Human-computer network interaction: Delay effects and its mediators

Enlie Wang, Purdue University

Abstract

Interaction delay in Human-Computer Network Interaction is one of the decisive factors influencing task performance and user satisfaction. This dissertation study investigated several important aspects of interaction delay in HCNI including delay perception, delay tolerance, delay effects on task performance, and the dynamics between interaction delay and Information Technology System usage. Three interrelated experiments were conducted to test 13 specific hypotheses about delay perception, tolerance and effects. Experiment 1 explored delay tolerance model in computer file downloading tasks using scenario-based interactive survey. Experiment 2 investigated delay effects and its mediators in HCNI as well as delay perception and tolerance using Website browsing tasks. Experiment 3 was an initial effort of studying the dynamics between interaction delay and ITS usage using simulation method. Results from previous delay studies including Experiments 1 and 2 were used in Experiment 3 to develop the simulation model. The major findings from this dissertation study are: (1) Delay tolerance is mainly determined by the ratio of perceived delay/expected delay; (2) Actual delay is the primary factor that influences perceived delay; (3) File size, network bandwidth, actual delay, the importance of the file content, and time availability for task completion can significantly affect user's delay tolerance; (4) Both information technology system factors (file size and network bandwidth) and task context factors (time availability and file importance) can significantly affect delay expectation; (5) Delay feedback (presence and format) can significantly improve task preference, but higher task preference is not always related to higher performance; (6) Actual delay can significantly affect users' delay tolerance and performance as well as system performance; (7) Delay perception conforms to Vierordt's law at the time scale of cognitive band (0.1--60 seconds); (8) The simulation modeling is a promising method for studying delay effects and their mediators in HCNI. A new delay tolerance model for file downloading was developed using the results of this study. Further study is needed to investigate the effect of delay feedbacks on delay tolerance and task performance. The simulation model also needs system performance data to calibrate parameter settings.

Degree

Ph.D.

Advisors

Caldwell, Purdue University.

Subject Area

Industrial engineering|Computer science

COinS