Five Tests of the Ability-as-Compensator Hypothesis in Diagram Comprehension Instruction

• In Event: Cognitive Processes in Comprehension of Visual Representations: Art, Diagrams, Graphs, and Models

Abstract

1) Purposes
When students are taught how to comprehend diagrams, do high-spatial students learn more than low-spatial students? Using short-term experiments (Höffler, 2010) and others have found spatial x treatment interactions. Some visual representation conditions are equally effective for students across a range of scores on spatial abilities whereas others benefit students with high spatial scores.

2) Theoretical framework
In Höffler’s ability-as-compensator hypothesis (2010), he speculates that high spatial ability allows participants to generate mental animations to compensate for the lack of animation in a static representation. Generalizing from Höffler’s (2010) findings, we argue that diagram comprehension requires the reader to engage in spatial processing, and that specific types of diagrams require specific spatial skills which can be tapped by different spatial measures (Hegarty & Waller, 2005). Furthermore, certain types of instruction should require specific types of mental transformations (Stieff, Hegarty, & Deslongchamps, 2011) and therefore draw on different spatial skills. This leads to a different perspective on Hoffler’s findings, an intervention-as-compensator hypothesis: namely, targeted interventions can compensate for lower student spatial ability.
Our primary research questions are therefore:
1) Are Conventions of Diagrams (COD), Coordinating Text and Diagrams (CTD), and Self-Explanation in Diagrams (Self-Expl) conditions equally equitable for students with different scores on spatial ability measures?
2) If so, does this pattern differ across different spatial measures?

3) Methods
We used a quasi-experimental design in five separate studies with random assignment of classes to pairs of interventions (COD, CTD, SelfExpl) in two different schools from Spring, 2009 through Fall, 2010. Please see below for participants, measures, materials, and procedures.

4) Data sources and materials
Participants were ninth-and tenth-grade students from intact biology classes.
Measures included demographics, biology knowledge, geoscience background knowledge, biology diagram comprehension, geoscience diagram comprehension, and one or more of three spatial ability measures: Mental Rotation Test, Paper Folding Test, and/or Hidden Figures Test. All spatial measures in prior research have shown correlations > .25 with diagram comprehension and also have excellent reliability (>.90).
Intervention Materials consisted of one of three workbooks, which students completed individually during class time with teacher scaffolding, followed by class-wide discussion. Topics were structure and bonding in DNA and other molecules and energy and molecular cycles though ecosystems.
After parent/guardian consent, student assent, and demographics, we pretested students and the intervention was implemented over 6 weeks (5 min / class workbooks and 5 min discussion).

5) Results
In every study, we found spatial x treatment interactions, but these were different across sites, interventions, and spatial measures (See Table 1).
6) Significance
In the present study, we have replicated Hoffler’s (2010) ability-as-compensator findings, but also complicated them. Consistent with the intervention-as-compensator hypothesis, the interactions vary depending on which treatment students are in; the interactions also vary according to which spatial measure and dependent variable are considered.

Authors

• Jennifer G. Cromley, Temple University

• Nora Newcombe, Temple University

• Theodore W. Wills, Temple University