The best critical examination of Kirkpatrick I have seen to date is in High Impact Learning (Brinkerhoff). I have yet to see anyone explain it better.

I'm with ShoreGirl. I've been asked to use L1-L4 evaluations, but most of the time, we stopped at L3. It's not because we didn't want to proceed, but just because it's way more complicated. We are talking about human beings, whose learning results can be impacted by a varity of factors, of which training is one.
We always want to measure the result in both quantitative and qualitative ways, but does the measurement really reflect the outcome? I don't know, but I guess Kirkpatrick provided a relatively easy and feasible way to evaluate training results.

Please read Brinkerhoff. I'm begging everyone in this profession or considering it to read his critique of Kirkpatrick.