A better idea about how to think about arguments from analogy is found here, from Richard Feldman:
I mentioned these ideas in class, and they were used in the handout on design arguments, below.
Here's the key ideas:
"What the analogy does is point us in the direction of the relevant generalization.
[W]hen we reconstruct analogical arguments in this way, we can avoid answering the hopelessly vague question, “Is the analogy a good one?” or the related question, “Which analogy is better?” and [Is this thing really "like" that other, or "similar" to the other, since "Everything is like everything else in some ways, not in others"?] Instead, we use analogies to help us identify generalizations to use in reconstructions of the argument. And then we can evaluate the resulting arguments in standard ways: are they well-formed (typically, “yes”)? is generalization reasonable? is the other premise reasonable?
We can spell out the process a little more precisely: there is an intended conclusion, which says that some object - the subject - has some property, A. The analogy says that some other object - the analogue - has that property, A. Usually, it is supposed to be very clear that the analogue has A. Moreover, there is some other feature, B, of the analogue that (allegedly) is sufficient for its having that property and which the subject has as well. The connecting generalization of the real argument will relate this second property to property A. In our example A = not protected and B = clear and present danger. The connecting generalization is: All Bs are As.
Thus, no good reconstructions will contain simple analogical arguments like the ones described above. Instead, they will be arguments invoking generalizations of the sort just described. (Third key point.)"