Effective Use of Multiple Representations by Novices in Learning Chemistry
David Corradi
Jan Elen
Geraldine Clarebout
Katholieke Universiteit Leuven
Abstract
Theories on learning with multiple external representations (MER) claim that novices in science have difficulties using MER (e.g., Ainsworth, 2006). Nevertheless such MER are argued to be essential to understand scientific concepts. In contrast, some research does find that novices can use specific combinations of representations for learning (e.g., Carney & Levin, 2002). In this study we investigate the effect of MER on novices’ learning gains in chemistry. Undergraduates (n=67) participated in a pre-post randomized experiment. Participants read texts, that were – depending on the condition – accompanied with symbols and/or pictures. Result found no significant differences in learning gains. Combining symbols and/or submicroscopic representations with text does not help novices learning with MER.
Aims
Understanding scientific concepts entails the ability to recognize and switch between multiple external representations (MER) of that concept (e.g., Rappaport & Ashkenazi, 2008). We focus here on texts, symbols (both descriptive representations) and submicroscopic representations (depictive representations). Symbolic representations are part of the systematic language of chemistry that communicates for example the composition of matter. Submicroscopic representations provide information at the level of invisible and untouchable particles (Johnstone, 2006).
Comparisons between novice students and more knowledgeable students show that novices tend to have difficulties understanding and translating between MER (e.g., Seufert, 2003). If learning with MER is to be successful, research concludes, prior knowledge of the domain and its representations is required (Ainsworth, 2006). However, other research found successful use of MER with novices. The main difference was that text was the main representation and other descriptive and depictive representations (illustrations) had a supportive function (e.g., Levie & Lentz, 1982).
We want to know whether combining symbolic and submicroscopic representations with texts of chemistry increases the level of conceptual understanding for novices (i.e., chemical literacy; Shwartz, Ben-zvi & Hofstein, 2006). Adding representations to texts has a positive effect for novices on learning because it supports memory, guides attention, organizes information and structures learning behavior (e.g., Carney & Levin, 2002). Additionally, combining descriptive and depictive representations benefits learning since both are processed differently. Depictions interact with the picture’s internal perception and the prior knowledge of the subject matter. A description’s surface structure is first mentally represented. Its semantic content then generates a propositional representation. That representation is formed into a mental model (Schotz & Bannert, 2003).
Therefore we think that learner’s chemical literacy increases significantly more when learning with texts, symbols and submicroscopic representations compared to when learning with only texts or with text and either a symbolic or a submicroscopic representation.
Methods
Participants
Participants were sixty seven undergraduates. 7% were male. Mean age was 18.61 (sd= 0.91).
Design
In a pre-post randomized experiment, participants were distributed over four groups. Group 1 (n=16) received texts about basic chemical concepts. Group 2 (n=17) received symbolic, group 3 (n=17) submicroscopic representations and group 4 (n=17) got both symbolic and submicroscopic representations about each concept (always with text; representations: Lagassé, 2007).
Instruments and procedure
Pre and post-tests were chemical literacy tests. One instrument in each test assessed recognition of chemical concepts (nominal literacy; self-report) and the second one the ability to correctly describe concepts (functional literacy; open questions) (Shwartz, Ben-Zvi, & Hofstein, 2006). The nominal literacy pre-test had Cronbach α =.89; post-test: α =.91. Coders scored the functional literacy test. Inter-rater agreement (kappa) scores for all tests were between .60 and .90. Coders compromised on the scores they did not agree on. Pre and post-test lasted 10 minutes each. The intervention was the assignment to read five texts (206 words average) on a computer. Only one representation was shown at the time and actions were logged. We used an ANOVA to measure difference in learning gains (posttest minus pretest results).
Results
Pre-test ANOVA did not find a difference between the conditions, F(3,63)= 0.069, p>.05, partial η² = 0.003 (i.e., successful randomization). Our hypothesis was that the condition with three representations per concept would have more learning gains compared to the other groups. Table 1 shows a higher mean for group 4. The ANOVA reveals no significant differences between the four conditions, F(3, 63) = 0.949, p= .42, partial η² = .043. Hence, our hypothesis that MER help learning is falsified. There is also no evidence that using MER negatively affects learning.
Table 1
Means and SD for learning gains (%).
conditions / Mean / SDtext + image + symbol / 10.73 / 14.72
text + image / 7.90 / 12.62
text + symbol / 6.56 / 11.77
text / 3.09 / 13.63
Discussion
That we found no significant difference between the groups can mean two things. First, combining text with other representations does not help novices more than when text is given without representations. Participants had low prior knowledge and this may have limited them in using the representations (Ainsworth, 2006). Second, even though some state that MER may cause negative learning gains (e.g., Schnotz & Bannert, 2003), using text combined with other representations does not lead to negative learning gains here. Based on these results we assume that representations had either a decorative function for learners (i.e., redundant) (Winn, 1993) or were ill conceived.
The value of these results is that it has shed light on an ambiguity of the literature concerning novice learning with MER. Conceptual understanding did not increase more than just with text. Hence, combining text with other representations does not seem to help novices make connections between them.
References
Ainsworth, S. (2006). DeFT: A conceptual framework for considering learning with multiple representations. Learning and Instruction, 16, 183–198.
Carney, R. N., & Levin, J. R. (2002). Pictorial illustrations still improve students’ learning from text. Educational Psychology Review, 14(1), 5–26.
Johnstone, A. H. (2006). Chemical education research in Glasgow in perspective. Chemistry Education Research and Practice, 7(2), 49–63.
Lagassé, P. (Ed.). (2007). The Columbia Encyclopedia. NY: Columbia University Press.
Levie, W. H., & Lentz, R. (1982). Effects of text illustrations: A review of research source. Educational Communication and Technology, 30(4), 195–232.
Rappoport, L. T., & Ashkenazi, G. (2008). Connecting levels of representation: Emergent versus submergent perspective. International Journal of Science Education, 30(12), 1585–1603.
Seufert, T. (2003). Supporting coherence formation in learning from multiple representations. Learning and Instruction, 13, 227–237.
Schnotz, W., & Bannert, M. (2003). Construction and interference in learning from multiple representation. Learning and Instruction, 13, 141–156.
Shwartz, Y., Ben-Zvi, R., & Hofstein, A. (2006). The use of scientific literacy taxonomy for assessing the development of chemical literacy among high-school students. Chemistry Education Research and Practice, 7(4), 203–225.
Winn, W. (1993). Perception principles. In M. Fleming & W. H. Levie (Eds.), Instructional Message Design (pp. 55–126). NJ: Educational Technology Publications.
5