中文摘要：

本研究以操作绩效和主观满意度为指标，对网页信息的单页和多页呈现方式进行研究。实验采用搜索任务，共有46名被试参加实验。研究结果表明：（1）在常用的页码翻页、下拉翻页、填写翻页和上页下页翻页四种多页操作方式中，页码翻页是最优的操作方式；（2）无论是在搜索绩效上还是主观满意度上，网页信息的多页呈现方式都要优于单页呈现方式。本实验结果可以为网页信息呈现设计提供科学的依据。

英文摘要：

The current study examines the differences in user performance and satisfaction within two general types of webpage information presentation （i.e., single-page and multi-pages）. The results provide several empirical suggestions related to webpage design. Two experiments were conducted with 46 participants. The materials used in present study were simulation of web pages, which contains picture and text organized in identical layout. In Experiment 1, we examined the effect of different types of multiple webpage presentation on participants’ task performance and subjective rating. Four types of multi-pages pagination were tested： pagination with a linear count of page chunks, pagination with drop-down selection of page chunks, pagination with direct-enter text box, pagination with forward and backward buttons. Each participant underwent 12 webpage relevant tasks in each page presentation condition first. Participants then were asked to judge these web pages based on their own preference with paired comparison paradigm. The results showed that no significant difference was observed in accuracy, however, there was significant difference among 4 types of pagination both in task completion time （p 〈 .01） and in subjective preference paired comparison（p 〈 .01）.The post hoc analysis of completion time indicated that participants spent shortest time in linear count of page chunks design （841ms, p 〈 .01）. Direct-enter text box had the longest completion time （1884ms, p 〈 .01）. No significant difference was found between forward and backward buttons and drop-down selection of page chunks （1484ms and 1537ms, p 〉 .05）. For the preference, linear count of page chunks was most preferred by subjects （81, p 〈 .01）. In Experiment 2, we further investigated the effect of page presentation on task completion performance and subjective rating by manipulating task difficulty. Each participant completed 14 search tasks, half of them were judged as easy tasks, while the other half belonged to