Subject Specialist:
Statistics and Data
U.S. Government Documents
814-865-0665
e-mail: sjw31@psu.edu
Subject Specialist:
Statistics and Data
U.S. Government Documents
814-865-0665
e-mail: sjw31@psu.edu
Note: [Penn State Libraries] indicates a resource available to researchers with a Penn State Access Account.
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).
iPOLL [Penn State Libraries]
Contains summary statistics from various public opinion polls, but primarily from the Roper Center
Odum Institute's Public Opinion Poll Question Database [Penn State Libraries]
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 [Penn State Libraries]
Surveys conducted by more than 700 polling organizations in the United States and more than 80 other countries from 1986 to the present
Some questions to ask yourself when looking at polls and polling data:
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 extremely important! For example, was this a poll of adults' perceptions of teenagers, or a poll of teenagers' 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 entire U.S. population.
How large was the sample?
A good figure to keep in mind for a national sample is 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).
