Particulate pollution has been clearly linked with adverse health impacts from open fire cookstoves, and indoor air concentrations are frequently used as a proxy for exposures in health studies. Implicit are the assumptions that the size distributions for the open fire and improved stove are not significantly different, and that the relationship between indoor concentrations and personal exposures is the same between stoves. To evaluate the impact of these assumptions size distributions of particulate matter in indoor air were measured with the Sioutas cascade impactor in homes using open fires and improved Patsari stoves in a rural Purepecha community in Michoacan, Mexico. On average indoor concentrations of particles less than 0.25 μm were 72% reduced in homes with improved Patsari stoves, reflecting a reduced contribution of this size fraction to PM2.5 mass concentrations from 68% to 48%. As a result the mass median diameter of indoor PM2.5 particulate matter was increased by 29% with the Patsari improved stove compared to the open fire (from 0.42 μm to 0.59 μm, respectively). Personal PM2.5 exposure concentrations for women in homes using open fires were approximately 61% of indoor concentration levels (156 μg m−3 and 257 μg m−3 respectively). In contrast personal exposure concentrations were 77% times indoor air concentration levels for women in homes using improved Patsari stoves (78 μg m−3and 101 μg m−3 respectively). Thus, if indoor air concentrations are used in health and epidemiologic studies significant bias may result if the shift in size distribution and the change in relationship between indoor air concentrations and personal exposure concentrations are not accounted for between different stove types.