物體輪廓圖形得評估研究

NORMS FOR 105 OBJECT CONTOURS

本研究延續汪曼穎（1997）的常見物體輪廓圖形之知覺及語意特性評估研究，增加建立105個不同於該研究的圖形常模，評估每個圖形之「叫名正確性」、「叫名一致性」、「熟悉度」「主觀複雜度」、及「心像一致性」的度量。結果顯示這些圖形對大學生而言大部分均為可辨認，可以做為一般以圖形為材料之研究使用。但由於重複先前研究的三個圖形中有二個在「心像一致性」度呈上的評估結果與該研究有顯著的差異，我們建議當研究者操控或平衡圖形的「心像一致性」時應該各自對應兩個圖形庫中的資料。此外，造成心像一致性差異的因素以及常模穩定性的議題也予以討論。

This study extends Wang's (1997) work on the perceptual and semantic characteristics of 132 object contours. We evaluated another set of 105 object contours that was examined in Snodgrass and Vanderwart's (1980) original work. College students evaluated each picture on five measures: naming accuracy, name agreement (information quantity H, percentage of the dominant response), familiarity, subjective complexity, and image agreement. The majority of these pictures was rated as highly recognizable and hence can be used for stimulus materials. Repeating three pictures from Wang's study, we however found subjective ratings for two contours on the scale of image agreement significantly different from the ratings observed by Wang. Factors that may cause the discrepancy on image agreement, issue of norm stability, and proper control of stimuli from two separate norms are also discussed.