This link is to a PDF of a paper by Ainley, Nardi and Pratt on the Institute of Education website.
Page 12 describes two tasks that were used in the research, ostensibly to engage students in the use of scatterplots. However, the students also need to consider the signal and noise in the data that emerge during the active graphing process, particularly in the helicopters task.
As students carry out the tasks, the bivariate data are plotted using a spreadsheet. For example, in the case of the helicopter task, the students might be aiming to find the ‘best’ helicopter – the one with the longest time of flight. They might consider the time of flight when helicopters of differing wing lengths are dropped.
The data are likely to be quite noisy given the need to measure lengths and durations of time. Nevertheless, the plot of wing length against time should gradually reveal a wing length that seems to offer the maximum time of flight. The signal that emerges through the noise is likely to be a humped shape with large or small wing lengths resulting in helicopters that drop very quickly.