Propensity score methods have frequently been used in the literature to estimate causal effects in observational studies. Typically, this type of problem has been assessed using propensity score techniques that require a binary treatment. In this paper, we apply generalized propensity score methods to estimate causal effects with a continuous treatment variable. A recently proposed method uses a boosting algorithm with a novel stopping criterion to determine the optimal number of trees to use in the algorithm, thereby optimizing covariate balance between treatment and control groups. Since this method operates under the assumption of a normally-distributed treatment, we modify this method using different distributions, including negative binomial and zero-inflated negative binomial, to consider continuous treatments that do not follow a normal distribution. The goal of this research project is to develop and assess generalized propensity score methods for analyzing a dose-response relationship in observational studies and as an application, we quantify the impact of visits to Supplemental Instruction on success in an undergraduate introductory Chemistry course. The results of our research show that our proposed methods perform well, although not to the same extent as some of the existing methods.