Perfecting language: experimenting with vocabulary learning.

Absalom, Matthew

ABSTRACT

One of the thorniest aspects of teaching languages is developing students' vocabulary, yet it is impossible to be 'an accurate and highly communicative language user with a very small vocabulary' (Milton, 2009, p. 3). Nation (2006) indicates that more vocabulary than previously thought is required to function well both at spoken and written discourse levels. With the recent spread and uptake of Language Perfect--'used by more than 1,000 schools around the world' (EducationPerfect, 2014), this paper reports on a trial study to explore how different approaches to vocabulary learning might affect vocabulary acquisition.

KEY WORDS

vocabulary acquisition, Italian, French, computer assisted language learning

BACKGROUND

Almost 20 years ago, Long and Richards (1997, p. ix) bemoaned the fact that '[t]he relative neglect of studies of vocabulary acquisition and related areas of lexical research in second language acquisition has often been commented on within the fields of language teaching and applied linguistics'. Thankfully, more recently Liz Giltner (2012, p. 169) notes that '[i] n the past two decades, research about learning and teaching second-language (L2) vocabulary has increased greatly'. Of note is the sustained contribution to this area of our understanding by Paul Nation and colleagues. Nonetheless, Nation and Webb (2011, p. 15) note that our knowledge of vocabulary teaching techniques remains 'a surprisingly under-researched area'. Between 2011 and 2013, I was once again confronted with designing and teaching a course for beginning students of Italian which stimulated a reflective approach to dealing with vocabulary teaching--I had been teaching only advanced students of Italian in the six years prior and, thus, the issue of vocabulary learning was of a different order. Turning to the scholarly literature on vocabulary teaching and learning for inspiration I was somewhat taken aback that there were not more experimental studies detailing successful approaches. At the same time, I was aware of the growth of LanguagePerfect, '... Australasia's most popular web-based vocabulary learning tool ...' (LanguagePerfect, 2014), in school language programs in Australia and internationally. Consequently, this consideration led to the design of an experimental comparison of vocabulary learning activities (see Nation and Webb, 2011, pp. 17-20), discussed in this article.

A further stimulus behind a targeted approach to vocabulary learning activities relates to the impact of the communicative language orthodoxy in our classrooms. Milton (2009, p.3) notes that due to the prevailing approaches to languages teaching and learning there has been a 'reduction both in the volumes of vocabulary presented to learners and in the volumes of vocabulary learned'. Indeed, with the progressive shift away from explicit teaching and the poor reputation of any learning activities that resemble rote learning or memorisation, approaches to teaching vocabulary need to be reviewed.

[ILLUSTRATION OMITTED]

[ILLUSTRATION OMITTED]

Experimental design and previous research

In their seminal volume, Research and Analyzing Vocabulary, Nation and Webb (2011, p.17) comment that '[w]hen we think of ways of deciding whether one vocabulary learning activity is likely to be more effective than another, our first thought may be to try them both out and see which one works best. If we want to do this in a careful way, then we can design an experiment'. This is the approach that was taken in exploring the impact of LanguagePerfect. As per Nation and Webb's conventional experimental design, utilising an independent variable, two groups were used to test vocabulary learning in different ways, with results for the two groups compared after testing. The experimental design followed very closely that identified by Nation and Webb, differing only in the number of treatments (see figure 1).

[FIGURE 1 OMITTED]

The research question was how is student learning affected by different approaches to vocabulary learning? The experiment took place over a four-week period in two beginners language subjects--French 1 and Italian 1--at The University of Melbourne in Semester 1,2013.The experiment compared the use of LanguagePerfect with traditional pen-and-paper list approaches to vocabulary learning. As noted above, due its growing and widespread uptake in schools, LanguagePerfect was considered an appropriate online tool for examination, and there was also a level of curiosity about its effectiveness that the researchers were keen to consider, to provide some preliminary data to language teachers considering using this resource. At the time of the experiment, there was a broad range of preloaded vocabulary lists in the LanguagePerfect tool, and also the capacity to customise lists to suit particular teaching contexts and content determined by the teacher. For this experiment we were able to have lists of vocabulary based on the respective subjects' textbooks created for use by our students. Aligning the vocabulary to be studied with the textbook and class activities addresses the students' need to learn given words/expressions. 'Need' is one of the three essential factors of Laufer and Hulstijn's (2001, p. 3) involvement load hypothesis, providing '[t] he best-known and best-researched way of analysing vocabulary teaching techniques' (p. 3). The other two factors are 'search' and 'evaluation'. Each factor may be absent (-), or present, in moderate (+) or full (++) strength. The total of the strengths of the three factors reflects the involvement load of the task. While Nation and Webb (2011, p.3) suggest that 'the greater the involvement load, the better the learning', they also indicate that 'other factors like time on task and repetition need to be considered' (p. 7).

In this paper, we will detail initial results of the comparison of four vocabulary teaching approaches. This was a blind experimental design in the sense that the students involved were unaware that they were participating in the project. Under The University of Melbourne's prevailing human ethics protocols, explicit approval is not required in the case of '[u]ndergraduate projects with an education, training, or a practical experience focus' (OREI, 2011). Selection of treatment groups was made from the existing tutorial groups in French 1 (13 groups to choose from) and Italian 1 (5 groups to choose from). Both tutorial group selection and assignment of treatment occurred using random selection, using the random number generator available at random.org. Tutorial groups were selected to ensure that each group would have a similar number of participants (around 20).

The four treatments were as follows:

* Custom: customised content (textbook and top 200 words) using LanguagePerfect

* Online: access to preloaded content and using LanguagePerfect

* Pen: vocabulary lists distributed weekly for four weeks (textbook and top 200 words)

* Control: normal instruction, without any provided vocabulary lists or tools.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

The lists of the top 200 words for each language were developed using frequency dictionaries: A Frequency Dictionary of French: Core Vocabulary for Learners (Le Bras & Lonsdale, 2009), and Lessico di frequenza dell' italiano parlato (De Mauro & Mancini, 1994). As indicated previously, vocabulary lists were also drawn from the relevant chapters of the subject textbooks, which, at that time, were Contatti 1 (3rd edition) (Freeth & Checketts, 2011) for Italian, and vAn introduction to French (5th edition) (Jansma & Kassen, 2011). Students in the LanguagePerfect groups were given unique login credentials using their institutional email addresses. The pen groups received their weekly vocabulary list in the first class of the week in hard copy and were also emailed the digital copy of the list. The control groups had no specific intervention in relation to vocabulary other than the usual teaching program.

A pre-test based on the total body of vocabulary was administered to all four groups. An identical post-test was undertaken by all Italian participants. Additionally, all groups sat a weekly 'pop' test. The final exam for Italian was analysed as a delayed post-test. In the following section, the preliminary findings based on the pre-tests and pop tests in both languages as well as the post-test for the Italian group re discussed.

Initial findings: French

Figure 2 shows the mean percentage of correct responses to the pre-test and the three pop tests by treatment group. What is notable is the clear change in performance of the custom group. All groups peak in the initial stages of the four-week project with pen, online and custom showing almost no increase in performance over the period of the experiment. Only the custom group demonstrates a clear change. Unfortunately, due to administrative hindrances, we were unable to administer either of the post-tests to the French group. Nevertheless, these preliminary results appear to indicate that students with online access to customised vocabulary showed a sustained increase in vocabulary retention.

Initial findings: Italian

If we compare the French findings (Figure 2) with the Italian results (Figure 3), we immediately note some differences. Firstly, for Italian, we have practically the full complement of tests and, again, one group stands out from the rest. Of note, is that all groups show improvement over time in the Italian case

Comparing the boxplots of change by group (Figure 4) based on the results of the pre-and post-tests, it is clear that the pen group is distinct from the other three treatments.

An analysis of variance (ANOVA) to compare the mean change in percentage score from pre to post between the four groups reveals that there were statistically significant differences between the mean scores across the groups (p=0.003) with the custom group being lower than the other three groups. While both LanguagePerfect groups showed improvement, the pen group performed the best.

Discussion and implications

It is not unusual to find outcomes to research that are less than black and white when it comes to the application of technology to language teaching and learning. For instance, in Oberg's (2011) comparison of Computer Assisted Language Learning (CALL) and flashcard approaches to vocabulary learning, he concludes that there were no significant differences between the two approaches. Similarly, Bagheri, Roohani, and Ansari (2012) conclude that both CALL-based and non-CALL-based methods of vocabulary teaching lead to similar outcomes. Motivationally, however, Oberg (2011) notes that there was a 'slight preference' (p. 118) towards a technology mediated approach. The trend in our French findings would be consonant with the claims of these authors: all groups performed in a similar manner but there may have been a positive effect from the use of the online environment with the custom group.

The Italian findings are, perhaps, slightly counterintuitive given that the pen group showed most gains overtime. Arguably, this is due to two interrelated factors. The pen group had the direct intervention of the instructor each week in relation to the distribution of that week's vocabulary list. This may have created a sense of prestige or importance around the vocabulary task. With the online groups (and the custom group), no such situation was created. While students were encouraged to go online, the same degree of in-class engagement was absent. The online groups, then, relied on student autonomy. We are, therefore, hypothesising that to some extent the instructional setting of the experiment had a positive effect on the pen group. Overall, however, we would claim that this experiment (at least for the Italian group) indicates that any type of programmatic intervention in relation to vocabulary teaching and learning will lead to positive change over time. The fact that even the custom group showed improvement lends strong support to this notion.

Based on these findings, we cannot say conclusively that LanguagePerfect led to improved vocabulary learning but there are clearly some positive findings in favour of an increased explicit focus on vocabulary. Informal feedback from students regarding the use of LanguagePerfect was very positive--indeed, a number of students of Italian were very keen to continue using the online platform for the remainder of the year.

Further research, and over longer timeframes, and taking into account other factors such as student-expressed motivation, and student responses to different vocabulary learning approaches should yield more informative data to build on this preliminary experiment.

Acknowledgements

I would like to thank the staff of LanguagePerfect, in particular Scott Cardwell and Tania Christie, for their professional and generous support of this project. We were provided with free access for our students and LanguagePerfect very kindly created a raft of customised vocabulary for the study.

This project would not have been possible without the collaboration of my colleagues in the French Studies Program, particularly Andrew McGregor. Invaluable statistical analysis was provided by Sandy Clarke of The University of Melbourne's Statistical Consulting Centre. This project was supported by funding from the Faculty Research Grants Scheme of the Faculty of Arts, The University of Melbourne.

References

Bagheri, E., Roohani, A., & Ansari, D.N. 2012. Effect of CALL-based and Non-CALL Based Methods of Teaching on L2 Vocabulary Learning. Journal of Language Teaching and Research, 3, 4, 744-752.

De Mauro, I & Mancini, F 1994. Lessico di frequenza dell'italiano parlato. Milan: Etaslibri.

EducationPerfect. Languages. Retrieved 30 August 2014 from http://worldseries.educationperfect.com/ languages.html

Freeth, M. & Checketts, G. 2011. Contatti 1, 3rd ed. London: Hodder.

Giltner, L. 2012. Review of Research and Analyzing Vocabulary. TESL Canada Journal/Revue TESL du Canada, 30, 1, 169-170.

Jansma, K. & Kassen M.A. 2011. Motifs: An Introduction to French, 5th ed. Boston: Heinle Cengage Learning.

LanguagePerfect. Reviews. Retrieved August 30 2014 from https://start.languageperfect.com/reviews.html

Laufer, B. & Hulstijn, J.H. 2001. Incidental vocabulary acquisition in a second language: The construct of task induced involvement. Applied Linguistics, 22, 1-26.

Le Bras, Y. & Lonsdale, D. 2009. A Frequency Dictionary of French: Core Vocabulary for Learners. London: Routledge.

Long, M.H. & Richards, J.C. 1997. Series editors' preface. In J. Coady &T. Huckin, Second language vocabulary acquisition: A rationale for pedagogy, ix-x. CUP.

Milton, J. 2009. Measuring second language vocabulary acquisition. London: Multilingual Matters.

Nation, I.S.R 2006. How large a vocabulary is needed for reading and listening? The Canadian Modern Language Review, 63, 59-82.

Nation, I.S.R &Webb, S. 2011. Researching and Analyzing Vocabulary. Boston: Heinle, Cengage Learning.

Oberg, A. 2011. Comparison of the Effectiveness of a CALL-Based Approach and a Card-Based Approach to Vocabulary Acquisition and Retention. CALICO Journal, 29, 1, 118-144.

OREI [Office of Research Ethics and Integrity]. 2011. When is approval needed? Retrieved 31 May 2011 from http:// www.orei.unimelb.edu.au/content/when-approval-needed

Matthew Absalom is a university teacher and researcher, linguist, Italian language coach, translator and published author. His current appointment is in the Italian Studies program at The University of Melbourne. He holds qualifications in music, education, languages and linguistics, and his research interests cover Italian linguistics, computer assisted language learning, and languages education. His university career in Australia spans three universities: the Australian National University, University of South Australia and The University of Melbourne. He is currently the immediate past president of the Australian Federation of Modern Language Teachers Associations.

Caption: Figure 1: Experimental design after Nation and Webb (2011)

Caption: Figure 2: Line plot of percentage correct over time by group French.

Caption: Figure 3: Line plot of percentage correct over time by group Italian.

Caption: Figure 4: Boxplot of change (post-pre) by group