Cross correlation applicable?

In deciding whether an athlete has recovered from a concussion, his
performance in cognitive tests taken after injury is compared to those
taken before injury.  The test results consist of around 100 data points
which measure reaction times under various conditions.  The usual method
for answering the question as to whether an athlete has recovered from a
concussion is to compare the means of the reaction times for the 2 tests.

My question relates to whether the sets of data points might be viewed as
digital signals and whether cross correlating the signals would reveal
anything useful.

On 12 Jun, 13:49, "cl" <cliff.ly...@gmail.com> wrote:
> In deciding whether an athlete has recovered from a concussion, his
> performance in cognitive tests taken after injury is compared to those
> taken before injury.  The test results consist of around 100 data points
> which measure reaction times under various conditions.  The usual method
> for answering the question as to whether an athlete has recovered from a
> concussion is to compare the means of the reaction times for the 2 tests.
>
> My question relates to whether the sets of data points might be viewed as
> digital signals

In principle, yes. One might assume that the athlete becomes
more tired as tests progress, and that a development in
reaction times during the experiment might tell you something
about his stamina.

> and whether cross correlating the signals would reveal
> anything useful.

Not unless the experiment was designed with correlation
analysis in mind. The experimentalist would have to control
factors like the environment where the experiment takes
place, the athlete's preparation in terms of food and sleep,
etc. If such factors were NOT acounted for, forget about
correlation analysis and stick with hypothesis tests.

Rune