179