If you consider the system zeolite- water, there is three kinds of water :an osmotic effect , a recovering water or water on surface and hydrated water depending of the relative humidity of the system.  Thise is the case for all porous materials (clays  HDL and zeolites)  see  works of Salles,  Blanc, Mercury  and my papers.

have you considered micro/nano focus x-ray CT to measure actual pore size?

See Works of Salles  using measurements  specific surface combined with thermoporometry . He tries to evaluate the amount of mesoporous water   at nanoscale  in clays.

If we consider the traditional means of pore size measurement, namely gas adsorption, then it is true that there are significant differences between the various legacy and modern methods of calculation.

Kelvin based models, such as BJH, assume bulk condensation behavior, i.e. formation of a liquid in the pore with a hemispherical meniscus. In micropores this property does not hold true. Semi-empirical methods (HK/SF etc) also fail to accurately measure micropores (though for many years represented a significant improvement over BJH). The important factors to recognize in pore size calculation from gas sorption data are adsorbate-adsorbent interactions (surface chemistry), adsorbate-adsorbate interactions (of the confined fluid), pore geometry, (and temperature).

Such details are taken into account in detail in molecular modelling approaches like density functional theory (DFT). Non-local DFT (NLDFT) represents the state-of-the-art for zeolitic materials (quenched solid DFT, QSDFT, is a preferred approach for heterogeneous carbons). The recent IUPAC Technical Report "Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution" by Thommes et al, says " Microscopic treatments

such as density functional theory (DFT) and molecular simulation, which can describe the configuration of the adsorbed phase at the molecular level, are considered to be superior and to provide a more reliable approach to pore size analysis over the complete nanopore range."

This paper also references the ISO standard 15901-3:2007 "Pore size distribution and porosity of solid materials by mercury porosimetry and gas adsorption -- Part 3: Analysis of micropores by gas adsorption" which echoes this sentiment with "Compared to the classical macroscopic thermodynamic models, the NLDFT methods describe behaviour of fluids confined in the pores on a molecular level. This application relates the molecular properties of gases to their adsorption properties in pores of different sizes. It follows that pore size characterization methods based on the NLDFT approach are applicable to the whole range of micro- and mesopores."

I hope that helps.

As many colleagues notice, depending on the method, we are employing different physical/chemical effects to determine pore-size distribution - so this is a sort of answer for question "why..."

Which results is "true" - none! :) However, depending on the purpose of the porous media (what is a plan for its practical application), choose the method (better to find a couple for comparison and to scan a wider pore size scale) that will mimic application conditions for the media... 

Ya that is  true sir. Its still an important area, Where we human still going with assumptions.  Which results is "true" - none! :) However, depending on the purpose of the porous media (what is a plan for its practical application),

Hope so in future we may find the procedure which reflect 100% correct result

 Agree with Dr Martin.