The health effects of pollutants are often modeled by a dose-response relationship where the dose may be the mass of the pollutant delivered to a target body compartment. Since the dosage cannot be measured, the subject's personal exposure or the ambient exposure is often used as surrogates for the dosage. As shown by derivations and simulations, the usage of personal exposure as a surrogate for dosage leads to an unbiased relationship. Since the use of ambient exposure as a surrogate for personal exposure can be treated as an observation of the personal exposure with measurement error, the use of ambient exposure as a surrogate for dosage leads to a biased estimate of dose-response effect. The simulations show that personal exposure as a surrogate for dosage can lead to an increased variance of the regression coefficients relating the measured responses to the scaled dosage.