Investigations into the near infrared analysis of intact tablets allow comparison of the reflectance and transmittance methodologies. The studies involved estimation of the effective sample mass and the effect of sample inhomogeneities. An empirical method for determining effective sample mass used tablets of varying thickness and constant compression force. Tablets containing a range of cimetidine concentrations were used to study the effect of drug concentration on effective sample mass as a function of wavelength of observation. Effective sampling depth was between 1.9 and 2.7 mm for reflectance measurements. With transmittance measurements, signal was completely attenuated when tablet thickness was greater than 3.4–4.9 mm, depending on the wavelength of observation. Multi-layered tablets provided a means of probing the effect of sample inhomogeneity. Both reflectance and transmittance measurements may be sensitive to sample inhomogeneity. Transmittance measurements are sensitive to pathlength variations.