When feeding a J-Pole is obviously makes sense to install a choke balun to ensure the outside of the coax shield does not conduct energy down the coax thus degrading the antenna pattern via additional radiator area. This article nicely discusses using beads to achieve the goal:

http://www.antennex.com/shack/Dec99/beads.htm

They suggest putting the balun at the antenna in response to a questioner who thought putting it at the radio might be just as good. So great we know to put the choke balun at the antenna. This article from ARRL…

http://www.arrl.org/tis/info/pdf/9409061.pdf

…which suggests placing the choke 1/4 wavelength from the feed point. Indeed more discussion on placing chokes every 1/4 wavelength along a coax exists here…

http://w2du.com/r3ch21a.pdf

The question is, why can’t we just put the choke right at the feed point? Some NEC simulations reveal high sensitivity of the antenna pattern to any currents flowing anywhere on the coax beneath the antenna. A choke placed a 1/4 wave from the feed point would dangle a 1/4 wave section beneath the antenna right where we don’t want it. I suspect the ARRL article’s suggestion to place the choke 1/4 wave length from the feed point comes from what you would do to use the coax outer shield to form the bottom half of a coaxial dipole. The J-Pole is already a full wavelength antenna and does not need, and simulations suggest is actually compromised by, any feed line radiation.

All this suggests a current choke balun is a wise thing to have as close to the J-Pole antenna feed point as practical.