This seems extremely difficult to me: truly water-soluble? Transparent "solutions"  of such polymers may in fact be polymeric surfactants that form nano sized micelles..

When you want to differentiate between  truly water-soluble and such transparent polymer micelles, which look soluble..., you have to do rheological measurement.

When the "solution" gives an eta-solids line as polymers similar to organic solvent based types, one may consider them to be water soluble.

When you get a eta-solids curve that has a sharp increase in viscosity somewhere between 20-35 wt-%  (depending upon the polymer),  you have nano dispersions, so may micelles and certainly not a true solution.

I wonder whether an ASTM test will cover such a difference..  Both systems I describe look like a solution, while the rheological measurements show they are not...

So reconsider what you want and what you really want to study!

good answer, I would like to add: one should make 

a) particle size measurements (Laser-Doppler or Light diffraction)

b) determine concentration in the "solution" 

c) then filter using nanosized pore membranes (I don't remember how small they are available)

d) then measure concentration and particle size again

if you lost something, you did not have a true solution.

A relatively easy way to know if your polymer will be soluble is knowing its solubility parameter. If you do not have this information, you can calculated it (Energy & Fuels 28.1 (2013): 409-416). How much more your molar mass distribution curve is narrow, more reliable is this method of comparison. But, due to the application of yours product, must be interesting see the size (nm).

Janaina, I agree, and I also used solubility parameters a lot, both such from literature as also self-calculated ones. However, one should finally always run exact analyses, I always did according to what I described above.

I'm very interested in exact (!) data (weblink to article?) that claim to be the border between water-soluble and no, based on SP calculations... as I never found them....

We also use the SP (Hansen/van Krevelen/Hoy) data  to at least try to establish an area where one may be interested in...  This will at least give much more directions than just trial and error.

I'm very interested in exact (!) data (weblink to article?) that claim to be the border between water-soluble and no, based on SP calculations... as I never found them....

We also use the SP (Hansen/van Krevelen/Hoy) data  to at least try to establish an area where one may be interested in...  This will at least give much more directions than just trial and error.

you can consult this standard ASTM D 3132 – 84.

Solubility Range of Resins and Polymers1

See an nice example of this method

http://www.dtic.mil/dtic/tr/fulltext/u2/a222938.pdf

and it reports exactly what you may expect:

"The determination of an exact borderline for solubility/insolubility is not feasible using this method ..."

When you are satisfied with such an approach, use it, but for water please read my first reaction that you may have still a problem that for the eye transparent systems turn out to be nano-dispersions so a thus do not behave as true water-soluble.

We have played around with this subject and we think that my proposed method gives you the best indication for having a true solution or not.

 Bernhard Wessling, I agree that if possible it is better to use techniques that demonstrate exactly the  polymer behavior.

Andre Roelofs, this article discusses what you want. (E. Keshmirizadeh et al. / European Polymer Journal 39 (2003) 1141–1150 - http://trl.lab.uic.edu/1.OnlineMaterials/NOLY/KeshmirizadehHard.pdf).