Public Policy Polling conducts its surveys through IVR telephone calls to landline phones and text messages sent to cell phones that direct respondents to a Survey Monkey link to fill out the survey. PPP conducts random surveys of registered voters using voter lists provided by Aristotle Inc. Quality control steps are taken with the texting sample on every survey to ensure that only those who were actually part of the sample have responses counted, and so that respondents are not able to answer the survey multiple times. Surveys are weighted with a raking method that incorporates variables like previous Presidential vote, gender, race, age, education level, party registration where available, and geography. Weighting targets are built based on a proprietary combination of census data, voter registration statistics, demographics of past voter turnout, and previous election results to be as reflective of the electorate being polled as possible. Margin of error is calculated at a 95% confidence level based on the size of the population being surveyed and the sample size. Weighting may introduce additional error and all surveys have additional unmeasured error. Most surveys are conducted in English with some also incorporating Spanish in areas with high concentrations of Spanish speakers.