Measuring Statistical Literacy Surrounding Confidence Intervals


Susan Lloyd (The Pennsylvania State University), Matt Beckman (The Pennsylvania State University)


Abstract

Background 

Having a tool to measure statistical literacy surrounding confidence intervals allows instructors to obtain a clearer representation of whether the corresponding learning objectives are being satisfactorily reached. This insight will help instructors make modifications to curriculum and instruction which could reap considerable benefits. 

A primary objective of the introductory statistics course is to develop statistically literate individuals. There are existing assessments that purport to measure statistical literacy on all major introductory statistics topics. Although this approach provides a comprehensive examination of students’ understanding, the wide-ranging information conceals important takeaways. 

This study develops an instrument, using a fractional factorial design, that measures students’ statistical literacy on one pertinent introductory statistics topic: confidence intervals. This screening design allows for more intentionality in the construction of items and illuminates factors that are influential in explaining students' understanding of confidence intervals. It also allows for evaluation of interaction effects within a balanced resolution.

Methods

The instrument will be evaluated by expert reviewers and will undergo a series of revisions. Student think-aloud interviews will be utilized to make final revisions to the instrument. The construction of this instrument will be outlined, along with discussion on the decisions made regarding the experimental design and item selection.

The final version of the instrument will be deployed to collect data from participating Penn State students enrolled in the introductory statistics course. The resulting data will be analyzed and presented using multivariate techniques. 

Implications

This instrument could then be used by statistics instructors in any environment as an effective gauge of students’ understanding of confidence intervals, with the caveat that open-ended items could limit scalability. Based on the survey results, instructors could make meaningful changes to the curriculum and instruction to enhance their students’ overall learning experience. 

Moreover, other researchers may be interested in investigating any misunderstandings that become apparent from this instrument and identifying ways to more effectively communicate about data using confidence intervals. Researchers may also see the benefit in assessing statistical literacy on a singular topic, which could prompt the formation of similarly designed instruments, such as an instrument measuring statistical literacy surrounding experimental design.
 


register