UPI/CVoter Poll Minorities Tracker 2016 Current projections are based on UPI/CVoter’s National & State tracking poll conducted online during the last four weeks among 18+ adults nationwide, including likely voters, details of which are mentioned beside the projections as of today. In our National Poll, we are tracking opinions of more than 200 people each day, leading to a national representative sample size of at least 1,400 people during any seven-day span. In our State Poll, we are tracking opinions of about 250 likely voters in each state every week, leading to a state representative sample size of about 500 likely voters during any two-week span. The latest 500 sample sub set is taken for each minority group for further analysis. Sometimes the table figures do not sum to 100 due to the effects of rounding. pg. 1 UPI/CVoter Tracking Poll Tracker2016 If the U.S. presidential election was held today, which candidate would you vote for? *LV Only* UPI/CVoter exclusive: Minorities Voting Trend. Target booster samples done in last four weeks. Latest 500 samples taken for each demographic group. Clinton beats Trump in every minority group Candidate Indian American African American Hispanic Other Asian Clinton 57.6 89.6 75.3 72.0 Trump 29.3 7.6 21.5 22.2 Lead of Hillary Clinton 28.3 82.0 53.8 49.8 Sample Size 500 500 500 500 Indians are not supporting Trump, neither in the United States nor in India. They are overwhelmingly pro-Democrat, like all other minorities groups. Candidate Indians In India Indian Americans Clinton 49.1 57.6 Trump 26.6 29.3 Lead of Hillary Clinton 22.5 28.3 Sample Size 1200 500 pg. 2 Methodology & Tracker Details Statistical margins of error are not applicable to online polls. The precision of online polls is measured using a credibility interval. The error due to sampling for projections based on the Likely Voter sample could be plus or minus 3 percentage points at the national level and plus or minus 5 percentage points at state level. All sample surveys and polls may be subject to other sources of error, including but not limited to coverage error and measurement error. In our State Poll, we are interviewing about 250 respondents in EACH state every week. This totals to more than 12,500 interviews every week and about 25,000 samples every two weeks. We make a rollover master file every day for each and every state from the samples collected in the last 14 days; this gives us a new sample of about 500 every day for each and every state. Thus, rather than a routine weekly tracker, the UPI/CVoter Poll is a daily tracker, giving the public opinion trends on a daily basis. We will analyze the Presidential Tracker over the rest of the campaign until Election Day. The State Poll data collection started Sept. 1 and will continue until Nov. 7. This gives us an opportunity to analyze fresh data daily through Election Day. We did multiple rounds of pilots with different platforms, both online and offline, and eventually decided that we would use the online mode for data collection. In the State Tracker, we employ multiple providers of panels to randomize and remove the contact bias of any one particular sample provider, if any. Just like in our 2012 presidential polls, our exclusive Psephometer algorithm will be updated every day. So we will have the national projection as well as the state-level projections on a daily basis. This will be a unique tracker from that perspective. Code of conduct & IPR Details All our tracking polls conducted for media are released in public domain and report archives are available on public platform. We are proud to follow the WAPOR/ESOMAR Code of Conduct and meet the AAPOR Level 1 standard of disclosure. All publication rights for worldwide UPI releases remain with UPI, and all survey instruments, design and data IPR remain with CVoter. Contact: Yashwant Deshmukh, Editor, CVoter International: yashwant@teamcvoter.com Charlene Pacenti, Chief Content Officer, UPI: cpacenti@upi.com pg. 3