Abt Associates conducted the survey of registered voters in Georgia’s 6th Congressional District on behalf of the Atlanta Journal Constitution. The survey included telephone interviews with a representative sample of 1,000 registered voters of Georgia’s 6th Congressional District. Telephone interviews were conducted by landline (n=399) and cell phone (n=601). Interviewing was conducted from June 5th to June 8th, 2017.
Sampling
The survey design used as the sampling frame a list of Georgia’s 6th Congressional District registered voters (RV) with their cell phone or landline numbers and a random sub-sample of records without a telephone number appended from that same list, purchased from Aristotle. This sample design is referred to as “registration-based sampling”.
The RV frame contains some information that people who register to vote in Georgia is asked to provide, including their landline or cellphone number. About 62% of the registered voters in Georgia’s 6th Congressional District provide either their landline or cellphone number. Therefore, it is important to also interview registered voters that do not have a telephone appended in the RV list to avoid a potential coverage problem. For this reason, we selected from the RV list a sub-sample of records without a telephone appended and sent for a telephone lookup using their address, which is another information appended to the list. We were able to successfully append a telephone number to about 20% of these records.
For the landline samples, interviewers were asked to speak with the registered voter selected in the RV list with the corresponding telephone number. For the cell samples, interviews were conducted with the person who answered the phone. Interviewers verified that the person was an adult and in a safe place before administering the survey.
All cooperating respondents were asked to confirm their voter registration status. To determine which respondents were likely to vote in the special election, all registered voters were further asked a series of likely voter questions including intention to vote in the June special election, attention to the race, and history of voting.
Weighting
The final weights produced for this survey aligned the sample to match population parameters of the registered voters in the 6th Congressional District of Georgia. To construct the weights, we used the full sample of 1,000 registered voters. The full sample was post-stratified (raked) to benchmark demographic distributions for the 6th Congressional District of Georgia registered voter population, as described below.
The weighting balanced sample demographics to estimated registered voter population parameters for the 6th Congressional District of Georgia. The sample was balanced to match the registered voter population parameters for sex, age, education level and race/Hispanic ethnicity. The population parameters were computed from the RV list of the 6th Congressional District of Georgia purchased from Aristotle.
The weighting was conducted using an operation known as raking ratio estimation, or “raking”. Raking is used to reduce the risk of biases due to nonresponse and non-coverage in sample surveys. The raking procedure uses an iterative technique that simultaneously calibrates the sample to target population distributions defined by socio-demographic parameters. After the raked weights were generated, we examined the distribution of values. The final weights were trimmed at the 2 and 98 percentiles to prevent individual interviews from having too much influence on the final results.
Margin of Error
The margin of error for an estimate is a measure of uncertainty that reflects the fact that the estimate is derived from a sample drawn from the population. If one were to draw a second sample in the exact same manner, the estimate would be different from the first simply due to the fact that the sample contains different members of the population. A third sample would be different from the first two, and so on. The margin of error measures how different estimates could be based on drawing different samples from the same population.
The error margin for the entire sample of 1,000 registered voters is +/-3.5 percentage points. For the sample of 755 likely voters, the margin or error is +/-4.0 percentage points. This includes a “design effect” of 1.27 for both the registered voter sample and the likely voter sub-sample. The design effect is the amount of variability introduced by the weighting.
