Eagleton
Poll Archive - FAQ |
 |
|
 |
|
|
What is a "frequency"?
What is a "crosstab"? What is a "random sample"? What are "general population" surveys? What are "registered voter" surveys? What are "likely voter" surveys? Why are all of the Eagleton polls numbered, and what do the numbers mean? What does it mean when survey data have been "weighted"? Why does the total number of respondents in my frequency or crosstab not match the total number in the survey? How can I tell if a survey is "good" or reliable? How do I download the survey questionnaire or the complete data file? Why isn’t the most recent poll here? MORE INFORMATION |
||||||||||||||||||||||||||||||
|
What is a "frequency"?
A frequency is the breakdown of answers to a specific poll question, usually expressed in percentages. The table below shows a hypothetical frequency for a question about Governor X’s job performance – 48 percent of respondents said they approved, 48 percent disapproved, and 4 percent did not have an opinion.
 What is a "crosstab"? A crosstabulation or "crosstab" is a table that shows how the answers to one poll question break down according to the answers to another poll question. For example, if you crossed a governor’s approval rating by respondent gender, the crosstab table would show you how each gender group in the survey feels about the governor’s job performance. In the hypothetical crosstab below, the first column shows the total frequency for the governor’s job rating among the entire sample. The second column shows the frequency for Governor X’s job approval among women and the final column shows the frequency among men
What is a "random sample"? The Eagleton Poll conducts its polls among a random sample of New Jerseyans. This means telephone numbers called for the survey are selected at random from the entire possible pool of working telephone numbers throughout the state (including unlisted numbers) – giving every household in the state an equal chance of being selected for each survey. Mathematical probability theory demonstrates that random sampling yields a representative sample of the population without having to talk to every person in the population. In other words, by using random sampling, the Eagleton Poll ensures it is talking to a representative sample of New Jerseyans in each survey. What are "general population" surveys? General population surveys are polls using a random sample of all adults (18 years of age and older) in New Jersey. What are "registered voter" surveys? Registered voter surveys are surveys that include a random sample of only those New Jerseyans who report that they are registered to vote at their current address. What are "likely voter" surveys? Likely voter surveys are surveys of those registered voters (see above) who are considered likely to participate in an upcoming election. Determination of likelihood to vote may vary from year to year, or poll to poll, but typically these voters are those who say they are likely to vote, show some level of interest in or attention to the election in question, and have voted in the past. Why are all of the Eagleton polls numbered, and what do the numbers mean? The Eagleton poll has numbered all of its public polls consecutively since it began polling in 1971. The numbers are a means of organizing the surveys chronologically. What does it mean when survey data have been "weighted"? After conducting a survey the random sample of respondents interviewed is compared to the general population of New Jersey on certain demographic characteristics (typically age and education), using U.S. Census figures. Typically there are small discrepancies between the sample in the survey and the general population. The Eagleton Poll corrects these small discrepancies by applying a small mathematical correction – or weight – to each respondent, to make the sample as representative as possible. These weights rarely change the results of the survey by more than one or two percentage points. Why does the total number of respondents in my frequency or crosstab not match the total number in the survey? Occasionally some questions in a survey are asked of only a sub-group of respondents rather than of the entire sample. For example, if the survey was asking a few questions of New Jersey commuters in a sample of all adults, it would first determine which respondents commute and then ask commuting questions of only that sub-group that does, since the commuting questions would not apply to those who do not commute. All of the survey questionnaires should make clear when and if this is occurring. If you have a question about the number of respondents in a frequency, check the questionnaire to see if questions that came before yours were used to filter out respondents to whom the your question did not apply. How can I tell if a survey is "good" or reliable? The National Council on Public Polls’ website has the twenty questions every poll consumer should ask about a survey to judge its quality. Visit the site at: http://www.ncpp.org/qajsa.htm How do I download the survey questionnaire or the complete data file? After you have completed a search or browse function, you will see a list of questions that can be selected to generate frequency or cross-tab tables. On the sidebar, there are two functions labeled "Questionnaire" and "Request File". Selecting "Questionnaire" will launch Adobe Acrobat Reader and allow you to view the complete survey questionnaire. Using the "File" menu in your browser, you can either print the questionnaire or download the pdf file to your local computer. Selecting "Request File" will allow you to download the complete SPSS numeric data file in .por (portable) format. This format can be read directly by SPSS or SAS software that is installed on a local workstation. To download the numeric data using Internet Explorer, you will need to right-click on the message "Statistical Data - (poll no)". In Netscape, you will need to do a shift and left-click on the message. Why isn’t the most recent poll here? The results from each survey are released over the course of a few months. As such, there is a six month lag between the time a poll is conducted and when it is posted to this website. For the most recent Star-Ledger/Eagleton-Rutgers Poll results, visit their website. *************************************************** For more specific questions about the Eagleton Poll’s methodology, please visit the poll’s current website at: http://slerp.rutgers.edu For more general polling information, visit the website of the American Association for Public Opinion Research (AAPOR) at: http://www.aapor.org AAPOR is a professional organization of public opinion researchers. All of Eagleton’s polling conforms to the methodological standards established by AAPOR. |