A clinical trial is run to compare two cancer therapies with respect to their ability to reduce tumor size during one year of treatment. The difference in efficacy between the two treatment groups (the "treatment effect") will be estimated using a linear regression of percent change in tumor size on treatment group. The research team recognizes that some tumor locations (e.g. lung, kidney, pancreas, etc.) may respond differently to treatment than others and that randomization may not adequately balance treatment assignments within the subsets of patients based on their tumor locations. What can you as the statistician do to get the most precise estimate of the treatment effect possible? a. Run a unique linear regression analyses of percent change in tumor size on treatment for the subsets of patients with each tumor location b. Include a factor for tumor location in a multiple linear regression of percent change in tumor size on treatment group to control for confounding c. Include an interaction factor of tumor location with treatment group in a multiple linear regression to assess for effect modification