English language proficiency and the effects of rewording semantic structures on P5 pupils' mathematical word problem solving

In our local primary mathematics curriculum, the solving of mathematical word problems is a major component both within the instructional programme as well as during formal assessments. It has been researched that both language and semantic structures play a part in determining pupils' performance in the solving of mathematical word problems, hence the interest in this study which aimed to determine if pupils' performance were affected through the rewording of mathematical word problems. The results of the findings would be useful for teachers to provide a course of action in enhancing pupils' comprehension and thus the solving of mathematical word problems. This study was based on a quasi-experimental approach, whereby a counter-balanced repeated measure design was used. Eighty-three Primary 5 EM2 pupils from two mixed-ability classes participated in this study. They were grouped into Good Performers (GP) and Marginal Performers (MP) based on their English Language performance in the P4 Streaming Examination (Band 2 and Band 3 respectively) for the purpose of determining whether their performance would differ in solving the 10 mathematical word problems used in this study. These 10 word problems comprised 5 original word problems and 5 reworded (reworded from the original and deemed to be parallel to the original) word problems and are termed as OWT (Original Word Problem Test) and RWT (Reworded Word Problem Test) respectively. The problems in the RWT were crafted based on 4 types of semantic structures, namely, "Chronological Order of Events", "Personalization", "Chunking" and "Repositioning the Givens". To determine if the semantic structures of the word problems would affect pupils' performance, the pupils were re-grouped into Low Achievers (LA), Average Achievers (AA) and High Achievers (HA) based on the results obtained from the OWT and then compared against their respective performance in the RWT. The findings revealed that there was no significant difference between the GP and MP groups of pupils in their attempt at the 10 word problems. This showed that language proficiency did not affect their performance. Rewording of the mathematical word problems based on semantic structures enabled the pupils to perform better as seen in the improvement in mean scores for the overall RWT against OWT for both the GP and MP groups. With respect to the specific types of semantic structures which affected the performances of the HA, AA and LA groups, the restructuring based on "Chronological Order of Events" positively impacted the LA group significantly while the AA group was quite close to registering a significant difference. "Repositioning the Givens" also positively affected the LA and AA to a significant extent in registering a significant difference in performance in one of the problems. Qualitative analysis of samples of pupils' solutions suggested that the rewording based on the two mentioned semantic structures had enabled the pupils to have a clearer mental construct and thus made the meaning more explicit for the pupils to comprehend. Rewording through "Chunking" did not show any significant difference in the pupils' performance and samples of the pupils' solution scripts indicated no difference in methodology used for both the OWT and RWT for that particular word problem. The surprise finding was the pupils' performance in the reworded problem based on "Personalization". Contrary to other research findings on the benefits of personalization, the HA group did significantly worse and the AA group also showed a decline in their performance mean. The result was attributed to the higher difficulty level of this word problem compared to the others due to its complex algebraic structure. The other reason was that personalization was not commonly used in our local instructional programme and that the pupils could have already an "over-developed schema" to the common word problems they were exposed to in class that they tried to read too much into the reworded problem in this instance.