1. A market survey company want to lcnow which social media platform people will choose when they have good news to share. They focus on three platforms: Twitter, Facebook, and Snapchat. They focus on three factors: age, gender (0: female, 1: male), and race (0: white people, 1: non-white people). They collect data from 1000 intemet users. a. Which model should the market survey company use? Why? (5 points) b. See the figure below for the results. Why are there no results shown for Twitter? (5 points) c. Using 90% condence level, can you explain the effect of age in social media choice? (10 points) insure Coefficient Std. err. z Phlzl {95% conf. interval] Twitter [base outcome) Flcehnuk age -.0111915 .0000915 -1.04 0.065 -.0231305 .0001415 mate .5139325 .2005221 2.35 0.004 .1009665 .9569935 nonwhite .1312659 .2109?! 3.34 0.001 .302017 1.160055 _cons .1567003 .2320509 0.55 0.500 -.39?E?73 .7110710 Snapchat age -.0050414 .0114114 -0.51 0.609 -.0202073 .0105245 male .5102231 .3639293 1.00 0.161 -.2031625 1.22361 nonwhite .4333101 .4105255 1.05 0.201 -.371407 1.230125 _cons *1.811165 .5348606 3.39 0.001 2.859413 -.7628518 3. Studies have shown that partisanship is a key factor for people to decide whether they want to blame an ofcial or not when there is a policy failure. James is interested in testing this in the ordinary people in United States. I-Iis argument is that, when there is a policy failure, people who know the ofcial's partisanship will blame the ofcial more because they will assume the ofcial makes mistake because of partisan bias. preople do not know the ofcial's name, they will assume the ofcial is neutral and blame the ofcial less for the failure. James decides to conduct an experiment on 1000 Americans. 1) For James' research question, how many experiment groups should we randomly assign participants into? What are those groups? (5 points} 2) Why is it very important to assign participants to those groups randomly? In the other words, why randomization is a must? {5 points) 3) Can you design an appropriate treatment for this research question? (10 points) 4) If James' argument is correct, what should we observe from the experiment? Why? (5 points)