Date of Award

1-1-2019

Document Type

Dissertation

Degree Name

Ph.D.

Organizational Unit

Morgridge College of Education, Research Methods and Information Science

First Advisor

Kathy Green, Ph.D.

Keywords

Data visualization, Graphs, PowerPoint, Presentations, Protocol analysis, Tables

Abstract

This dissertation reports results of a study with a quasi-randomized experimental component and a protocol analysis, or think aloud, component. The experimental component was designed to determine if people with no statistical training and people with some statistical training differed in their understanding and recollection of statistical information with varying degrees of complexity. Information was presented using data visualization techniques based on cognitive theory and compared to presentations using APA-style numerical tables of statistical output. The focus was on using empirically-supported graphical displays in PowerPoint presentations such as one might see at a research conference. Classroom groups of beginning and more experienced statistics students (n = 194) were randomly assigned to watch one of two scripted PowerPoint presentations; one presentation predominantly utilized graphs while the other depended on tables to present the same information. Participants were tested for understanding immediately after viewing the presentations and two weeks post viewing to test their recall of the material. Protocol analysis was used to illuminate the thought processes of individuals with advanced statistical training as they interpreted either the graphs or tables.

Experimental results indicate large effects for complexity and time, and a small positive effect for the graphs treatment. Significant interactions in favor of the graphs treatment were found with novices on easy items in round 1 and for advanced beginners on difficult items for the advanced beginners in round 2. Protocol analysis found that advanced statisticians use the slide title to cue processing and interpretation of the slide content regardless of presentation type, however, they reached the interpretation stage more rapidly and directly when presented with graphs. Results support the use of graphs to enhance understanding and recall of empirical research presentations and present new findings to advance researchers', statisticians', and evaluators' impact, and enhance communication in the classroom and boardroom.

Publication Statement

Copyright is held by the author. User is responsible for all copyright compliance.

Rights Holder

Holly L Roof

Provenance

Received from ProQuest

File Format

application/pdf

Language

en

File Size

160 p.

Discipline

Statistics



Share

COinS