Figure 1 indicates the corresponding percentages of students’ correct answers to comprehension tasks at different thermal environments. According to it, we can tell that students' accuracy in answering questions changes when their environment gets hotter. As can be seen, they answered with higher accuracy at 23. 5? C, and kept the percentages of correct answers around 80%. A peak of approximately 85% was reached when the sessions went for 2 hours (120 minutes). Although during the first 1. 5 hours (90 minutes) students responded even more correctly at 20? C, the accuracy percentage then decreased to about 65% in the following time. As for situations at 27? C, it seemed that students showed a relatively lower degree of accuracy in the whole sessions, and the correct answers accounted for around 65% and never went beyond 70%.
After this, another experiment was conducted with all factors staying the same, except for a change in the tasks dealt from English comprehension to mathematics. The results are displayed in Figure 2. It’s easy to notice that at 23. 5? C students answered also around 80% questions correctly during the 3-hour math session and performed better 2 hours (120 minutes) later at an accuracy of 85%. However, the highest accuracy in the first 1. 5 hours (90 minutes) was corresponding to the environment temperature setting at 20? C, and 85% it represented. With time passing, there were more and more wrong answers and the accordingly down-sloping trend of accuracy stopped at the end of the session, at approximately 75%, which were still 10% higher than the best correct answer percentage at 27? C students could get. The corresponding ratio of right answers began at 60%, then followed an increase, and mounted at 65% 2 hours (120 minutes)later. Then they tended to drop slightly and finally fell to 63%.
From what has been observed in both Figure 1 and Figure 2, it is quite clear that the thermal environment do have an effect on students' performance. Moreover, what particularly noteworthy is that 23. 5? C proves to be better linked with better performance on accuracy at the most time. This temperature ensures 80% of responses correct during a 3 hours (120 minutes) session, and even a higher percentage when matched with an appropriate session time setting less than 3 hours (180 minutes). For those sessions lasting for less than 2 hours (120 minutes), environment temperature of 20? C can also guarantee the same high percentage of accuracy, with shorter the lasting time, higher the percentage. Also another thing worthwhile noting is that 20? C clearly beats 23. 5? C when sessions are less than 1. 5 hours (90 minutes). In conclusion, higher performance may be achieved when classroom temperature ranging from 20 to 23. 5? C. Also limitations of the whole experiments exist, such as the weakness of experiment samples in representing the general students, considering they were all NUS students.