左右腦側化功能的比較實驗研究與腦側化能力測驗和自陳量表的效度考驗

An Experiment on the Hemispheric Differences in the Perception of Visual Stimuli and the Validation of Tests and Scales Hypothetically Measuring Cerebral Laterality

林清山(Chen-Shan Lin)

《教育心理學報》 19期 (1986/05) Pp. 37-54

https://doi.org/10.6251/BEP.19860601.3

摘要

本研究分兩部分。在第一部分有關腦側化功能的實驗裹，研究者利用視野控制法呈現「中文詞」、「英文字」、「黑方塊」、和「完形圖」放受試者的「右視野－左腦」及「左視野－右腦」，以考驗「左腦處理分析性、系列性語文訊息，而右腦處理整體性、同時性圖形空簡訊息」的說法是否篇真。71名大學生參加實驗，每人都接受上述四種刺激項目，每一個刺激項目出現時均與凝視點成2°至6°視角，其曝露時簡篇40msec。實驗採完全受試者內設計來進行。多因子變异數分析後的單純主要效果考驗結果發現：(1)封「英文字」的認識正確率，左腦顯著高於右腦；這與預期的結果相一致。(2)就「中文詞」而言，左腦與右腦的認識正確率幷無顯著差异，無法支持「中文詞因具有系列性訊息，係由左腦處理」的說法。原因可能是曝露時間太長，使這類刺激項目鑑別力降低；也可能是左右兩腦均有中文詞的知識貯存之故。(3)就「黑方塊」的認識正確率而言，左腦顯著高於右腦，正與預期相反。原因可能是本研究的「黑方塊」亦包含分析性、系列性訊息在內，有賴左腦處理；也可能是凝視點上出現的阿拉伯數字引起左腦的激發狀態，使有利於左腦。(4)左右腦在「完形圖」的認識正確率並無不同，亦與預期不一致。原因可能是「完形圖」不但需受試者認爲圖形而且需能産生完形的知覺，在40msec的曝露時間限制下，難度太高所致。第二部分以腦側化實驗的左右腦分數篇效標，考驗腦側化能力測驗及自陳量表的建構及同時效度。受試者仍爲參加實驗的71名學生。(1)以Campbell & Fiske (1959)的方法處理多項特質（左腦功能、右腦功能）－多種方法（實驗法、測驗法、量表法）相關係數矩陣後，發現用腦側化能力測驗或自陳量表所測得的左腦分數與腦側化實驗所求得的左腦分數之間，以及能力測驗所測得的右腦分數與實驗所得的右腦分數之間的相關均未顯著大放0。用腦側化自陳量表所測得的右腦分數與實驗所得的右腦分數之相關爲.24，達.05顯著水凖，但不具實用上的價值。整體言之，幅合效度或區別效度皆甚不理想。(2)再利用典型相關分析法考驗腦側化能力測驗及自陳量表以實驗結果爲效標時的同時效度的結果，也發現四個腦側化測量變項與兩個腦側化實驗變項之間無典型相關存在。測量變項解釋實驗變項總變異6%，仍侮爲機遇所造成。基于這些研究結果，筆者建議能力測量或量表編製者不要輕易宣稱他的測驗或量表可以測量左腦功能或右腦功能，或可以區分受試者爲「左腦型」或「右腦型」。

並列摘要

This paper consists of two parts. Part one describes an experiment on the hemispheric asymmetries in the perception of visual stimuli. It was designed to test the accuracy of the well-known notion that ”the left hemisphere is superior for analytical or serial verbal processing, and the right hemisphere is superior for holistic or parallel visuospatial processing”. Four kinds of visual materials, named Chinese Characters. English Syllable, Latticed Square, and Gestalt Closure, were tachistoscopically exposed in the LVF (left visual field) and the RVF (right visual field) of 71 college students, through visual half-field presentation, with visual angle of 2°to 6° from fixation point, and with exposure time of 40 msec. The subject was required to do immediate recall by writing down what he or she had perceived. Recall score, the percentage of correct recall, was used as response measures. Data thus obtained were analyzed by within-Ss factorial ANOVA. The results were as follows: (1) Items of English Syllable presented in RVF to LH (left hemisphere) were more accurately identified than they were presented in LVF to RH(right hemisphere). (2) No significant laterality differences was found in regard to the recall scores of Chinese Characters. The hypothesis that ”since the item of Chinese Characters is sequential in nature, it would be more superiorly processed by left hemisphere” was not supported The nonsignificant differences was possibly due to t act that exposure time of 40 msec was too long for each item of Chinese Characters and therefore the item discrimination power was not good enough, or due the possibility that both the left and the right hemispheres had the storage of Chinese lexical knowledge. (3) Contradictory to what was predicted, Ss yielded significantly more accurate responses to items of Latticed Square when they were exposed to LH than when they were exposed to RH. This fact suggested that the items of Latticed Square might possibly contain in it some sequential information which the LH is superior to process, and that the digit presented at the fixation-point might cause the activated state of the LH and then produced an artifactual advantage for the LH. (4) In regard to items of Gestalt Closure, there was no lateral superiority for the RH. The hypothesis that the RH is superior for holistic or simultaneous information processing was not supported in this case.
Part 2 describes the results of the study on validation of tests and scales hypothesized to measure cerebral laterality. The present author felt strongly that such kinds of tests or scales should be validated by using the same subjects who participated in the experiment on cerebral laterality. Therefore the same 71 college students also served as subjects of this part of study. For each subject, a LH score and a RH score were obtained through experiment, testing, and scaling. The results were as follows: (1) The application of a multitrait-multimethod approach (Campbell and Fiske, 1959) to the construct validity failed to show promising evidence for either convergent or discriminant validity of the tests and scales supposed to measure hemisphere differences in this study. (2) The canonical correlation between variables of measurements (the LH and the RH scores obtained from testing and scaling) and the variables of experiment (the LH and the RH scores obtained from experiment) was not significantly different from zero both statistically and practically. Thus, the concurrent validity of the tests and the scales employed in this study was also not promising, when the LH and the RH scores obtained from experiment were used as criteria in this validation study. Based on these findings, the present author pointed out that it is inappropriate to claim that a test or a scale can be used to assess hemispheric brain functions, or to classify an individual into left-brain or right brain type, unless the test or the scale has a significant construct or concurrent validity based upon correspondence with a criterion obtained from experiment of hemispheric laterality.