Computer-mediated textual and audio glosses,perceptual style and L2 vocabulary learning

Abstract

The current study examined the effects of first language (L1) textual and audio glosses along with the moderating effect of learners’ visual and auditory perceptual style on second language (L2) vocabulary learning. The participants were first divided into two groups of auditory and visual learners based on their scores obtained from a learning style questionnaire. Then, each style group was further subdivided into three groups who received a text via a screen including a number of unfamiliar vocabulary items with one of the following L1 gloss conditions: textual glosses, audio glosses or no glosses. Learners could either hear or see the L1 definitions of target words by clicking on hyperlinks provided under each target word. The results provided evidence that both gloss types promoted vocabulary learning and indicated that audio glosses were more effective than textual glosses. Moreover, the results suggested the moderating role of perceptual learning style in vocabulary learning. Among the treatment groups, the auditory learners who received audio glosses indicated the highest rate of vocabulary learning.

Keywords

Audio glosses auditory style textual glosses vocabulary visual style

I Introduction

One of the most challenging issues facing second language (L2) learners to communicate effectively in an L2 is the acquisition of L2 vocabulary. In fact, because of inadequate vocabulary knowledge many learners of English as a foreign language (EFL) have difficulty comprehending and also producing English. Due to the central role that vocabulary learning plays in L2 acquisition, investigating innovative techniques for enhancing L2 learners’ vocabulary knowledge deserves more attention. There are conflicting views regarding how vocabulary should be learned by L2 learners. It is believed that most essential vocabulary knowledge is acquired incidentally during reading (Hulstijn, 2001). However, dealing with unfamiliar lexical items while reading a text can be taxing and demotivating for L2 learners. According to Abraham (2008), guessing unfamiliar words appropriately in a text depends on several factors such as learners’ proficiency, learners’ use of appropriate strategies, the frequency of unknown words in a text and also learners’ vocabulary knowledge. One solution for this is to provide learners with different forms of glosses along with the target vocabulary. Glossing therefore provides an additional advantage for incidental vocabulary learning as learners may ignore unfamiliar words while reading (Qian, 2004) or even may make incorrect inferences and guesses regarding unknown words (Kondo-Brown, 2006).

There is now ample evidence from previous research that providing target words with glosses enhances vocabulary learning (e.g. Hulstijn, Hollander, & Greidanus, 1996; Yanguas, 2009; Yoshii, 2006). As noted by Yoshii (2006), the question nowadays is not whether glosses are effective or not but what form of glosses would be more effective for vocabulary learning. Marginal glosses associated with unfamiliar vocabulary in a text can take different forms including textual (first language and second language), graphical and auditory. Recently, researchers have also become interested in investigating computer-mediated multimode glosses that include a combination of textual and visual glosses such as pictures and videos (e.g. Al-Seghayer, 2001; Plass et al., 2003; Sydorenko, 2010; Yanguas, 2009; Yoshii, 2006). The results of a second-order meta-analysis by Plonsky and Ziegler (2016) suggested that computer-mediated multimodal glosses are more effective than traditional paper-based glosses for enhancing L2 vocabulary knowledge.

Given the increasing interest in implementing computer-mediated glosses for L2 vocabulary learning, one issue that needs to be addressed is the extent to which individual differences modulate the efficacy of different glossing modes for vocabulary learning. More specifically, little is known about the relationship between the effectiveness of different gloss types and learners’ cognitive and learning preferences. As noted by Abraham (2008), one central concern for designing computer-mediated tasks for vocabulary teaching is to understand the relationship between individual differences and the features of such instructional contexts. To shed some light on this issue, the current study aims to examine whether learners with auditory and visual learning styles benefit equally or differently from textual and auditory glosses.

1 Different forms of glosses and L2 vocabulary learning

Reading comprehension is one of the major sources of vocabulary learning for L2 learners (Hulstijn, 2001). Since learners’ processing resources are limited to coordinate lower and higher cognitive processes involved in reading comprehension, providing various forms of print-based and computer-mediated glosses can help learners manage reading comprehension and at the same time learn new vocabulary items encountered in a text. Previous studies predominantly investigated the effects of various forms of textual glosses on L2 vocabulary learning. Some studies examined the effectiveness of first language (L1) and L2 glossing conditions compared to no gloss condition. These studies provided evidence for the importance and effectiveness of glosses for incidental vocabulary learning (e.g. Hulstijn, Hollander, & Greidanus, 1996; Jacobs, Dufon, & Fong, 1994; Laufer & Shmueli, 1997; Yoshii, 2006). Meanwhile, conflicting evidence exists as to what form of glossing is more effective than others. Some studies provided evidence that L1 glosses are more effective than L2 glosses for vocabulary learning during reading comprehension (e.g. Laufer & Shmueli, 1997; Yee, 2010), while other studies indicated no significant difference between the effectiveness of L1 and L2 glosses (Chen, 2002; Jacobs et al., 1994; Yoshii, 2006). Empirical evidence also exists that L2 glosses can be more effective than L1 glosses for vocabulary learning (e.g. Miyasako, 2002). One explanation for the mixed results of previous research is the role of individual differences that can affect learners’ comprehension of a reading text or their perceptions of gloss types provided. In this regard, Miyasako (2002) reported that low proficiency learners benefit more from L1 glosses while high proficiency learners benefit more from L2 glosses for vocabulary learning.

A number of researchers also investigated the effects of L1 and L2 glosses in different forms such as sentence level glosses (e.g. Grace, 1998, 2000) single versus multiple choice glosses (e.g. Duan & Yan, 2004; Rott, Williams, & Cameron, 2002; Watanabe, 1997) and also dictionary form glosses (e.g. Gettys, Imhof, & Kautz, 2001) on L2 vocabulary learning. All these studies revealed that different forms of glossing assist vocabulary retention and also some forms of glossing can provide better results. Researchers have reported several major factors contributing to the differential effects of different forms of glossing on vocabulary learning, namely individual differences (e.g. Miyasako, 2002) and also the amount of cognitive processing triggered by different glossing formats (e.g. Gettys et. al., 2001).

Another issue that has sparked some recent studies is the use of technology for presenting glosses in different modes such as pictures and videos and also the combination of textual and graphical modes often known as multimedia glosses. The outcome of these studies suggests that a combination of glosses in different modalities is advantageous over glosses in a single modality for vocabulary acquisition (e.g. Al-Seghayer, 2001; Chun & Plass, 1996; Kim & Gilman, 2008; Kost, Foss, & Lenzini, 1999). Research has also shown that the use of technology in language classrooms has several advantages for learners including catering instruction to learners’ instructional preferences, increasing learners’ motivation and decreasing their anxiety and also providing more opportunities for interaction and negotiation of meaning (Brinton, 2001; Sydorenko, 2010).

Meanwhile, previous research often focused on the effects of different forms of single mode textual glosses such as L1 and L2 glosses (e.g. Miyasako, 2002; Yoshii, 2006) or multimedia glosses consisting of textual, graphical and audio glosses on vocabulary learning (e.g. Al-Seghayer, 2001; Kim & Gilman, 2008; Sydorenko, 2010) and therefore failed to provide evidence what specific gloss modality is more advantageous for vocabulary learning. In other words, each of the multimedia gloss conditions examined in the previous studies included several input modalities that make it difficult to discern which specific gloss modality is more beneficial for vocabulary learning. For example, little is known about whether audio glosses are more effective than textual glosses or vice versa. Moreover, one additional limitation of research on multimedia glosses is the possible interaction between learners’ perceptual style and different gloss modalities that learners experience at the same time. To shed some light on this issue, it is first important to establish whether learners benefit differentially from glosses presented in specific modalities (i.e. textual or audio) and also whether matching learners’ perceptual learning styles with gloss modalities further enhances vocabulary learning.

2 Visual vs. auditory perceptual style and learning

It is generally believed that students’ learning preferences or learning styles have considerable effects on their ultimate achievement. Oxford (2003) defined learning styles as ‘the general approaches that students use in acquiring a new language or in learning any subject’ (p. 2). According to Slack and Norwich (2007) individuals differ in the way they learn and this calls for various forms of teaching which take into account individuals differences and preferences. Krätzig & Arbuthnott (2006) defined learning styles as a combination of individuals’ cognitive and affective factors that predispose people to process information and interact with environment in different ways.

There are conflicting views regarding the benefits of matching learning and teaching style. On the one hand, some researchers and educators believe that matching teaching style with learning style has several benefits such as improving students’ achievement and also enhancing their motivation (e.g. Othman & Amiruddin, 2010; Stitt-Gohdes, 2001; Thomson, Watt, & Liukkonen, 2015). Naimie et. al. (2010) suggest that the mismatch between learning and teaching style may result in learners’ demotivation and confusion in classrooms. According to Peacock (2001), ‘when student and teacher styles are better matched, students are likely to work harder both in and outside the classroom and also to benefit much more from their EFL classes’ (p. 4). Proponents of matching teaching and learning style argue that instruction is more effective when it is provided in a format that matches students’ learning style. For example, visual learners should benefit more from visual data while auditory learners benefit more from auditory information. Based on a meta-analysis of 76 experimental studies, Lovelace (2005) reported that her findings support the hypothesis that matching students’ learning style preferences with instruction improves achievement. Meanwhile, other studies did not provide evidence for the benefits of matching learning and teaching style. For example, after reviewing about 20 studies, Rohrer and Pashler (2012) concluded that there is no compelling evidence in the literature to argue in favor of style-based instruction. According to Peacock (2001), despite a wealth of theoretical arguments in favor of matching learning style with instruction, no empirical evidence can be found to support the issue in L2 acquisition research. However, as noted by Liu and Todd (2014), the contradictory results of previous studies regarding the benefits of instruction which is catered to learners’ style preferences do not mean that style-based instruction is ineffective and instead it indicates a need for further studies to shed further light on this issue.

Although there is a large number of learning style taxonomies, the visual/auditory taxonomy has received much attention in different areas of education (for reviews of learning styles, see Coffield, Moseley, Hall, & Ecclestone, 2004; Kozhevnikov, 2007). The visual/auditory taxonomy is also referred to as perceptual learning styles because they are associated with how individuals perceive and consequently process incoming environmental information. The visual/auditory styles taxonomy is in particular important for L2 learning research because L2 learners receive requisite data for L2 acquisition in visual or auditory modes or their combination. Therefore, it can be hypothesized that visual learners benefit more from visual data (i.e. written language, pictures and images) while auditory learners would benefit more from aural data (i.e. spoken language).

One way to test the hypothesis that matching learning style with instruction mode would further enhance L2 development is to investigate the effects of different modes of glossing on L2 vocabulary learning when glossing modes match or mismatch individuals’ perceptual style. More specifically, it can be hypothesized that learners with visual learning style would benefit more from visual or textual glosses while learners with auditory style would benefit more from audio glosses. In fact, little is known about the benefits of matching instructional technique and learning style on L2 vocabulary learning. Few L2 vocabulary learning studies that investigated if matching learners’ perceptual learning styles would enhance vocabulary learning provided mixed and conflicting results. Some studies reported that certain instructional techniques or a combination of several techniques are more effective for vocabulary learning irrespective of learning style (e.g. Tight, 2005, 2010; Yeh & Wang, 2003). Other studies, on the other hand, indicated the beneficial effects of matching perceptual learning styles and vocabulary instruction (e.g. Bascome, 2004; Koppelman, 1999). As noted by Tight (2010), although previous literature provides evidence for the role of learning style preferences in L2 instruction, the results of previous studies are far from conclusive. Due to its importance, the investigation of visual and auditory learning styles in L2 vocabulary learning literature deserves more attention. In fact, little research has been done in L2 acquisition literature to examine if tailoring instruction to learning styles, and in particular visual and auditory styles, is beneficial for L2 vocabulary learning. To shed further light on the effects of matching instruction and individuals’ perceptual style on L2 learning and in particular vocabulary learning, the present study investigates the effects of textual and audio glosses on L2 vocabulary learning by EFL learners with visual and auditory perceptual style. To this end, the following research questions were formulated for the current study:

Do EFL learners benefit from L1 textual and audio glosses for vocabulary learning?

Which glossing condition is more effective for L2 vocabulary learning?

Does learning style moderate the effects of glossing on L2 vocabulary learning?

II Method

1 Design

The present study employs an experimental design with participants being randomly assigned into different treatment conditions. At first, the participants were assigned into two groups of visual and auditory learners based on the scores they obtained in a learning style questionnaire (Slack & Norwich, 2007). Then, each learning style group was randomly subdivided into three gloss conditions (textual, audio, & no glosses) resulting in the following six treatment conditions:

visual learners / textual glosses (n = 15)

visual learners / audio glosses (n = 15)

visual learners / no glosses (n = 15)

auditory learners / textual glosses (n = 16)

auditory learners / audio glosses (n = 16)

auditory learners / no glosses (n = 16)

A vocabulary knowledge test was assigned at the beginning of the study before the treatment session to identify a number of target words that were unfamiliar to learners. One week later the learners participated in one treatment session. Immediately after the treatment, the participants took a post-test. Two weeks later, a delayed post-test was administered.

2 Participants

Six EFL upper intermediate level classrooms in Iran provided the participants for the current study. The total number of participants enrolled in these classrooms was 161. After taking the learning style questionnaire, learners who did not meet the criterion for being assigned into visual and auditory groups were excluded from the study. As a result, 93 participants were selected as the final sample and were assigned into visual and auditory conditions. There were 52 females and 41 males ranging in age from 21 to 37. All participants had experienced 5 to 6 years of learning English as a foreign language and were studying English at an upper intermediate level in the language teaching institute from which they were recruited to participate in the study.

3 Learning style questionnaire

In order to assign learners into visual and auditory learners, a learning style questionnaire developed by Slacks and Norwich (2007) was used (Appendix 1). The original questionnaire includes 18 items measuring three style preferences, namely visual, auditory, and kinesthetic style. For the purpose of the current study, only 12 items which are related to measuring auditory and visual styles were selected and administered to the participants. The items were scored based on a 5-point Likert scale ranging from never to almost always (never; occasionally; sometimes; usually; almost always). The Cronbach alpha reliability of the questionnaire as reported by Slacks and Norwich was .63 and .75 for the visual and auditory subscales. They also reported .90 and .95 test–retest reliability for the two subscales. For the purpose of the current study, learners’ visual and auditory scores based on the two subscales were separately estimated for each learner by adding up the scores associated with the items comprising each subscale and then dividing the aggregate score by the number of items constituting the subscale. In this way, for each learner two scores ranging from 1 to 5 were obtained, representing the learners’ visual and auditory scores. In order to assign learners into either visual or auditory styles, based on the 1–5 scores range for each subscale, learners’ whose mean scores were at or above 3.6 in either visual or auditory subscales were considered as visual or auditory learners respectively. In other words, learners whose mean scores in the visual subscale were at or above 3.6 were considered as visual learners and those whose mean scores in the auditory subscale were above 3.6 were considered as auditory learners. This cutoff point was selected because it approximately represents ‘often’ (point 4 on the Likert scale) and ‘almost always’ (point 5 on the Likert scale) answers to the questionnaire’s items. Several learners who scored above the cut-off point in both subscales were excused from the study. As a result, 45 learners were assigned into visual group and 48 learners assigned into auditory group. As mentioned above, the auditory and visual learners were each subdivided into three glossing conditions. Table 1 indicates descriptive statistics regarding the participants’ visual or auditory scores for each of the three glossing conditions.

Table 1.

Descriptive statistics for learners’ perceptual style scores.

Condition	n	Mean	SD
Visual/audio	15	4.0	.26
Visual/textual	15	3.9	.16
Visual/no gloss	15	3.9	.21
Auditory/audio	16	4.0	.30
Auditory/textual	16	3.9	.25
Auditory/no gloss	16	3.8	.17

4 Target vocabulary

In order to examine the effects of different treatment conditions on L2 vocabulary learning, it was necessary to identify a number of words which were unfamiliar to learners. As a result, the learners of the current study were provided by a list 30 words which were considered by the researcher to be unfamiliar to learners. The learners were asked to write a Persian or English definition or synonym of each target word in a space provided along with each word. The words were selected from a text that was also used as the treatment material for the study. Based on learners’ answers to this test, those words for which learners provided no correct definition or synonym were selected as the target words for the current study. The target words were: hassle; nondescript; porch; beckon; implore; gaze; wrinkle; senile; swallow; persuasion; worthwhile; secluded; shady; dreary.

III Materials

For the purpose of this study, two reading passages were employed as reading materials to provide learners with target words. One of the texts was selected and modified by the researcher from ‘Chicken soup for the soul’ (Canfield & Hansen, 1993). The second text was developed by the researcher based on the content of the first text and incorporated the same target words. In other words, the second text was a paraphrase of the first text. The length of each text was around 600 words. Each text contained all target words. Therefore, all learners were exposed to each target word twice. The computerized versions of the selected reading texts, with hypermedia links to separate audio and textual files for target words definitions, were prepared by the researcher using Microsoft PowerPoint.^® The hyperlinks provided learners with annotations in ether textual (written) or audio definitions for the target vocabulary while reading each text. The same texts without hyperlinks to audio or textual definitions were developed for the control conditions.

1 Treatment procedure

After being assigned to a specific treatment conditions, the learners of each condition were placed in a laboratory with each learner having access to one computer. The learners had access to the two treatment texts with the target words marked with hyperlinks in each text. Before providing the learners with the texts, in order to facilitate learners’ comprehension of the texts, a research assistant narrated the story that the texts represented for learners in their L1. Then, the learners of the experimental groups were asked to read the two texts via their computer screens and answer 6 L2 true/false comprehension questions that followed each text. The comprehension questions were different for each text and did not include any target word. The learners were also asked to click on the hyperlink provided for each target word only once to receive the Persian definitions of the target words either in textual or auditory mode. The learners of the textual glosses conditions (textual glossing/visual style; textual glossing/auditory style) could view the Persian textual definition of the target words in separate slides whenever they clicked on the hyperlinks provided below each target word. They were instructed to get back to the reading text as soon as they saw the definition by once clicking the backspace key on their keyboard. The participants of the audio glossing (audio glossing/auditory style; audio glossing/visual style) conditions could hear the L1 definitions of the target words via headphones by clicking on the hyperlinks associated with each target word. In order to keep the frequency of the exposure to the target words definitions equal for all treatment conditions, the learners were instructed to check the definition of each target word by clicking on its hyperlink only once. The audio definitions of the target words were recorded by the researcher. The same Persian definitions were provided in textual or auditory modes for learners of the textual and audio gloss conditions. The learners were allowed to ask the definition of any non-target word in the text that they did not know from the research assistant who was monitoring the procedure. Regarding the control conditions, the learners were just asked to read the texts via their computer screens without having access to the target word definitions, and answer the follow-up comprehension questions. The average time taken for reading the two texts and completing the task was 43 minutes for the textual condition and 40 minutes for the learners of the auditory condition. Table 2 indicates descriptive statistics for the participants’ comprehension scores for the two gloss conditions and the control group. Because there were 6 comprehension questions following each text, the maximum possible score for an individual was 12. One-way ANOVA indicated no significant difference among the participants of the three glossing conditions in terms of comprehension of the texts, F (2, 90) = .95, p = .39.

Table 2.

Learners’ reading comprehension scores.

	n	Mean	SD
Textual gloss	31	10.9	0.79
Audio gloss	31	11.0	0.83
No gloss	31	10.7	0.98

2 Testing instruments

In order to examine the effects of the treatment conditions on vocabulary learning, two vocabulary tests that measured two aspects of vocabulary knowledge were prepared and assigned to learners as post-test and delayed post-test. They were (1) a cued production test to measure learners’ productive knowledge of target words, and (2) a multiple choice test that depended more on learners’ recognition of word forms. The cued production test was administered before the multiple choice test.

3 Cued production test

For this test, which aimed at measuring the learners’ productive knowledge of the target words as a result of treatments, the learners were provided with an English text developed by the researcher that included the target vocabulary. The length of the text was about 600 words. The target words along with several non-target words functioning as distractors were removed from the text. The learners were provided instead with Persian equivalent of the removed words in parentheses after each slot. L1 equivalents were used instead of L2 synonyms in order to avoid any confusion due to learners’ possible unfamiliarity with the L2 synonyms. The L1 equivalents were different from those used as textual glosses during the treatment session. The number of letters of each word was also specified. The learners were asked to read the text and fill in the slots using the cues provided. Two different texts of equal difficulty were used for post-test and delayed post-test. The two parallel forms of the texts were constructed by carefully changing some non-target words and also some sentences in the texts. Estimating the test–retest reliability of this test indicated .76 correlation coefficient between the control group’s scores in the post-test and delayed post-test. Learners received one point for correctly incorporating each target word in the text. Spelling errors were not considered as far as the word was intelligible. Learners’ scores were obtained and reported in percent by estimating the number of target words correctly incorporated by the learners. A second rater was also asked to randomly rate 50 percent of the texts in the post- and delayed post-tests which indicated .96 inter-rater reliability.

4 The multiple choice recognition test

This test included 22 items. Each item consisted of a Persian word followed by four English synonyms. Among the four English options one was a target word. Therefore, learners needed to identify the correct English word equivalent to the Persian definition provided. L1 to L2 recognition test was used instead of L2 to L1 recognition items because it can be argued that the former which requires learners to identify the L2 word form among four L2 words is more demanding than the latter that only requires recognizing the L1 meaning of an L2 word. Among the 22 items, 14 were associated with target words and the rest were distractors. As a result, the highest possible score for each learner was 14. No strict time limit was set for learners to complete the test. Two parallel forms of this test were developed for the post-test, and delayed post-test by changing the distractor options and also the order of items. The Persian equivalents used were not those used as textual glosses during the treatment session. Furthermore, most of the L1 definitions were different from those used in the cued production test. The reliability of this test was measured by correlating the control’s group scores in the post- and delayed post-tests which indicated .77 test–retest reliability. An example of the test items is provided below.

1. درخواست کردن

a. frustrate b. implore c. condole d. perpetuate

IV Results

Tables 3 and 4 present descriptive statistics for the three glossing conditions in the production and recognition tests. As these tables indicate, participants in all glossing conditions significantly outperformed no glossing conditions in both production and recognition tests. Among the glossing conditions, the mean scores for the audio glossing conditions are higher than other conditions.

Table 3.

Descriptive statistics for the three glossing conditions in the cued production test.

	n	Post-test		Delayed post-test
	n	Mean	SD	Mean	SD
Textual glosses	31	6.1	0.5	6.0	0.5
Audio glosses	31	7.4	0.9	7.1	0.9
No glosses	31	2.5	0.8	2.3	0.4

Table 4.

Descriptive statistics for the three glossing conditions in the multiple choice recognition test.

	n	Post-test		Delayed post-test
	n	Mean	SD	Mean	SD
Textual gloss	31	7.6	0.8	7.0	0.8
Audio gloss	31	9.2	1.2	8.7	1.6
No gloss	31	3.1	0.7	2.3	0.8

In order to see if the differences among the three glossing conditions is statistically significant, one-way ANOVA was performed on learners’ post- and delayed post-test scores with glossing condition (visual, audio, no glosses) as an independent variable and learners’ scores as the dependent variable. It should be noted that a preliminary examination of the dependent variables (i.e. scores) before running one-way ANOVA indicated that the distribution of the scores is normal.

With respect to the production post-test, one-way ANOVA indicated a statistically significant difference among the groups, F (2, 90) = 321, p < .001. Tukey’s post hoc comparisons revealed that both treatment conditions significantly outperformed the control condition (p < .001) while the audio glossing condition significantly outperformed the textual glossing condition (p < .001). Likewise, the analysis of the production delayed post-test indicated statistically significant differences among the groups, F (2, 90) = 321, p < .001. The results also revealed the significant advantage of both glossing conditions over the control condition (p < .001) with the audio glosses condition outperforming the textual glosses condition (p < .001).

With regard to the recognition post-test, the analysis indicated a statistically significant difference among the three treatment conditions F (2, 90) = 268, p < .001. Tukey’s post hoc analysis revealed that the participants of both glossing conditions outperformed the control condition (p < .001). Furthermore, the audio glossing group also significantly outperformed the textual glossing group (p < .001). The same results were also obtained regarding the delayed recognition test. That is, a statistically significant difference was found among the three glossing conditions, F (2, 90) = 211, p < .001, and the analysis indicated that both glossing conditions outscored the control condition significantly (p < .001). Furthermore, the difference between the two glossing conditions was also statistically significant in favor of the audio gloss condition (p < .001).

Finally, in order to examine the moderating effects of the participants’ preferred perceptual styles on learning, two-way ANOVAs were performed with glossing conditions (textual vs. audio) and learning style (visual vs. auditory) as two independent variables and learning scores as a dependent variable.

With respect to the production post-test, two-way ANOVAs indicated significant effects for glossing, F (1, 58) = 50, p < .001, and learning style F (1, 58) = 4.5, p < .05. Furthermore, the interaction between glossing and learning style was found statistically significant F (1, 58) = 4.7, p < .05. These findings suggest that besides textual and auditory glosses, individuals’ preferred learning styles influenced learners’ scores and that learning style had differential effects on the efficacy of glossing conditions. Regarding the production delayed post-test, two-way ANOVA revealed statistical effect for glossing F (1, 58) = 63, p < .001. However, no significant effect was found for learning style F (1, 58) = 1.2, p = .26. Meanwhile, the interaction between learning style and glossing type turned out to be statistically significant, F (1, 58) = 7.9, p < .01.

Table 5 displays the mean scores for the four treatment conditions in the production post- and delayed post-test. As the table indicates, both audio gloss conditions have been superior over the textual conditions in post-test and delayed post-test. Moreover, the audio gloss/auditory condition proved to be the most effective condition for vocabulary learning. One-way ANOVA on learners’ scores obtained from these four conditions also revealed statistically significant differences among the groups in the post-test, F (3, 58) = 20.3, p < .001, and also delayed post-test, F (3, 58) = 24.7, p < .001. With respect to the post-test, Tukey’s pairwise comparisons revealed that the audio gloss/auditory style group significantly outperformed both textual glossing conditions (p < .001) and also the audio gloss/visual style group (p < .05). Furthermore, the audio gloss/visual style group also significantly outperformed both textual glossing conditions (p < .01) while no difference was found between the two textual glossing conditions. Similar results were obtained regarding the production delayed post-test. More specifically, the audio gloss/auditory style group significantly outperformed the two textual glossing conditions (p < .001) and also the audio gloss/visual style group (p < .05). Moreover, the audio gloss/visual style condition outscored the textual visual (p < .005) and textual auditory condition (p < .001).

Table 5.

Group comparisons for the production post- and delayed post-tests.

	Post-test		Delayed post-test
	Mean	SD	Mean	SD
Textual gloss/visual style	6.1	0.5	6.0	0.5
Textual gloss/auditory style	6.1	0.9	5.6	0.8
Audio gloss/visual style	7.3	1.0	7.1	0.9
Audio gloss/auditory style	8.3	1.2	8.0	1.0

Regarding the recognition post-test, two-way ANOVA indicated statistical effects for glossing, F (1, 58) = 41.8, p < .001, and also learners’ perceptual style, F (1, 58) = 7, p < .05. Furthermore, the interaction between glossing condition and style preferences was also found statistically significant, F (1, 58) = 4.1, p < .05 suggesting that learners’ perceptual style differentially affected learners’ scores in the two glossing conditions. With regard to the recognition delayed post-test, two-way ANOVA results also indicated the significant effects of glossing conditions, F (1, 58) = 35.6, p < .001, and also learners’ perceptual styles, F (1, 58) = 6.1, p < .05 on learners’ scores. However, the interaction between glossing conditions and learning styles was non-significant, F (1, 58) = 2.6, p = .10.

Table 6 displays group means in four different treatment conditions as a result of glossing and perceptual style levels for the recognition post- and delayed post-tests. As this table indicates, the audio gloss/auditory style group obtained the highest mean score in both post- and delayed post-tests. Moreover, as Table 6 shows, the audio gloss/visual style condition obtained the second highest mean score among the groups in both post-test and delayed post-test. One-way ANOVA also confirmed statistically significant differences among the four conditions in the post-test, F (3, 58) = 18, p < .001. Pairwise comparisons indicated that the audio gloss/auditory style group significantly outperformed the other groups including the audio gloss/visual style condition (p < .001). This finding corroborates the significant interaction found between glossing and learning style as provided by the two-way ANOVA analysis. Moreover, the audio gloss/ visual condition significantly outperformed the two textual gloss conditions (p < .001). Similarly, one-way ANOVA performed on learners’ delayed post-test scores indicated significant differences among the groups, F (3, 58) = 15, p < .001. Tukey’s pairwise comparisons indicated that the audio gloss/auditory condition significantly outscores other conditions (p < .001). The only non-significant difference found was the one between the two textual glossing conditions (p = .93).

Table 6.

Group comparisons for recognition post- and delayed post-tests.

	Post-test		Delayed post-test
	Mean	SD	Mean	SD
Textual gloss/visual style	7.6	0.8	7.0	0.8
Textual gloss/auditory style	7.8	1.4	7.3	1.0
Audio gloss/visual style	9.1	1.3	8.6	1.6
Audio gloss/auditory style	10.6	1.5	10.1	1.5

V Discussion

The current study examined the effects of glossing mode and learning style on L2 vocabulary knowledge. Two glossing modes investigated in the current study were written (textual) and oral (audio) while the learning styles examined were visual and auditory perceptual styles. The results indicated that both types of glosses are meritorious for vocabulary acquisition. Furthermore, the results also suggest that learning style preferences also affect vocabulary leaning. The first important finding of the current study was that computer-mediated glosses both in textual and audio modes promote vocabulary learning as participants who were exposed to either textual or audio glosses irrespective of their learning style outperformed the participants of no gloss conditions. Therefore, with regard to the first research question, we can state that both glossing conditions were effective for vocabulary learning. Therefore, the results of the current study in line with previous research (e.g. Hulstijn, Hollander, & Greidanus, 1996; Yanguas, 2009; Yoshii, 2006) provided further evidence for the efficacy of glosses in vocabulary learning. However, it should be remembered that the current study examined the effects of L1 textual and audio glosses on L2 vocabulary learning and thus it may not be appropriate to generalize the results of this study to L2 gloss conditions.

Regarding the second research question, the findings provided evidence that audio glosses are more effective than textual glosses for vocabulary learning. This finding is an important issue because previous studies largely focused on the effects of textual glosses on vocabulary learning. In fact, the possible effects of audio glosses that turned out to be more effective than textual glosses in the current study were often ignored by previous research. To the best of the researcher’s knowledge, no previous study has directly compared the effects of audio and textual glosses on L2 vocabulary learning. It should be noted that in the current study both textual and audio glosses were presented to learners in their L1. So, why were L1 audio glosses more effective than L1 textual glosses in the current study? One speculation is that learners paid more attention to the audio glosses (than the textual glosses) because they were presented in a different modality. In other words, for the audio glosses conditions, the text and the glosses modalities were different and this could trigger more attention to the glosses. Another possible explanation as noted by an anonymous reviewer of Language Teaching Research is related to the fact that the participants who received textual glosses were directed to new slides to see the definitions of the target words while in the audio glossing condition the reading text remained visible while the learners received the audio glosses. Therefore, it is possible that learners’ exposure to word definitions without disruption of the flow of reading comprehension for the audio glossing condition resulted in more efficient processing of the target words while reading the text.

Regarding the third research question, which is the focus of the current study, two-way ANOVA indicated the significant moderating effect of learning style on vocabulary learning. A closer examination of the results based on one-way ANOVAs on learners’ post- and delayed post-tests scores indicated that the combination of auditory learning style and audio glosses resulted in the most effective condition for vocabulary learning. This issue further corroborates the results of the two-way ANOVAs that indicated significant interaction effects between gloss type and learning style for post- and delayed post-tests. Therefore, the answer to the third research question is also affirmative. Previous studies provided mixed results regarding the role of learning style in different aspects of L2 learning. In the context of child L1 learning, Slack and Norwich (2007) found that children were better able to retain words spelling when the instructional method matched their perceptual style. With regard to L2 vocabulary learning, Tight (2010) also found that matching vocabulary instruction mode with learners’ perceptual style is advantageous over mismatching instruction and style. The results of the current study appear to support the general view in L2 acquisition literature that learning style is a factor that affects L2 learning. Other taxonomies of learning style such as field dependence/independence have also been shown to affect L2 acquisition (e.g. Johnson, Prior, & Artuso, 2000; Rassaei, 2015).

Another issue that also merits further attention is the superiority of auditory/audio condition over visual/audio condition. In other words, while the results indicated that audio glosses were more effective than textual glosses, the findings also revealed that matching auditory style with audio glosses is more effective than the visual/audio condition for all testing occasions. This finding corroborates the hypothesis that matching instructional style to learning style enhances learning. Therefore, the results of the current study supported the results of several studies that indicated the benefits of matching instruction with learning style for vocabulary learning (e.g. Slack & Norwich, 2007; Tight, 2010).

Another finding of the current study is that while the auditory/audio group outscored other treatment conditions, the visual/textual condition was not as effective as other treatment conditions that involved audio glosses. The point here is that matching instruction with perceptual style did not work for the visual learners as effectively as it did for auditory learners. One possible explanation for this finding is that the effect of gloss type was more significant than the effect of learning style on vocabulary learning. Another possibility is that the combination of certain perceptual styles and instruction modes enhances learning while other adjustments provide no further advantages for instruction.

VI Conclusions

1 Conclusion and implications

Several major conclusions can be drawn from the results of the current study. First, the results suggested that both textual and audio glosses are effective for promoting vocabulary learning. Second, it was found that audio glosses are more effective than textual glosses for vocabulary learning. Moreover, the findings of the study provided evidence for the moderating role of learning style in vocabulary learning. In particular, the results suggested that matching learning style and instructional mode can benefit vocabulary learning. Meanwhile, the findings also suggest that the effect of matching learning style with instruction is not always straightforward and this issue can be undermined by other factors that can exert stronger influence on learners’ scores. Based on these conclusions, several pedagogical implications can be inferred. One important implication is that audio glosses can be incorporated into language teaching materials for enhancing vocabulary learning given the increase in the use of technology in language classrooms. This study provided further proof for the important role of different forms of technology in language classrooms and in particular vocabulary teaching.

Another implication of the results of the current study is that teachers should consider individual differences as a factor that might influence the outcome of instruction. In particular, the current study provided evidence that the combination of auditory style and audio glosses would create a significantly effective condition for vocabulary learning.

2 Limitations and suggestions for further research

There are a number of issues that need to be considered when interpreting the results of the present study. The first limitation is related to the control group. While the participants of the two experimental groups were allowed to ask the meaning of the unfamiliar non-target words in the reading texts, the participants of the control group were not allowed to do so because the definitions of the target words were not glossed for them, and this could prompt the learners of the control condition to ask the meaning of target words. However, this was not a big problem since very few non-target words turned out to be unfamiliar to several participants of the experimental groups. Moreover, analysing learners’ reading comprehension scores suggested no difference in terms of the comprehension of the texts between the control group and the two experimental conditions.

Moreover, as noted by an anonymous reviewer of Language Teaching Research, the participants of the control group did not have any support to notice the target words while they were reading the texts. In other words, not only the control group had no access to the glosses but also the target words were not underlined in the texts presented to the participants of the control condition, resulting in little chance for them to compete with the participants of the experimental conditions. Furthermore, as noted by the same reviewer, the fact that the participants were provided with an L1 description of the texts before reading the L2 texts could have affected the strategies the learners employed in reading the texts because they were no longer reading for meaning. However, it should be noted that the L1 description of the texts was very general and the learners needed to read the texts to answer the follow-up true/false questions. Finally, as argued above, due to the program used for presenting the audio and textual glosses, the participants of the textual glossing conditions could view the L1 glosses for the target words on a separate page rather than on the same page after clicking on each word, and this could negatively affect the benefits of textual glosses in the current study.

Therefore, one suggestion for future research is to conduct a similar study and address the above limitations. Furthermore, as mentioned before, the perceptual learning style for this study was considered as a categorical variable, and learners whose mean scores on the two subscales of the learning style questionnaire were at or above 3.6 based on the Likert scale were assigned into visual and auditory groups. In this way, a number of learners whose mean scores fell below the cutoff point were discarded from the study. Another suggestion for further research is therefore to replicate the current study and incorporate the leaning style as a continuous variable. Moreover, as mentioned above, since L1 glosses were used in the current study, the results cannot be generalized to L2 gloss conditions. Thus, another suggestion for further research is to investigate the effects of learning style and L2 (audio and textual) glosses on vocabulary learning. Moreover, a similar study can be conducted using other learning style inventories. As mentioned before, since the evidence for the benefits of matching learners’ style and instruction is far from conclusive, further studies that incorporate other perceptual style measures will shed further light on the possible benefits of style matching on vocabulary learning and other aspects of L2 knowledge.

Footnotes

Appendix 1 Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Abraham

L.B.

(2008). Computer-mediated glosses in second language reading comprehension and vocabulary learning: A meta-analysis. Computer Assisted Language Learning, 21, 199–226.

Al-Seghayer

(2001). The effect of multimedia annotation modes on L2 vocabulary acquisition: A comparative study. Language Learning & Technology, 5, 202–232.

Bascome

E.G.

(2004). Effects of learning-styles instructional resources on short- and long-term vocabulary retention and attitudes of seventh-grade language arts students. Unpublished doctoral dissertation, St. John’s University, Queens, NY.

Brinton

(2001). The use of media in language teaching. In: Celce-Murcia

(Ed.), Teaching English as a second or foreign language (pp. 459–476). Boston, MA: Heinle & Heinle.

Canfield

Hansen

M.V.

(1993). A 2nd helping of chicken soup for the soul. London: Vermilton.

Chen

(2002). Investigating the effects of L1 and L2 glosses on foreign language reading comprehension and vocabulary retention. Unpublished paper presented at the annual meeting of the Computer-Assisted Language Instruction Consortium, Davis, CA, USA.

Chun

Plass

(1996). Effects of multimedia annotations on vocabulary acquisition. The Modern Language Journal, 80, 183–198.

Coffield

Moseley

Hall

Ecclestone

(2004). Learning styles and pedagogy in post-16 learning. A systematic and critical review. London: Learning and Skills Research Centre.

Duan

Yan

(2004). Impact of multiple choice note on English incidental vocabulary acquisition. Foreign Language Teaching and Research, 3, 213–218.

10.

Gettys

Imhof

L.A.

Kautz

J.O.

(2001). Computer-assisted reading: The effect of glossing format on comprehension and vocabulary retention. Foreign Language Annals, 34, 91–106.

11.

Grace

(1998). Personality type, tolerance of ambiguity, and vocabulary retention in CALL. CALICO Journal, 15, 19–45.

12.

Grace

C.A.

(2000). Gender differences: Vocabulary retention and access to translations for beginning language learners in CALL. The Modern Language Journal, 84, 214–224.

13.

Hulstijn

J.H.

(2001). Intentional and incidental second-language vocabulary learning: A reappraisal of elaboration, rehearsal and automaticity. In: Robinson

(Ed.), Cognition and second language instruction (pp. 258–286). Cambridge: Cambridge University Press.

14.

Hulstijn

J.H.

Hollander

Greidanus

(1996). Incidental vocabulary learning by advanced foreign language students: The influence of marginal glosses, dictionary use, and reoccurrence of unknown words. The Modern Language Journal, 80, 327–339.

15.

Jacobs

G.M.

Dufon

Fong

C.H.

(1994). L1 and L2 vocabulary glosses in L2 reading passages: Their effectiveness for increasing comprehension and vocabulary knowledge. Journal of Research in Reading, 17, 19–28.

16.

Johnson

Prior

Artuso

(2000). Field dependence as a factor in second language communicative production. Language Learning, 50, 529–567.

17.

Kim

Gilman

D.A.

(2008). Effects of text, audio, and graphic aids in multimedia instruction for vocabulary learning. Educational Technology & Society, 11, 114–126.

18.

Kondo-Brown

(2006). How do English L1 learners of advanced Japanese infer unknown kanji words in authentic texts? Language Learning, 56, 109–153.

19.

Koppelman

V.N.

(1999). A comparison of vocabulary instruction outcomes of urban seventh grade students who have been taught through their learning style preference and those who have been taught in a traditional fashion. Unpublished doctoral dissertation, St. John’s University, Queens, NY, USA.

20.

Kost

C.R.

Foss

Lenzini

J.J.

(1999). Textual and pictorial glosses: Effectiveness on incidental vocabulary growth when reading in a foreign language. Foreign Language Annals, 32, 89–113.

21.

Kozhevnikov

(2007). Cognitive styles in the context of modern psychology: Toward an integrated framework of cognitive style. Psychological Bulletin, 133, 464–481.

22.

Krätzig

Arbuthnott

(2006). Perceptual learning style and learning proficiency: A test of the hypothesis. Journal of Educational Psychology, 98, 238–246.

23.

Laufer

Shmueli

(1997). Memorizing new words: Does teaching have anything to do with it? RELC Journal, 28, 89–108.

24.

Liu

Y.T.

Todd

A.G.

(2014). Dual-Modality input in repeated reading for foreign language learners with different learning styles. Foreign Language Annals, 47, 684–706.

25.

Lovelace

M.K.

(2005). Meta-analysis of experimental research based on the Dunn and Dunn model. The Journal of Educational Research, 98, 176–183.

26.

Miyasako

(2002). Does text-glossing have any effects on incidental vocabulary learning through reading for Japanese senior high school students? Language Education & Technology, 39, 1–20.

27.

Naimie

Siraj

Piaw

C.Y.

Shagholi

Abuzaid

R.A.

(2010). Do you think your match is made in heaven? Teaching styles/learning styles match and mismatch revisited. Procedia-Social and Behavioral Sciences, 2, 349–353.

28.

Othman

Amiruddin

M.H.

(2010). Different perspectives of learning styles from VARK model. Procedia-Social and Behavioral Sciences, 7, 652–660.

29.

Oxford

R.L.

(2003). Language learning styles and strategies: Concepts and relationships. International Review of Applied Linguistics in Language Teaching, 41, 271–278.

30.

Peacock

(2001). Match or mismatch? Learning styles and teaching styles in EFL. International Journal of Applied Linguistics, 11, 1–20.

31.

Plass

J.L.

Chun

D.M.

Mayer

R.E.

Leutner

(2003). Cognitive load in reading a foreign language text with multimedia aids and the influence of verbal and spatial abilities. Computers in Human Behavior, 19, 221–243.

32.

Plonsky

Ziegler

(2016). The CALL–SLA interface: Insights from a second-order synthesis. Language Learning & Technology, 20, 17–37.

33.

Qian

D.D.

(2004). Second language lexical inferencing: Preferences, perceptions and practices. In: Laufer

Bogaards

(Eds.), Vocabulary in a second language: Selection, acquisition and testing (pp. 155–169). Amsterdam: John Benjamins.

34.

Rassaei

(2015). Recasts, field dependence/independence cognitive style, and L2 development. Language Teaching Research, 19, 499–518.

35.

Rohrer

Pashler

(2012). Learning styles: Where’s the evidence? Online Submission, 46, 634–635.

36.

Rott

Williams

Cameron

(2002). The effect of multiple-choice L1 glosses and input–output cycles on lexical acquisition and retention. Language Teaching Research, 6, 183–222.

37.

Slack

Norwich

(2007). Evaluating the reliability and validity of a learning styles inventory: A classroom-based study. Educational Research, 49, 51–63.

38.

Stitt-Gohdes

W.L.

(2001). Business education students’ preferred learning styles and their teachers’ preferred instructional styles: Do they match? Delta Pi Epsilon Journal, 43, 137–151.

39.

Sydorenko

(2010). Modality of input and vocabulary acquisition. Language Learning & Technology, 14, 50–73.

40.

Thomson

Watt

Liukkonen

(2015). Cognitive style and teaching style influences on the motor skill performance of 11 and 12 year old physical education students. Journal of Physical Education and Sport, 15, 509–517.

41.

Tight

D.G.

(2005). Picture drawing, learning style preferences, and retention of L2 vocabulary. Unpublished paper presented at the Second Language Research Forum, New York, USA.

42.

Tight

D.G.

(2010). Perceptual learning style matching and L2 vocabulary acquisition. Language Learning, 60, 792–833.

43.

Watanabe

(1997). Input, intake, and retention: Effects of increased processing on incidental learning of foreign language vocabulary. Studies in Second Language Acquisition, 19, 287–307.

44.

Yanguas

(2009). Multimedia glosses and their effect on L2 text comprehension and vocabulary learning. Language Learning & Technology, 13, 48–67.

45.

Yee

S.W.

(2010). Short-term and long-term retention of new words: Investigating the role of L1 glossing in vocabulary learning among Hong Kong ESL learners. Unpublished masters dissertation, The University of Hong Kong, Hong Kong.

46.

Yeh

Wang

C.W.

(2003). Effects of multimedia vocabulary annotations and learning styles on vocabulary learning. CALICO Journal, 21, 131–144.

47.

Yoshii

(2006). L1 and L2 glosses: Their effects on incidental vocabulary learning. Language Learning & Technology, 10, 85–101.