Josephson Sheree, Holmes Michael E
Communication Department, Weber State University, Ogden, Utah 84408-1605, USA.
Behav Res Methods Instrum Comput. 2002 Nov;34(4):539-48. doi: 10.3758/bf03195483.
The scanpath theory of visual perception was tested using Web pages as visual stimuli. Scanpaths are repetitive sequences of fixations and saccades that occur upon reexposure to a visual stimulus. Since Internet users are exposed to repeated visual displays, the Web provides ideal stimuli to test this theory. Eye movement data were recorded for subjects' repeated viewings of three Web pages over three sessions. Resemblance of eye path sequences was measured with a string-edit method; multidimensional scaling and cluster analysis were used to group sequences. Support was found for the scanpath theory; some clusters included pairs of sequences from the same subject. A repeated measures analysis of variance revealed a statistically significant main effect for stimulus type, with a text-intensive news story page generating more similar sequences than a graphic-intensive advertising page. There was a statistically significant main effect for cross-viewing comparisons, reflecting a linear trend in which eye paths for the same subject became more alike over time.
视觉感知的扫描路径理论通过使用网页作为视觉刺激进行了测试。扫描路径是再次接触视觉刺激时出现的注视和扫视的重复序列。由于互联网用户会接触到重复的视觉展示,网络为测试该理论提供了理想的刺激。记录了受试者在三个阶段对三个网页的重复观看的眼动数据。用眼动轨迹序列相似性的字符串编辑方法进行测量;使用多维标度法和聚类分析对序列进行分组。研究结果支持扫描路径理论;一些聚类包含来自同一受试者的序列对。重复测量方差分析显示刺激类型有统计学上显著的主效应,与图形密集的广告页面相比,文本密集的新闻报道页面产生更多相似序列。交叉观看比较有统计学上显著的主效应,反映出一种线性趋势,即同一受试者的眼动轨迹随时间变得更加相似。