Reliable characterization of the pore structure is essentially important for transport-related durability studies of cementitious materials. Mercury intrusion porosimetry has been commonly used for pore structure measurement while the ink-bottle effect significantly affects the trustworthiness of pore size features of cementitious materials. Pressurization-depressurization cycling mercury intrusion porosimetry (PDC-MIP) is an alternative approach previously reported with the purpose to provide better estimates of pore size results. It is found however that the PDC-MIP greatly overestimates the ink-bottle pore volume owing to the incomplete extrusion of mercury in throat pores after the pressurization-depressurization cycle. Intrusion-extrusion cyclic mercury porosimetry (IEC-MIP), as a further improvement, is then described, which can reliably capture the ink-bottle effect and obtain a clear picture of the distribution of the ink-bottle pores in cementitious materials. The ink-bottle effect of cement pastes is observed being pore size-dependent and the role of critical pores is emphasized. Water-cement ratio primarily changes the effective porosity while plays a minor role in the ink-bottle porosity. The addition of reactive blends substantially enhances the ink-bottle effect during mercury penetration into small pores. IEC-MIP tests, together with a unique data analysis, enable to obtain a more truthful pore size distribution.