中文摘要：

人员疏散既有疏散运动，也有行为和心理反应，因此对同一人群在疏散前后进行两次问卷调查，可以有效地对第3类人员（一般人员）和第2类人员（参加疏散演习和试验人员）的问卷结果进行对比，更准确地分析人员疏散时的心理行为特征。在某高校教学楼进行模拟疏散演习试验，用火灾报警器做疏散警报，重复8次。在演习前5天和演习结束时分别进行问卷调查。调查结果表明，在出口发生堵塞时，67％的人在第1份问卷中选择其他出口，而演习结束时的问卷中，这个选择的人数降至35％。这说明很多人往往会高估自己在面临危机时的反应。担任过一定社会职务的人比一般人群更倾向于独立证实火灾的发生。

英文摘要：

The present paper is to provide a comparative investigation of evacuation behavior and psycho-analysis of the college students be- fore and after evacuation test. As is known, a questionnaire can be made to be an effective solution to making a psycho-analysis and sur- vey of the behaviors of the people in a fire excavation or in any other accidents. In our case, the individual subjects who were to be sur- veyed are composed of three groups of people, that is, people who have had some real experience in emergency evacuation, people who have some kind of experience of simulated evacuations and the people who have no experience at all with such events. The response of gen- eral public surveyed is not enough accurate because they could only guess what the emergencies are. It is just for this mason that we have surveyed twice the same group of people before and after the evacua- tion test. Thus, it makes it possible for us to compare the results be- tween the general public and those who have experienced simulated e- vacuations. Taking the college students for example, they were sup- posed to evacuate from a teaching building in a simulative emergency, when they heard the ring of fire warning. Such experiments were re- peated for eight times. And five days later, a questionnaire was dis- tributed to 109 students for their response to the fire warning experi- ence. When they were asked to feed back the emergency test, feed- backs were received from 97 students. The feedback information data to the questionnaires were twice put into the computer with Epidata software to ensure the statistical quality. The statistical feedback re- suits show that there were no significant difference between the first and second composition of the subjects. From the analysis data of questionnaires, 67 percent of student-subjects would choose other dif- ferent exits when the exit was congested in the first questionnaire, but only 35 percent of second group of subjects did such choice. That in- dicated that a good many of subjects overrate