Abstract: Many experiments had been made for interpreting the cause of dual-task performance decrease, but findings were beginning to provide different answers. Three of the most influential classes of explanations were structure limitation (bottlenecks), resourse competition and outcome conflict (information interference). Based on previous experiments, this research attempted to combine structure limitation with outcome conflict and explore new explanations for dual-task performance decrease. 16 subjects (8 females and 8 males, aged 20- 23 years) were seated in front of a computer screen to complete two searching tasks with two hands simultaneously. Both stimuli were separately displayed on the left and right side of the screen one by one. Subjects were asked to search target words from a series of stimuli, which were made up of figure words, orientation words and strokes of Chinese character words. The target words of one task were figure words, another were orientation words. The SOA(stimulus onset asynchrony) between two tasks were 0 ins, 300 ms and 500 ins. All subjects took part in three SOA condition tasks. The results of the data analysis supported the following conclusions: (1) The SOA between two tasks is a major factor that affected dual-task performance. The longer the SOA, the better the dual-task performance. When SOA was 500 ins, the performance of dual-task was the best. This result confirmed structure theory. (2) The interference between information-processing of the two tasks is another factor that affected dual-task performance. The higher the interference, the worse the dual-task performance. In particularly, when a word was presented for one task that required a "no" response, but was a target for another task, the latency of response time increased. Thus the stimuli for both tasks were not processed independently. This result supported outcome conflict proposition. (3) Although the SOA between two tasks had been prolonged, it couldn't avoid the interference between two tasks. The SOA between two tasks is independent of the effect of information interference. Based on this experiment and previous experiments, a three-factor hypothesis was proposed. Structure limitation, information interference and resourse competition all influenced dual-task performance.

Key words: dual-task performance, structure limitation, information interference, outcome conflict, resourse competition