Vocabulary acquisition is the backbone of second language learning. If a student uses the correct words with the wrong grammar, he or she will probably still be able to convey the message (Folse, 2004). Therefore, students and teachers of ESL or EFL are constantly searching for the best ways to learn vocabulary faster and retain it in long-term memory. Definitions, pictures, and translations are three of the most common ways to present and practice unfamiliar vocabulary. Definitions in the target language are very popular in ESL teaching, pictures are widespread in children’s or beginners’ classes, and translation is heavily relied on in many EFL settings. The present study aimed to test and compare the effectiveness of these three techniques, with the assumption that all three of them would produce optimal learning effects if applied at the same time.

Paivio’s Dual Coding Model

Paivio’s (2006) dual coding theory suggests that nonverbal stimuli, such as pictures and images, support verbal memory, including bilingual memorization of vocabulary. The stimuli enter the brain’s processing systems along two parallel channels, which, however, may also compete with each other (see Figure 1):

Figure 1. The two parallel channels of Paivio’s (2006) Dual Coding Model.

Kroll’s Bilingual Model

Kroll’s (1993) bilingual model represents the different ways in which less fluent and more fluent second language speakers access concepts through their native or second language. It can be seen from Figure 2 that more concepts are easily accessible through the native language, and this happens especially at low or intermediate levels of second language proficiency:

Figure 2. Kroll's (1993) Bilingual Model

According to the models in Figures 1 and 2, images and using the native language can help students learn vocabulary and retain it in their long-term memory. This study examined how memory works in vocabulary acquisition, in particular, and whether definitions alone or when accompanied with images and translations better aid intermediate ESL learners in remembering English vocabulary. A quasi-experimental research design exposed the same class of 19 high-intermediate intensive English program students (Chinese, Arabic, and Japanese) in the United States to three different vocabulary treatments—the first explained and practiced vocabulary using definitions only; the second definitions and pictures; and the third definitions, pictures, and translations. The means and standard deviations of participants’ pretest, immediate posttest, and delayed posttest scores were compared to find out if a combination of definitions, pictures, and translations would be the most effective way to present and practice new vocabulary.

Research Questions

Are there significant differences in the effectiveness of teaching vocabulary using definitions only; definitions and pictures; or definitions, pictures, and translation?

Are there significant differences in the short-term and long-term retention of vocabulary taught using definitions only; definitions and pictures; or definitions, pictures, and translation?

Procedures

Pretests
Immediately before each treatment, the students were asked to complete, individually, a crossword puzzle consisting of 10 level-appropriate target vocabulary items and three or four distractor words, which were not part of the target vocabulary list (see Appendix A).

Treatment
The above-mentioned class (listening and note-taking class) met twice a week, on Tuesday and Thursday for 75 minutes. During three consecutive weeks, every Tuesday the students received one of three different types of treatments (see Appendix B): the first presenting and practicing vocabulary in context with definitions; the second adding images to the context and definitions; and the third adding translation to the context, definitions, and pictures. The target vocabulary words, definitions, pictures, and translations were cut up into cards and the students matched them, working in groups of three or four. The teacher walked around the room, observing the students as they participated in their groups. She answered students’ questions and provided clarifications and explanations to help students identify the correct pictures and translations when necessary. This activity was seen as presentation of new vocabulary, so the teacher helped a lot, but also let some students who already knew or had guessed the meanings of new words explain them to their group. The practice context consisted of academic lectures including the target vocabulary. After the matching activity, the class listened to a recorded lecture containing the vocabulary, took notes, and answered comprehension questions.

Immediate Posttests
Two days after each treatment, on Thursdays, students were asked to complete a different crossword puzzle consisting of the target vocabulary.

Delayed posttests
Three weeks after each treatment, students were asked to complete yet another crossword puzzle containing the target vocabulary. They were told it was a review that would not be graded.

Data Analysis

The scores of the nine tests (three pretests, three immediate posttests, and three delayed posttests) were tabulated for each of the 19 participants. Mean scores and standard deviations were calculated for each test. The scores were also entered in the statistical software program PASW, Predictive Analysis Software (formerly known as SPSS, Statistical Package for the Social Sciences) and a one-way repeated measures ANOVA, Analysis of Variance, as performed to see if there were significant differences among the different treatments and the timing of the tests.

Results

The nine tests that the students took before and after each of the three treatments (teaching 10 vocabulary items explained first only in definitions; then, teaching 10 different vocabulary items with definitions and pictures; and finally, teaching 10 more vocabulary items using definitions, pictures, and translations in matching activities) were scored. The mean scores and standard deviations for all tests and treatments are presented in Table 1and Figure 3.

Table 1. Test Results

As can be seen from Table 1, the students generally did not know most of the three sets of 10 target words before they were first introduced in class, as the mean scores of all three pretests are very low, between 0.25 and 2.50, with relatively low standard deviations, between 0.50 and 0.82. Surprisingly, the best scores on the immediate posttests, with a mean score of 8.50 and the low standard deviation of 1.73, were achieved after using only definitions in English to teach the new vocabulary. When definitions and pictures were combined, the mean score on the immediate posttest was much lower at 4.25, with a higher standard deviation of 3.10, similar to when definitions, pictures, and translation were presented simultaneously (mean score 4.75 on the immediate posttest with standard deviation of 3.69). On the delayed posttests, the students also did better when vocabulary was taught using only English definitions, even though they forgot many of the words they knew on the immediate posttest, as the mean score went down to 5.00 with standard deviation of 2.16. The use of definitions and pictures together yielded a slightly lower mean score of 4.25 on the delayed posttest, with standard deviation of 2.87. However, the application of definitions, pictures, and translations together resulted in the lowest delayed posttest score of 0.75, which is lower even than the pretest, before any teaching had occurred. (See the bar graph in Figure 3 for a better visual representation of the results).

Figure 3. Test Results

A repeated-measures one-way ANOVA was run in PSAW statistical software. The results were computed using alpha = 0.05. As for the treatment, there were no significant differences among the effectiveness of the three treatments (sig = 0.297) but for the second factor, time, there were significant differences among the scores on the pretests, immediate posttests, and the delayed posttests (sig = 0.037).

Discussion

The results of the tests for the first and second vocabulary sets show a pattern. Students got the highest scores in the immediate posttests and the lowest scores in the pretest. The scores for the delayed posttests were lower than the immediate posttests but higher than the pretests. This pattern was expected, because students should be able to remember more words immediately after the treatment. Two weeks after the treatment, students forgot some of the words learned, as they did not revise or do any practice with the words between the treatment and the delayed posttest.

However, the test results of vocabulary list 3 show a different pattern. The students still got the best results for the immediate posttest, but did worse in the delayed posttest. The mean score of the delayed posttest was very low (0.47), much lower than the pretest (1.38). Many of the test papers were turned in blank. One student received two points, and nobody scored higher than that. In the pretest, three students were able to obtain three points, which was the highest score. As for the number of students who got zeros, there were five in the pretest but ten in the posttest.

In general, the overall results contradicted the hypothesis that vocabulary is learned more effectively with images and translations. The students performed better in both posttests for word list 1, which was taught with definitions only. The students continued to do badly in the following tests, which show that the images and translations did not help them to learn the words. On the contrary, it appears that they even had adverse effects on students’ vocabulary learning.

As Paivio’s (2006) dual coding model might suggest, perhaps the pictures and the definitions, which were both in print, competed for the students’ attention. Pictures may also be confusing, especially for more abstract words: for instance if a smiley face is used to illustrate the word happy, students might think that it means smile. The use of pictures, however, seemed to help improve long-term memory as the results of this particular immediate posttest and delayed posttest were almost exactly the same. In other words, the students did not learn more words with pictures and definitions than with definitions alone, but whatever they learned seemed to stay in long-term memory without any further practice. Therefore, pictures may be helpful only after the meaning of a new word has already been clarified in a different way, and then a supplementary image association may aid memory and long-term retention.

On the other hand, when translations were added to the definitions and pictures, the students did the worst. One possible reason could be that they were matching words in their native language to images, as one student reported doing, completely leaving out the English vocabulary and definitions. Again, perhaps the three different techniques again competed for the students’ attention and cognitively overloaded them. Translation may be a good, efficient technique on its own, but not combined with definitions in the target language and pictures at the same time.

Conclusion

In conclusion, simpler and shorter explanations seemed to be the best formula for successful vocabulary acquisition. Certainly more research is needed in this most fundamental, indispensable, and to some extent neglected area of second and foreign language learning.

REFERENCES

Folse, K. (2004). Vocabulary myths: Applying second language research to classroom teaching. Ann Arbor: The University of Michigan Press.

Kroll, J. F. (1993). Accessing conceptual representations for words in a second language. In R. Schreuder & B. Weltens (Eds.),The bilingual lexicon (pp. 54–81). Amsterdam/ Philadelphia: John Benjamins.

Paivio, A. (2006). Mind and its evolution: A dual coding theoretical interpretation. Mahwah, NJ: Lawrence Erlbaum Associates.

Appendix A
Sample Crossword Puzzle Test (click to enlarge)

Appendix B
Sample Treatment Material (click to enlarge)