A Question of Measurement - Treadmill vs Elliptical

The concept of measurement without error is nonsensical. But when does the error become so large as to render the measurement nonsensical? This was a question that started troubling me in Jan'09. 

Over the last 2 months at the gym, I had relied on the measurements on the gym machines to set targets for myself and going about achieving and improving on them. Soon, however, I started doubting their accuracy. 

Initially, my problem used to be the difference in the numbers between the treadmill and the elliptical. On the treadmill, after a 20 minute workout, I used to feel quite drained. Overall, I used to feel the strain. On the other hand, the elliptical did not feel that difficult - yes, I used to have resistance etc - nevertheless, the workout was never that intense as a run on the treadmill. When one looked at the calories burnt figure however, the elliptical would always show more by atleast 20%. 

I could not somehow reconcile myself to this. Sure - my search across the net on this subject informed me that the elliptical was a very efficient machine. It also placed less stress on the knee joints - and hence, it did not tire one out as much as treadmill.

But, I just did not feel convinced. Maybe I was wrong, and being unscientific. This skepticism however, played a very important role in transforming my attitude to cardiovascular exercise.