If I have a GST 26kDa polyclonal antibody, will it bind to the GST 28kDa protein as well?
If the sequence of the 26 kDa form is included within the sequence of the 28 kDa form then the answer is 'probably yes'.
However if the N- or C-terminus of the short form is an important/dominant epitope and it is masked in the long form (by being extended) then the activity of the serum against the long form could be significantly reduced.
As Alan said, it depends but I'd guess that, yes, it probably does.
Given that you are using polyclonal antibodies against a not so dramatic extension, I'd say it's even more likely that you'll see signal. The various antibodies within the polyclonal mixture will bind different parts of the smaller protein, so unless the additional segment causes a dramatic change in shape or somehow blocks most of the binding sites, you should see binding.
I don't think you can directly compare binding signals between the different protein forms, but it should work for simple detection or purification