Usually we plot graphs of our numbers, and get a nice curve from them.
Then we compare the curve with results from other observations and see if they match up. If they don’t then that tells you something.
Either you are wrong, or the data is incomplete, or there is something you haven’t factored in.
Usually we don’t know what the “correct” answer is, but we know what it is likely to be roughly.
Julians right, much easier to figure out if things are right from plots rather than just numbers.
Its also good to run the same analysis on some test data where you “know” the answer, that way you can check your method is working correctly before you do it on the real data.
We don’t ever really know if our work is right, we just have to factor in errors from all the things we know about, make our very best estimates, and compare to things other people have done to see if it looks ‘sensible’.
Comments