Name: Influence of Wording on Survey Responses


Abstract: This study is an experiment done by a Stats 310 group in the spring of 2007. They created two surveys that asked peoplesí opinions (on a scale from 1-9) about abortion and affirmative action (see below for an example). They both gave a scenario before they asked the questions, but one survey contained a liberal bias (adding more emotional words and further development of the scenario). They handed these surveys out in the entrances to both dining halls at Hope, and participation was voluntary. There were 556 responses.












Should be illegal


Should be illegal with some exceptions


No opinion either way


Should be legal with some exceptions


Should be legal

A woman was raped and is now pregnant.What is your opinion on abortion?

Linda Vayer is a 16-year old who was brutally raped and beaten. Now she is pregnant and unable to support a baby from a pregnancy she had no control over. What is your opinion on abortion?



Variables (column):

  1. Survey (0=Bias, 1=Unbiased)
  2. Gender (0=Female, 1=Male)
  3. Abortion (rating, 1-9, see above scale)
  4. Affirmative (rating, 1-9, see above scale)
  5. Combined (Abortion rating + Affirmative rating, 2-18)


Possible Uses:Independent Sample T Test, Box Plots, Bar Graphs, Histograms


References: Stats 310 Project, Spring 2007