To further test the Involvement Load Hypothesis proposed by Laufer & Hulstijn (2001), the present study, adopting user-behavior tracking technology (Collentine 2000), recorded the on-line learning behaviors of Chinese non-English-major students when they were performing tasks that varied in task-induced involvement load. In addition, one immediate posttest and two delayed posttests were administered to examine the respective association between the initial learning and retention of the target words on the one hand, and task type (in terms of involvement load) and on-line learning behaviors on the other. Through a computer program specially written for the present study, 81 participants were randomly assigned to 4 tasks that varied in involvement load (from 0 to 4) but all contained 9 target words to be learned. The program also recorded the target words that were clicked, the number of clicks, the time spent on the target words, and the target words that were repeatedly clicked. Statistical analyses show that (1) what affected the on-line learning behaviors of the participants was the nature of the task, rather than the task’s involvement load; (2) on-line learning behaviors were to some extent associated with the initial learning of the target words, but they were not correlated significantly with the retention; and (3) involvement load did not appear to have a significant influence on both the initial learning and retention of the target words. Thus, the Hypothesis is not verified. Further examination of the results reveals that the causes behind the failure to verify the Hypothesis are: (1) as a construct, some aspects of "involvement load" are difficult to operationalize, and (2) it is almost impossible to manipulate the learners’ attention through instructions of tasks.
