Use statistical analysis to determine if the phrasing of a yes-no question has an effect on the answer.


Procedure: Easy

Concept: Easy


In this experiment, we want to determine if the connotation of a question affects the answers given by a random sample of people. For example, consider the following questions:

Do you think the current president has implemented effective foreign policies?

Do you think the current president has implemented ineffective foreign policies?


Polls are a very common method to gather data from a population. However, measures should be taken to gather unbaised data. For more information on statistical biases, click here.


Do you think that the connotation of the question will influence the answers given? Why or why not?


  1. Formulate a question to test your hypothesis.
  2. Determine the number of people who you will poll for the positive and negative connotatation of the question.
  3. Conduct your poll (be sure to get a random sample!). Also, be sure to fill out the appropriate consent forms.
  4. Enter the data into excel and run a t-test.


Does the p-value of the t-test show a significant difference?


Try running this experiment with other questions. You may be able to design an experiment that will work for multiple-choice questions or questions that require people to rate on a certain scale. Use our Statistics Guide if necessary.