Concept Maps and Feedback in Statistics Learning: Exploring the Effect of Expert Map Feedback and Peer Feedback on Concept Map Structure

Date of Award

Spring 5-15-2018

Document Type


Degree Name

Doctor of Philosophy in Education (PhD)



First Advisor/Committee Member

Dr. Nyaradzo Mvududu

Second Advisor/Committee Member

Dr. William Nagy

Third Advisor/Committee Member

Dr. David Denton


concept map, feedback, statistics, expert map, education, instruction


This mixed methods exploratory study examined the effects of two types of feedback – Peer and Expert Map – on Concept Maps used as learning tools in a statistics classroom. Of interest were possible effects of feedback on Concept Map structure (determined by structural scores), on structural classification (Discrete, Integrated), and on student choice of starting concept (General, Specific). Student perceptions of feedback and the role of mapping in statistics learning were elicited using a 12-item questionnaire. Two open-ended responses were coded by themes. All data were assessed separately, then merged to enrich findings.

Results from a mixed repeated measures ANOVA of Concept Map scores indicate a significant main effect of Time (F(1, 26) = 4.92, p = .035, partial eta-squared = .159). The main effect of Feedback Group was not significant at the p = .05 level (F(1, 26) = 3.50, p = .073, partial eta-squared = .118). Given the study’s low power, outright dismissal of this effect might result in a Type II error.

Questionnaire responses indicated that Concept Maps and feedback were positively perceived. Disaggregation by structural classification revealed that all students with Discrete maps agreed that feedback sessions aided conceptual understanding (100%), while those with Integrated maps were mixed (42% Agreed, 33% Disagreed). Students with Discrete maps agreed that feedback sessions helped them see new relationships (82%), while students with Integrated maps responded more neutrally (25% Agreed, 58% Neither, 17% Disagreed). Students with Discrete maps planned structure and relationships revisions (85%), while the group most focused on presentation revisions had Integrated maps to start (67%).

Analysis of starting concepts before and after feedback indicated students improved in their starting choices. Results from a 2-sided, exact McNemar test indicate a statistically significant difference in the proportion of students using General starting terms pre- and post- feedback, p = .016.

Regarding classifications, results from 2-sided exact McNemar tests indicate the proportion of Discrete and Integrated maps before and after feedback were not significantly different for the Peer group (p = 1.0), but might be considered significant for the Expert group (p = .065).


Dissertation removed at the request of the author.

Copyright Status

Additional Rights Information

Copyright held by author.

This document is currently not available here.


Copyright Status