Monitoring data streams: Using judgment to detect regime change and to assess the effectiveness of interventions

The research will investigate people's judgments of whether a change has occurred in time series data. Judments of this type are ubiquitous. Doctors monitor diagnostic indicators for signs of disease onset and for evidence that treatments are effective; farmers monitor soil conditions to determine whether additional irrigation is necessary; probation officers moitor probationers' behaviour for evidence of a return to crime. Many other examples could be given.
Eighteen experiments will address five questions:

Relative to formal methods, how good are people at detecting changes in different features of a time series (eg, level, trend, variance, and autocorrelation)?
What factors influence how well people can detect such a change in a time series (eg, characteristics of the series and how it is presented, speed of the change itself)?
How is ability to detect a change in a series related to performance in tasks that require it? Such tasks include forecasting from the series and assessing the effectiveness of treatments.
How does the nature of the agent producing the change affect how easily it is detected?
What psychological processes are responsible for change detection and how, if at all, do they vary with the factors listed above?