I've written a paper that a) demonstrates an equivalence between conditional complexity $K$($Y$|$X$) in information theory and the random component of an effect size estimate $r_{xy}$, and then b) shows that certain metrics of conditional complexity related to Hamming distance can be interpreted as indicating the sign and magnitude of a certain random component of $r_{xy}$. In other words, a function of the Hamming distance between two vectors of scores is indicative of the accuracy of the scores' sample correlation (not to be confused with the precision of $r_{xy}$, its standard error.) A reasonable statistician (or editor of a statistics journal) might reject this claim out of hand because the conventional interpretation of estimation theory has no mechanism for accuracy estimation: it only recognizes one kind of information, parameter (Fisher) information, whereas the random component of a particular estimate from a particular sample isn't a parameter. Or they might just not be used to thinking of "information" in two different ways. So the paper would be very important if correct, but its correctness may be difficult to evaluate.
Indeed, I've submitted the paper to two journals, Psychometrika and Statistical Papers, and I've gotten two form-letter rejections. I fully accept that the respective editors may have given the paper keen consideration, bringing to bear both an open mind and the necessary interdisciplinary knowledge, and they simply found it to be wrong. This isn't me just complaining about rejections, but I don't want to just keep submitting it blindly either way. Assuming it isn't junk, can you recommend a journal that would be receptive to this kind of paper?
Best Answer
IEEE transactions on Information Theory comes to mind. Their specifications of topics is broad.
Fisher information regularly appears in papers there.