Power law in a probplot

I have a log-log probability plot of some data points. I was hoping to derive a power law from this, i.e. I was hoping to fit a line to the probability curve such that the probability of achieving a data point of size X would equal p(x)=x^a, where a would give the slope of the line. I have constructed a log-log probplot, used the basic fitting feature but I'm not sure how to derive an exponent representing the slope. Any ideas?

Thank you

[Richard Brown]

Before doing that, I suggest you read this http://cscs.umich.edu/~crshalizi/weblog/491.html

[Brandon]

I'm trying to develop data that mimics a -3/2 power law. Right now I'm just not sure how to pull out any kind of law from the probplot I have.

[Brandon]

Ive also switched to a ksdensity graph