Download Full Text (1.3 MB)
Reported protocols include the methodologies employed by the most recent iterations of the Public Opinion Survey, Exit Poll, and the Riots Survey. Methods for the previous versions of these instruments have been similar and included minimal variations. Details for each year’s methodologies are available upon request.
LOS ANGELES PUBLIC OPINION SURVEY
As part of Forecast LA’s unique approach to forecasting in the Los Angeles region, the Center for the Study of Los Angeles (StudyLA) conducted an outlook survey. The Los Angeles Public Opinion Survey involved 20-minute telephone sessions and online surveys with more than 2,400 adults (1,800 phone and 600 online) living in Los Angeles County. Survey respondents were asked about quality-of-life perceptions, personal economic wellbeing, economic concerns, overall life satisfaction, and various civic issues.
Since the primary purpose of this study was to gather representative input from adult residents within the Los Angeles region, an initial random digit dial (RDD) sample was employed. The RDD sample was drawn by determining the active phone exchanges (the first three numbers of a seven-digit phone number) and blocks with a given sampling area (in this case, by the zip codes that comprise the county). A random list of all active residential and cell phone numbers in the area was produced. This method included both listed and unlisted phone numbers. Listed samples were used to meet particular quotas for racial/ethnic categories and geographic location. The online portion was comprised of responses from double opt-in respondents who have agreed to participate in surveys. Real-time sampling and survey publishing services were also used to target respondents outside of the initial reach. Finally, listed samples were used to fill gaps within racial/ethnic and geographic quotas. For all methods, adult respondents in LA County were targeted and then randomly selected within their group.
The protocol for this study involved asking potential respondents a series of questions, referred to as screeners, which were used to ensure that the person lived within the county and was at least 18 years old. The target sample size was 1,200 residents from the city of Los Angeles and 1,200 residents from Los Angeles County who live outside the city of LA. The first quota was a random digit dialing of approximately 750 residents (with 66% cell phone). The online survey ran concurrently with a target sample size of 600 respondents. The remaining racial/ethnic and geographic quotas were determined based on the fallout: 400 African American residents, 600 Asian residents, 400 residents from the San Fernando Valley (only within the city of Los Angeles) and 400 residents from the San Gabriel Valley. Given the demographic proportion of Latino and white residents in the region, as expected, both groups naturally fell out from the initial wave of online and phone respondents.
Telephone surveys were conducted the first four full weeks in January 2017 and first two weeks in February between the hours of 4:30pm and 9pm during the week, 10am to 4pm on Saturday, and 10am to 5pm on Sunday. The survey was translated into Spanish, Mandarin, and Korean. Translators who spoke Spanish, Mandarin, and Korean were available to conduct interviews for residents who only spoke, or were more comfortable speaking any of those languages. The online survey ran from January 13 to January 26 and was available in English and Spanish. The margin of error is ±3.0%. The 2,404 responses collected in 2017 for this survey are weighted based on respondents’ self-identified racial/ethnic group, gender, age, and geographic location matching them to the most recently available population parameters from the American Community Survey (ACS) estimates.
LOS ANGELES VOTES EXIT POLL
As cities like Los Angeles grow in size and diversity, accurately polling voters becomes a vital tool for understanding decisions that improve voters’ perceptions about their quality of life and their choices at the ballot box. Since 2005, the StudyLA “LA Votes” project has conducted eleven exit polls in the Los Angeles area, performing the largest per-capita exit poll in the nation for over ten years. The Los Angeles Votes project began as a response to exit poll discrepancies in the 2000 and 2004 Presidential Elections. Ethnic breakdowns from exit polls conducted during these elections skewed Latino voters towards the Republican candidate while no other data supported this trend. The racially stratified homogeneous precinct approach was designed by StudyLA researchers to address this sampling anomaly and more accurately reflect ethnic voting patterns discovered by exit polls. In contrast to conventional sampling methodologies, the racially stratified homogenous precinct approach places an emphasis on location when it LAPD | 15 thomas and dorothy leavey center for the study of los angeles loyola marymount university, 1 lmu drive, suite 4119, los angeles, ca 90045comes to racial and ethnic voting patterns, since voting patterns of a racial or ethnic group living within a racial and ethnic enclave are different from voting patterns of racial and ethnic groups living in racial and ethnic enclaves dissimilar to their own. The 2016 Presidential General Election Exit Poll asked voters whom they voted for and why, as well as how they voted on various propositions. In addition, voters were asked about quality-of-life perceptions, personal economic wellbeing, economic concerns, overall life satisfaction, and various civic issues. Voters were also asked for general demographic information such as income, race/ethnicity, political ideology, religion, education, etc.
POLLING PLACE SELECTION
All polling places were located in the city of Los Angeles. Twenty-five polling places were selected via the racially stratified homogenous precinct approach. Geographic location matching for the racially stratified homogenous precinct approach used the most recently available population figures from the ACS estimates. This sampling methodology resulted in five polling places of each racial/ethnic category in the city of Los Angeles: white, black/African American, Latino, Asian, and mixed racial and ethnic groups.
Field researchers were advised to ask every other person who exited the polling place to participate in the survey. If that person refused to participate, researchers were instructed to maintain the same skip pattern and continue to ask every other person who exited the polling place to participate in the survey. Some polling places served multiple precincts.
The 2016 Presidential General Election Exit Poll was conducted on November 8, 2016. More than one hundred LMU students served as field researchers, distributing anonymous English and Spanish exit poll surveys in the city of Los Angeles from 7:00 am to 8:00 pm on Election Day. Field supervisors collected completed surveys throughout the day and a majority of the data were entered by 8:00 pm. The remaining surveys were entered in the days following the election. Over the course of Election Day, 2,829 surveys in English and Spanish were collected by field researchers. The margin of error is ±3.0%. The 2,829 responses collected in 2016 for this survey are weighted based on respondents’ self-identified racial/ethnic group, gender, age, and geographic location matching them to the most recently available population parameters from the ACS estimates.
LOS ANGELES RIOTS SURVEY
In observance of the 5th (1997), 10th (2002), 15th (2007), 20th (2012) and 25th (2017) Anniversary of the LA Riots and Civil Unrest, StudyLA conducted cross-sectional phone surveys of Angelenos and tabulated them to effectively create a longitudinal study of residents’ attitudes toward the city in light of the 1992 Riots. In 2017 StudyLA conducted a survey of 1,203 randomly selected and ethnically represented residents in the city of Los Angeles. The survey was part of the 2017 Public Opinion Survey, using the same methodology. The margin of error is ± 3.0%. The 1,203 responses collected in 2017 for this survey are weighted based on respondents’ self-identified racial/ethnic group, gender, age, and geographic location matching them to the most recently available population parameters from the ACS estimates.