Bahr, S & Dasereau, D. (2005). Bilingual knowledge maps (BiK Maps) as a presentation format: delayed recall and training effects. Journal of experimental Education, 73(2), 101-118.
Problem/Purpose
This article was selected because it represents the latest study in an almost decade long bilingual knowledge map research by the same authors. The focus of research within the bilingual language map domain is two-fold. The first is to develop new strategies that help learners of foreign language form the so-called lexical association between native and foreign language words. The second is to make semantic foreign language concepts more accessible to novice learners with highly limited vocabulary and almost nonexistent knowledge of grammar.
Previous research on monolingual as well as BiK maps suggest that BiK maps are likely to create a very positive learning context by allowing for additional tolerance for the lack of basic foreign language vocabulary and grammar and by adding both multidimensional pictorial and visio-spatial information.
The purpose of this study was to investigate two factors related to BiK map research, the presentation format, meaning the use of either BIK maps or standard word pair list, and training on how to learn using BIK maps or word-pair lists. Within these two factors, based on previous studies by the authors, the following three objectives were explored: (1) to test the prediction that when learners’ recall s is tested following a two day delay without prior immediate recall, BiK map learners will show greater recall than word-list learners (2) to asses recall strategies and the use of cues using processing inventory of vocabulary recall (PI-VOR), and (3) to asses the training effects on both BiK map and word-list trainees performance using both PI-VOR and processing inventory of behavioral intentions for textbook study (PI-BIT).
Methods
Bahr and Dansereau recruited 82 undergraduates from several psychology classes from an unknown university to participate in this study in exchange for academic credit. The participants were treated according to guidelines of the American Psychological Association and were randomly assigned to one of the four groups. The study consisted of three sessions ranging from 35 minutes to 2 hours and was conducted within a seven day period on Monday, Wednesday and Friday. In session 1 participants were randomly assigned to conditions where their previous German language knowledge, their motivations were assessed, then received either BiK map or traditional word-list training.
In session 2, both groups warmed up by using Hindi and English word-pairs and then were trained with 16 word sets of German English word-pairs with an average of 6.65 letters per foreign word reflecting tourist vocabulary. In session 3, the authors collected memory measures on both free and cued recall, conducted two additional personal inventory post-tests and ended the session with a debriefing of the participants.
Results
Bahr and Dansereau conducted four 2 X 2 multiple analyses of covariance (MANCOVA) primary analyses on four sets of dependent measures. The first set consisted of the memory measures, which included the free recall scores
FR-wp and FR-add) and cued recall scores. The second, third, and fourth sets consisted of the personal inventory (PI) factors retained from the PI-VOR, PI-BIV, and PI-BIT, respectively.
In the memory measure (Training X Presentation) format MANCOVA on the memory measures revealed a significant effect for study presentation format, F(3, 75) = 5.16, p < .002. The authors found the univariate effects on all memory measures: cued recall, F(1, 77) = 4.78, p < .031; FR-wp, F(1, 77) = 6.61, p < .012; and FR-add, F(1, 77) = 13.57, p < .001.
The adjusted means of participants who studied BiK maps were higher than those of word-list students: cued recall, 9.03 vs. 6.47, respectively; FRwp, 3.47 vs. 1.87, respectively; FR-add, 5.55 vs. 2.90, respectively ( see Table 5 below). The standardized mean differences were dCuedRecall = .497, dFR-wp = .565, and dFR-add = .826. Here the authors found that no other effects, including set effects, were significant.
In the PI-VOR. (Training X Presentation) format MANCOVA on the two PIVOR measures revealed a significant effect for presentation format, F(2, 76 ) = 7.43, p < .001, which was reflected at the univariate level as a main effect for
the Breadth Processing factor, F(1, 77) = 14.40, p < .001. Here the authors discovered that the adjusted mean score was greater for BiK-map students than for the word- list students (4.51 vs. 3.49, respectively; d = .847).
MANCOVA analysis on the two PI-BIV measures revealed a main effect for training approaching significance, F(2, 76 ) = 3.08, p < .051. At the univariate level, this effect was apparent in the Surface Processing factor, F(1, 77) = 5.20,
p < .025. Here the authors found word-list trainees indicated greater intentions than BiK map trainees to engage in this mode of processing (5.63 vs. 5.08, respectively; d = .495).
MANCOVA analysis on the PI-BIT measures revealed no significant differences. However, given the exploratory nature of the study, a univariate finding for training is reported with the Theme (Breadth) Processing factor, F(3, 75) = 4.33,
p < .040. Here the authors found that BiK map trainees indicated greater intent than word-list trainees (5.57 vs. 5.13, respectively; d = .453).
Discussion
Bahr and Dansereau state that the general results of the study are comparable with others obtained by previous foreign language acquisition researchers. This means that large standard deviations and relatively low recall scores and correlation coefficients ( in this study Cronbach’s alpha is ranging from .613 to .969) characterize it. In this respect the authors concluded that their experimental data provided support for the argument that BIK map-based learning does have significant advantage over word-list learning, in particular within short terms. However, Bahr and Dansereau also acknowledged that students’ performance in the present study was significantly lower (ranging from 10.8% to 28.8% in the BiK group and 5.8% to 20.2% in the list group) than those of students from earlier studies by the same authors in 2001.
The authors state that the difference in the students’ performance in the present study is likely due to the fact that delayed recall in previous experiments was preceded by immediate recall, which apparently caused a rehearsal effect. Although their objective in the current study was to eliminate the possibility of differential rehearsal effects, it seems that, without rehearsal, a consequent suppression of memory in general is unavoidable when testing for BiK-map effects.
The authors also discussed the possibilities of taking BiK map out of the academic realm in order test the durability of findings in realistic settings and expressed hope That future investigation into foreign language learning connectivity and multilingual memory will lead to the development of more sophisticated learning materials.
Critique
Bahr and Dansereau’s research paper, although relatively well written, is presented in a nontraditional format that can be hard to follow as both the result and discussion sections are found at the beginning of the paper. The article could have been improved by providing additional information about the location of the study and by limiting references to the authors themselves as their work has been replicated by others on numerous occasions. Other areas of weakness include the very low Cronbach’s alpha values in the processing inventory section on page 107-108 and the very low number of participants (well below the usual 300+ that is recommend for factor analysis). In spite of these critiques and in light of the experimental nature of this investigation I conclude that the Bahr and Dansereau article is both clear written and its aim is well directed.
Theoretical and practical implications
One of several theoretical implications of this study is that it highlighted the need for additional investigation to better understand the connections between of BiK maps and the foreign and the native language lexica. Although there is previous research (see page 102) that seems to support the hypothesis that a strong semantic or conceptual association exists between foreign language words and their meaning, future research should not only confirm these hypothesis’ but should test new presentation formats that would facilitate these connections in novice learners.
Practical implications of this study include using both BiK and monolingual knowledge maps in all kinds of academic settings in order to promote a shift toward a more beneficial top-down processing strategy among the learners. The use of both bi and monolingual knowledge maps in various settings also raises the possibilities that the textbook publishers would include BiK supplements in their new editions of foreign language textbooks and that BiK map creation as a subject would be added to teachers of foreign language degree course offerings.
