South Carolina ETV

Winthrop/ETV Poll

Methodology - August 2008 Poll

This Winthrop/ETV Poll was conducted among 1088 randomly sampled Likely Voters in the South between August 1 and August 17, 2008. Data utilizing all respondents has a margin of error of plus/minus 2.97 percent. As is true with all survey data, any results that use a subset of the respondents will have a higher margin of error.

For the purposes of this survey, “the South” is defined as: AL, AR, FL, GA, LA, MS, NC, SC, TN, TX, VA. Data is weighted by state.

States are sub-divided and divided as follows:

Deep South: AL, GA, LA, MS, SC
Peripheral South: AR, FL, NC, TN, TX, VA

Further, Southern African Americans in this sample were overwhelmingly supporters of Obama (only 3.2% identified themselves as McCain supporters). Therefore, breakdowns by “Christian” and “Evangelical” are done exclusively with white respondents. Many pundits have questioned whether the Democratic candidate can make in-roads with these groups.

For the purposes of this analysis, “working class whites” are defined as white respondents with less than a college education and an annual household income no greater than $50,000.

Not all tables sum to 100% due to rounding.

The sample of phone numbers for registered voters for the August 2008 Winthrop/ETV Poll of the southern states was purchased from an outside vendor specializing in providing survey samples. Some states provide regular updates of their registered voter rolls to the vendor while a third-party compiler provides updated information for the remaining states. The primary outside vendor matches voters with phone records for all states and regularly updates and appends (when necessary) these files. So called “Do-Not-Call” phones are included. Do-Not-Call rules do not restrict these surveying and polling applications.

The name and registration status of each respondent was verified and then respondents were screened based on likelihood of voting. Additionally, the data were supplemented with Random Digit Dial sample where households were screened for the presence of registered voters. If more than one registered voter lived in the household, respondent selection was randomized. Finally, these respondents were also screened based on likelihood of voting. This additional step ensures that the final sample includes a representation of the most recent registrants who, by definition, could not appear on recently released voter rolls (which, at the minimum, are several months old). Surveys that do not include this step (either by only using listed registered voters or voters who have voted in some specified previous electoral contest) contain a built-in bias against newly registered voters.

Viewed 13856 times

ETV encourages readers to comment on our posts. We seek comments that are brief, on topic, civil, truthful and not abusive. We pre-moderate comments, so it might take some time for your comment to appear. Thanks for your input and comments.
Add Your Comments

Name:

Email:

Location:

Remember my personal information

Notify me of follow-up comments?

Submit the word you see below:


 

SCETV Webmaster
E-mail: webmaster@scetv.org For more information about Winthrop/ETV Poll, contact:

Television:
Aimee Crouch, Coordinating Producer
E-mail: acrouch@scetv.org

Media/Press:
Rob Schaller, Director of Communications
E-mail: rschaller@scetv.org

NEWSLETTER

Either Javascript is turned off in your web browser or you have encountered a server error. We truly apologize for the inconvenience. Please submit your request for our email newsletter by email. Thank you for your interest.

  • Home
  • |
  • ETV Stations
  • |
  • Network Status
  • |
  • Weather and Emergency Alerts
  • |
  • Jobs and Internships
  • |
  • Pressroom
  • |
  • Privacy Policy
  • |
  • About ETV
  • |
  • Contact Us
    • 1101 George Rogers Boulevard, Columbia, SC, 29201 — Tele: (803) 737-3200


    NPR - National Public Radio    PBS - Public Broadcasting Service