2. Panel Data Models Suppose you collect three years of data on state-level soda demand. You estimate a pooled cross sections 1, first difference, and a fixed effects regression based on the following regression model: qsoda,itd=+sodapsodait+i+2015t2015+2016t2016+it The results from these regressions are shown in the table below. 1. Using the pooled cross sections results, by how much does soda demand decrease if the price of soda increases by $1 ? 2. Using the first difference results, by how much does soda demand decrease if the price of soda increases by $1 ? 3. Using the fixed effects results, by how much does soda demand decrease if the price of soda increases by $1 ? 4. Using the fixed effects results, is soda demand greater in 2014 or 2015 ? 5. Using the fixed effects results, is soda demand greater in 2014 or 2016 ? 6. Why is the estimate of soda different for the first difference and fixed effects models? 7. Would you present clustered or robust standard errors for all three models given that you observe soda demand for 50 states over a three vear period? 8. Given that there are three years of data and 50 groups, which model is preferred? 2. Panel Data Models Suppose you collect three years of data on state-level soda demand. You estimate a pooled cross sections 1, first difference, and a fixed effects regression based on the following regression model: qsoda,itd=+sodapsodait+i+2015t2015+2016t2016+it The results from these regressions are shown in the table below. 1. Using the pooled cross sections results, by how much does soda demand decrease if the price of soda increases by $1 ? 2. Using the first difference results, by how much does soda demand decrease if the price of soda increases by $1 ? 3. Using the fixed effects results, by how much does soda demand decrease if the price of soda increases by $1 ? 4. Using the fixed effects results, is soda demand greater in 2014 or 2015 ? 5. Using the fixed effects results, is soda demand greater in 2014 or 2016 ? 6. Why is the estimate of soda different for the first difference and fixed effects models? 7. Would you present clustered or robust standard errors for all three models given that you observe soda demand for 50 states over a three vear period? 8. Given that there are three years of data and 50 groups, which model is preferred