2016 Election Eve Poll Methodology

 

Methodology

Latino Decisions completed 5,600 interviews with Latinos with extremely high propensity voters who reported having already voted, or reported being certain to vote in the November 8, 2016 presidential election.  Interviews were conducted either on-line or via telephone with live callers, all of whom were bilingual, and both phone and web interviews were completed in the language of preference of the respondent. Respondents were reached on landline (40%) and cell phones (35%), or invited to complete the survey on the web (25%), between November 2-7, 2016 and interviews averaged 10 minutes in length.

Respondents were randomly selected from the voter file to match a statewide representative sample of Latinos. Sample was pre-screened based on vote history in previous presidential elections, and date of registration to include a mix of new registrants and first-time voters.  All respondents confirmed their Hispanic identity at the start of the survey, and non-Hispanic respondents were screened out. Respondents were then asked if they had already voted early, and if not, if they were 100% certain they would vote on November 8th.  Any respondent who was not certain was terminated. In past cycles, this methodology resulted in over 90%% of respondents validated as voting in the election, and the distributions on variables of interest did not vary between the total and those validated.

For 11 individual states, a minimum of 400 interviews were completed to provide state-specific reliable estimates.  In Florida only, 800 interviews were completed.  For the remaining 38 states and the District of Columbia, an additional national sample of 300 respondents were collected, and then combined with the 12 stand-alone state samples, weighted to proportional population percentages, to achieve an overall combined nationally proportionate sample.  The national sample of 5,600 is directly proportional to the Latino voter population nationwide by geography, and is weighted to reflect the known Census demographics for Latino voters, most critically age, education, gender and nativity.

The national sample carries an overall margin of error of 1.8%. This margin-of-error is adjusted to account for the design effect resulting from 12 unique sample strata of varying size, mode differences, and post-stratification weighting used to derive the national estimate. Florida has 800 completed interviews and carries a margin of error of 3.5%. The remaining individual states sampled  -- Arizona, California, Colorado, Illinois, New York, Nevada, North Carolina Ohio, Texas, Wisconsin, and Virginia – have 400 completed interviews and carry a margin of error of 4.9%. Interviewing were administered and overseen by Pacific Market Research.

Back to 2016 Election Eve Poll Interactive Results

Back to 2016 Election Eve Poll Toplines and Crosstabs