Traditionally the autocorrelation function doesn't do this. In fact, it almost looks like a triangular window has been applied to the data. How can I account for this so my results match the theory? Thanks, Joanna
I would actually expect the data to taper towards zero at both ends of the output of the autocorrelation function. You can think of an autocorrelation as passing a signal over itself and integrating the product of the amplitudes where they overlap. As you slide the signal over itself there won't be much overlap in the beginning and towards the end - hence the small values at the ends of your output data set.
