issue of replication

[hardmann]

Dear Tom:

If I want to repliacte the published paper, would I get agreement from authors or press.

Best regard

[TomDoan]

Dear Tom:

If I want to repliacte the published paper, would I get agreement from authors or press.

Best regard

I'm not sure what exactly you're asking. It's very common for a paper to use the same data set and (start out with) estimating a published result. Sometimes the new paper reproduces the old results; sometimes not. If you come up with something different, you'll probably have a hard time getting the paper published unless you can show how your results are correct when the published ones are not. That's much easier if you have the co-operation of the original authors (for instance, they admit that they made a mistake), but that's not necessary, particularly if you can demonstrate how to reproduce the original results (for instance, if the published results can be seen to be a local mode, when you've found one with a higher likelihood). If you reproduce the original model and extend it, then obviously, that's a new result.

In general, you do not need to get permission to replicate an analysis, assuming the data have been made available. Quoting extensive portions of the text of a paper is a different matter, but the models themselves have basically been put into the public domain.