Polls and Surveys

• Penn State Pulse
  Provides summary statistics from surveys of Penn State students.

• PollingReport.com
  Many free survey and opinion poll statistics, but you must be careful about evaluating the results [see Evaluating Polls and Surveys below].

The following databases are available to researchers with a Penn State Access Account:

• iPoll
  Contains summary statistics from various public opinion polls, but primarily from the Roper Center.

• Gallup Brain
  Contains reports and summary statistics from public opinion polls conducted by the Gallup Organization since 1935.

• Odum Institute's Public Opinion Poll Question Database
  The Odum Institute for Research in Social Science maintains one of the oldest archives of polling data in the United States and is the exclusive repository for Louis Harris public opinion data.

• Polling the Nations
  Surveys conducted by more than 700 polling organizations in the United States and more than 80 other countries from 1986 to the present.

• Who is the sponsoring agency?
  Often the organization or agency sponsoring and/or conducting the poll has a particular bias and/or agenda.

• What is the purpose of the poll?
  This cannot always be easily ascertained, but often if you will take the time to look at the mission statement of the sponsoring agency the purpose will become self-evident.

• What other questions were asked and in what order?
  Often the purpose of the poll can be more clearly defined by looking at the other questions that were asked. The order can also be important. Are there "leading" questions?

• Who was polled?
  This is iextremely important! for example, was this a poll of adult's perceptions of teenagers, or a poll of teenager's perceptions of teenagers?

• When was the poll done?
  The timing of a poll is also important. For example, if you were asking U.S. citizens for their perspective of the Middle East before and after the the terrorist act of 9/11, responses would be dramatically different.

• How were the interviews conducted?
  How the polls were conducted can also be important. For example, were these polls conducted in "American Idol" fashion? These polls are conducted by surveying only those who happened to be watching the show that evening - not exactly representative of the U.S. population.

• How large was the sample?
  A good figure to keep in mind for a national sample is around 1,000. Anything less is suspect. However, a smaller sample may be appropriate if it is only trying to measure a smaller population [e.g., Penn State teaching assistants],