Illustrating the Effect of Leading Questions in an Introductory Statistics Course

• Activity Number: 654
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Education
• Author(s): Ulrike Genschel*
• Companies: Iowa State University
• Keywords: survey data ; response bias ; leading questions ; proportions ; hypothesis testing

Abstract

In introductory statistics classes the influence of leading questions on survey responses is often discussed in the context of observational studies, survey data or the collection of data in general. In this study we randomly assign students to one of two questions on the same topic - a neutrally worded question and a leading question. We repeat this experiment twice, once using a topic that students are generally familiar with versus a second example that is based on a topic that most students in the classroom are unfamiliar with. At the time of taking the survey students are unaware of the differences in wording and the main purpose of the survey. We present the survey responses and illustrate the effect both leading questions had on the students themselves by analyzing the data using a two-sample test for the difference in proportions. Because tests for proportions are a topic in most introductory statistics classes as well the experiment and the resulting data additionally provide an illustration of the method that students can directly relate to.