The impact of topic interest,L2 proficiency,and gender on EFL incidental vocabulary acquisition through reading

Abstract

This study investigated the impact of topic interest, alongside L2 proficiency and gender, on L2 vocabulary acquisition through reading. A repeated-measures design was used with 135 Korean EFL students. Control variables included topic familiarity, prior target-word knowledge, and target-word difficulty (word length, class, and concreteness). Participants read both high- and low-interest topic passages and took vocabulary posttests (word-form recognition, translation recognition, and translation production) immediately and four weeks after reading. Analyses revealed significant effects of topic interest and L2 proficiency, and a significant interaction between topic interest and gender. These results were maintained over time. The article concludes by discussing the facilitative role of topic interest, expanding on the motivational factor considered in the involvement load hypothesis.

Keywords

Gender involvement load hypothesis L2 incidental vocabulary acquisition L2 proficiency topic interest

I Introduction

Vocabulary knowledge and its acquisition play a critical role in second language (L2) literacy development (Nation, 2001; Schmitt, 2008). As a result, L2 researchers have sought to identify how various factors affect new word gains through reading such as frequency (Schmitt, 2010), background knowledge (Pulido, 2003), and L2 proficiency (Pulido, 2004). One neglected factor, however, is topic interest: interest that learners have in the texts they read. This omission is relevant given that affective and cognitive variables are intertwined in second language learning (Ortega, 2014). Interest-based motivation can significantly impact learning because ‘interest impacts students’ attention and memory for tasks and their depth of processing’ (Hidi, Renninger, & Krapp, 2004, p. 106). For literacy activities, learner interest may have a direct bearing on how readers allocate attention to information encountered within texts, such as new vocabulary. Given that first language (L1) and L2 studies have found robust relationships between reading comprehension and vocabulary acquisition (Pulido, 2004), topic interest should influence engagement in reading tasks, and such engagement can, in turn, impact vocabulary acquisition.

With regard to the mechanism underlying L2 incidental vocabulary acquisition, Laufer and Hulstijn’s (2001) involvement load hypothesis holds that motivational factors affect the learning process. Expanding on the depth-of-processing theory with its focus on cognitive factors in word retention (Craik & Lockhart, 1972), Laufer and Hulstijn (2001) argued that both cognitive and motivational components play pivotal roles in human learning. They suggested that when induced by tasks, learners display a combination of degrees of involvement both cognitively and motivationally, which can predict the subsequent retention of words encountered while engaging in such tasks. This theory predicts that the higher the involvement load when processing a new word, the better the chances of its retention. For learner interest, when intrinsic interest in a text’s topic is high, this component – reflecting overall motivation – can lead to a higher degree of involvement while processing a text. Thus, combined with cognitive components, learner interest can influence overall involvement in reading tasks, which can affect vocabulary gains.

Although little research exists on the role of topic interest in L2 incidental vocabulary learning, previous L1 empirical studies have reported strong effects of topic interest on reading comprehension by young learners, sometimes with the effects of this factor being mediated by other individual difference factors, such as language proficiency and even gender. Graham, Tisher, Ainley, and Kennedy (2008) explain that boys and girls engage in literacy tasks differently, in that boys and adolescent males are less likely to engage positively and therefore spend less time when approaching reading tasks compared to girls and adolescent females. Thus, it is vital that L2 research concerned with learning outcomes from literacy activities consider the added effects of gender.

To date, there is a lack of empirical L2 research examining the nature of the interaction between the variables of topic interest, L2 proficiency, and gender on vocabulary learning through reading. Furthermore, the extant research on the various factors listed above has largely been conducted with adult reader participants, involving reading comprehension as the outcome variable of interest (e.g. Carrell & Wise, 1998). Yet many L2 learners in educational settings are children and adolescents. The following section highlights the relevant findings in the existing literature on L1 and L2 reading comprehension and vocabulary acquisition and pinpoints the existing gaps and limitations, which justify the present investigation in an attempt to shed new light on learner involvement in L2 incidental vocabulary acquisition through reading.

II Review of research

1 The effects of topic interest, reading proficiency, and gender on L1 reading comprehension

Research on L1 reading has demonstrated that topic interest significantly facilitates various aspects of the reading process. For instance, Ainley, Hidi, and Berndorff (2002) operationalized the psychological processes that mediated interest and learning and, through structural equation modeling, they reported that topic interest positively influenced students’ affect, which increased the degree of students’ persistence, the latter directly influencing reading comprehension.

L1 research has also shown that topic interest interacts with other learner variables, such as reading proficiency, verbal ability, and gender. For example, early study by Walker, Noland, and Greenshields (1979) investigated the combined effects of L1 reading ability and topic interest on reading comprehension. The study reported significant effects of topic interest on reading comprehension in that participants better comprehended passages depicting more interesting topics than those depicting less interesting topics. Additionally, differences in comprehension across the varying topic interest conditions were more prominent in less proficient readers than in more proficient readers. However, a study by Stevens (1980) reported that more proficient readers comprehended more from the more interesting passages than less interesting ones, though no differences in comprehension due to topic interest were obtained for less proficient readers.

More recently, in a study on L1 reading comprehension assessment, Bray and Barron (2004) investigated the relationships between topic interest, verbal ability, and another relevant reader-based factor (i.e. gender). Their results revealed a consistently strong positive relationship between verbal ability and reading comprehension for all participants regardless of their level of interest in the reading materials. In addition, they reported better comprehension by girls compared to boys, and a stronger positive relationship between interest level and comprehension for girls compared to boys. In other words, topic interest affected girls’ performance more than boys’ performance.

Yet, other studies considering the dual effects of gender and topic interest on L1 reading comprehension obtained different results (Ainley, Hillman, & Hidi, 2002; Walker et al., 1979), whereby comprehension was significantly affected by topic interest, but only for boys. In these studies, boys scored significantly higher on high-interest texts compared to low-interest texts, whereas girls’ comprehension performance remained consistent regardless of their level of interest in the topics. Ainley, Hillman, et al. (2002) further illustrated the mechanism of the effects noted above using a computer program to examine the relationships between gender, topic interest, and comprehension. This study obtained a ‘dynamic assessment’ (p. 417) of topic interest by not only recording interest levels via a Likert-type scale, but by also recording students’ online affective responses to passages (i.e. choice of emoticons to represent feelings evoked by a text), persistence in reading (i.e. decisions to continue reading, time spent reading, and number of passage sections read). Comprehension was also assessed online. Through path analyses, they concluded that the degree of topic interest was significantly associated with affective responses to texts, which contributed to the level of persistence in reading, the latter being correlated with comprehension. Partitioning the higher- from the lower-interest texts, the researchers also reported a direct effect of gender on topic interest and persistence, whereby only the girls persisted with less interesting texts. This study helps to highlight the mechanism by which topic interest may differentially affect reading comprehension based on other factors, such as gender.

On the whole, the L1 literature conducted primarily with children and adolescents has reported robust effects of interest on reading comprehension as well as significant interactions with gender and reading or verbal ability. The L1 research findings are significant because this study is also concerned with grade school students’ learning through reading. Yet, the second language acquisition literature has been less concerned with the impact of these factors on young learners’ reading engagement.

2 The effects of topic interest, reading proficiency, and gender on L2 reading comprehension

L2 reading research has largely focused on post-pubescent and adult learners and has been less successful in obtaining significant effects of topic interest and gender on reading comprehension. For instance, Carrell and Wise (1998) and Joh (2006) found no significant effects of topic interest on their measures of L2 reading comprehension. However, there were methodological limitations to these investigations. For example, Carrell and Wise’s measurement of interest had questionable validity in that students rank ordered their interest in 10 encyclopedia-based topics, regardless of their actual interest in each topic. Although this measure might have represented students’ interest in the given topics relative to others, it did not reflect their absolute interest in each topic. Carrell and Wise found no significant main effect of gender on reading comprehension in their study, corroborating reports by Young and Oxford (1997) and Brantmeier (2003). Interestingly, Young and Oxford found that females more often than males used local strategies, such as lexical inferencing to resolve the meaning of unfamiliar words. Thus, it is possible that there are differential effects of topic interest on L2 reading comprehension based on the gender of the readers.

However, in separate investigations, LeLoup (1993) and Erçetin (2010) obtained significant effects of topic interest on L2 reading comprehension. These studies demonstrated that learners recalled significantly more information from high-interest passages compared to low-interest passages. LeLoup (1993) also found significant effects for the variables of L2 ability and gender (females recalled more information than males). Through regression modeling, LeLoup’s study attributed the majority of the observed variance in L2 reading comprehension to the variable of L2 ability (28%–41%), followed by topic interest (9%), and gender (6%). LeLoup concluded that females may be more facilitated by high-interest topics than males, and that males may be more hindered when reading low-interest topics compared to females.

3 The effects of topic interest, reading proficiency, and gender on L1 and L2 vocabulary acquisition through reading

L1 research has reported on the impact of topic interest and reading proficiency on vocabulary learning through reading, however not on the nature of gender effects. In terms of incidental learning, Elley (1989) found that, on both short- and long-term measures, learners remembered twice the number of new words heard in more appealing stories than from less appealing stories. Including the variable of reading proficiency, Endo (2010) also reported positive effects of topic interest on vocabulary learning through reading, however via explicit vocabulary instruction tasks. In this study, new vocabulary was explicitly connected with a high-interest topic and a low-interest topic. From multiple-choice and semantic judgment measures, the results revealed a significant interaction between topic interest and reading proficiency, whereby only the more proficient readers gained significantly more vocabulary from the high-interest passage than the low-interest passage. These studies, however, have methodological limitations. In Elley’s (1989) study, the difficulty of target words across high- and low-interest stories was not controlled. Thus, it remains unclear whether it was target-word difficulty or interest-level that contributed to the outcomes. Though Endo’s (2010) study showed a significant effect of topic interest, it was only tested in connection with the recognition of word meaning rather than other aspects of vocabulary knowledge. Furthermore, this study did not control for topic familiarity as the majority of the students were already familiar with the high-interest topic (i.e. saw or read Harry Potter), whereas none of the students were familiar with the low-interest topic (i.e. Ancient Mesopotamia). Such variations in prior knowledge thus confound the results connected with the variable of topic interest. Finally, the existing studies are limited to the area of L1 vocabulary research and did not consider the additional role of gender in the assigned literacy tasks.

To date, no L2 studies have considered the interactions between the factors of topic interest, L2 proficiency, and gender in L2 vocabulary learning through reading. The results from the L1 literature on reading comprehension with young readers, and the L2 literature on reading comprehension with adult readers suggests that these factors should also be involved, to some extent, independently or shared, in the process of learning new words from reading because motivational and cognitive components are hypothesized to interact in the vocabulary acquisition process (Laufer & Hulstijn, 2001). Whether or not topic interest differentially affects L2 vocabulary acquisition through reading with adolescent males and females of varying levels of reading proficiency is a matter that remains to be explored. An investigation into these topics could advance language educators’ understanding of incidental L2 vocabulary learning process, thereby facilitating individual learners’ successful L2 reading and vocabulary development. Based on the results of the L1 and L2 literature and the limitations outlined above, the following hypotheses were investigated:

Topic interest will significantly impact L2 vocabulary acquisition through reading.

L2 reading proficiency will significantly impact L2 vocabulary acquisition through reading.

There will be a significant interaction between topic interest and reading proficiency, whereby the effects of topic interest would be greater for the less proficient readers.

Gender will have no significant effect on L2 vocabulary acquisition through reading.

There will be a significant interaction between topic interest and gender, whereby males will be more affected by topic interest than females.

Significant effects and interactions will be observed over time.

III Method

1 Participants

Seven intact ninth-grade classes in a Korean middle school participated in the study. Because this school had a three-to-one ratio of females to males, participants included 101 females and 34 males (n = 135) who had studied English as a Foreign Language (EFL) for more than seven years on average. The participants demonstrated a range of general English proficiency (i.e. 24%: high; 46%: intermediate; and 30%: low) based on the results of the National Academic Achievement Test of English (Korea Institute for Curriculum and Instruction, 2013).

2 Materials

a Passages

This study used two expository passages, one portraying a high-interest topic and one portraying a low-interest topic (Appendix S1 in the Supporting Information). The high-interest passage, ‘Psy’s Gangnam Style’ (Psy), was adapted from a published article¹ and focused on the Korean rapper Psy’s global hit, ‘Gangnam Style’. The low-interest passage, ‘Medieval life’ (Medieval), was also an adapted text² and described the feudal system in European societies in the Middle Ages. That the two passages portrayed high- and low-interest topics, respectively, was confirmed by participants’ self-reported ratings of topic interest on a topic interest inventory administered prior to and immediately after reading (Appendix 1). The results of the topic interest inventory revealed that, prior to reading, ‘Psy’s Gangnam Style’ was rated as the passage topic of highest interest (M = 5.26, SD = 1.37) and ‘Medieval life’ as the passage topic of lowest interest (M = 2.10, SD = 1.31). The levels of interest for each passage topic were maintained immediately after reading (Psy: M = 4.78, SD = 1.39; Medieval: M = 2.50, SD = 1.40). Paired-samples t tests revealed that, prior to reading, students demonstrated significantly higher levels of topic interest in the Psy passage than in the Medieval one, t (134) = 20.27, p < .001, d = 1.75, and these differences were maintained after reading the passages, t (134) = 19.01, p < .001, d = 1.64.

Both passages were of comparable length and linguistic difficulty (Table 1). In addition, participants had a generally high level of prior knowledge regarding the topics depicted in the passages. This was determined by their formal instructional experiences – ‘Medieval life’ was covered in their social studies class prior to the experiment, and ‘Psy’s Gangnam Style’ was internationally well known at the time of this study – and then confirmed via an 8-item true–false prior knowledge measure³ on the contents of the passages. The measure, written in the participants’ L1 (Korean), contained sentences describing content that pertained to each of the two texts’ topics. The reliability of the prior knowledge test for ‘Psy’s Gangnam Style’ and ‘Medieval life’ was .67 and .68 (Cronbach’s α), respectively. Results of the true–false test revealed similarly high levels of prior knowledge for each passage (Psy: M = 6.70, SD = 1.01; Medieval: M = 6.61, SD = 1.40), further confirmed by a paired-samples t test, t (134) = 0.90, p = .096. Through classroom observation and the true–false instrument, we were confident that all participants had similarly high prior knowledge of the topics depicted within the passages, regardless of the differences in participants’ self-reported interest level. In other words, this study controlled for the potentially confounding effect of background knowledge on vocabulary acquisition through reading.

Table 1.

Linguistic characteristics of the texts.

Text	Words	Sentences	Paragraphs	Average characters per word	Average words per sentence	Sentences per paragraph
Psy	318	27	5	4.9	11.8	5.4
Medieval	317	25	5	4.8	12.7	5.0

b Target words

Thirty lexical items were chosen from the two passages, with 15 target words (TWs) drawn from each passage (Appendix 2 and Table S1 in the Supporting Information). The TWs were selected based on prior field test responses from 10 advanced non-participant students who were asked to read the two passages and to underline every unfamiliar word. This task sought to ensure that the actual study participants were unlikely to know the TWs. After this process, a rigorous set of criteria was applied to establish similar difficulty of TWs between the high- and low-interest passages. The criteria for TW selection included frequency, word class, word length, and concreteness based on recommendations by Schmitt (2010), and were considered in the following manner. First, words appearing only once in the passage were selected to control for effects of frequency.⁴ Next, the ratio of content words was balanced between the two passages (i.e. five nouns, five verbs, and five adjectives from each passage). Last, word length and concreteness⁵ were matched between the two passages. The TW selection method outlined above guaranteed both their unfamiliarity and equivalence in difficulty between the high- and low-interest passages. For target words, real words were employed instead of pseudowords to give pedagogical value to the study’s young EFL learners. Two weeks prior to reading the passages, participants’ prior knowledge of the 30 TWs was assessed via a measurement combining a yes/no checklist with a translation component (Pulido, 2003). In this pretest, students were instructed to answer a yes/no question to indicate their familiarity with the words. They were also required to translate, define, or explain in their L1 (Korean) the words for which they had indicated ‘yes’. All translations were scored as follows: 0 = incorrect; .5 = partially correct; 1 = correct. Results from the TW pretest revealed virtually no prior knowledge and no significant differences in scores between the TWs depicted in the Psy and Medieval passages (Psy: M = 0.05, SD = 0.25; Medieval: M = 0.02, SD = 0.15; t (134) = 1.58, p = .117). Based on the measurements and comparisons reported above, we were confident that, with the exception of topic interest, the passages were equivalent in terms of linguistic difficulty, TW selection, and portrayal of familiar topics.

3 Measures: Independent variables

a Topic interest

A topic interest inventory consisting of a 5-item questionnaire was used to assess learners’ level of interest in the passage topics and three additional topics (Appendix 1). Students were instructed to rate their level of interest in each topic on a scale of 1 (‘not interesting’) to 7 (‘very interesting’). Immediately after reading each passage during the experimental sessions, participants again completed the topic interest inventory to ascertain whether students’ initial interest level in each of the target passages was maintained throughout the reading tasks.

b L2 reading proficiency

The L2 reading proficiency measure consisted of the reading section of the National Academic Achievement Test of English (Korea Institute for Curriculum and Instruction, 2012), which is a 35-minute timed test containing readings of a functional (e.g. advertisements) and educational (e.g. persuasive essays) nature. The test was composed of 19 reading passages, each approximately 100 words long and followed by one multiple-choice question designed to test students’ ability to comprehend main ideas. This test was administered two weeks prior to the passage reading sessions, and the variable was included as a covariate (KR 21 = .83).

c Gender

The variable of gender was included to examine its impact and interaction with topic interest in L2 vocabulary acquisition through reading. Gender information was gathered from a background questionnaire.

4 Measures: Dependent variables

Three vocabulary measurements were administered to assess vocabulary gain (immediately after reading) and retention (four weeks after reading). The three tests, administered in the following order, included word-form recognition, translation production, and translation recognition. Three measures were administered in order to assess various dimensions of receptive word knowledge from a partial-to-full range. To control for presentation order effects, the 30 TWs were randomized. Two versions of each test were made by switching the order of words from one version to the next.

a Word-form recognition

The word-form recognition test assessed recognition memory, that is, the ability to recognize new word forms, which is an initial step in vocabulary acquisition. This recognition measure of memory serves as a rough estimate of how deeply textual information was processed because more correctly identified words indicates greater active memory for the input (Pulido, 2004). This test required the student to respond ‘yes’ if they remembered having seen the word in the texts or ‘no’ if not. The word recognition test consisted of 30 TWs and 20 distractors, which were included to correct for guessing (Mochida & Harrington, 2006). The test counted each correct answer as one point and each false recognition as one point. However, the false recognition response totals were subtracted from the total score to correct for guessing (Waring & Takaki, 2003) (Cronbach’s α = .82).

b Translation production

The translation production measure contained the 30 TWs in randomized order and tapped learners’ receptive word knowledge. This task required students to translate, define, or explain the TWs in their L1 (Korean). Word gain was scored as follows: 0 = incorrect; .5 = partially correct; 1 = correct (Cronbach’s α = .85).

c Translation recognition

The translation recognition measure was a multiple-choice test to assess learners’ receptive knowledge of meaning. Each item included the correct meaning, three distractors, and a fifth option: ‘I do not know.’ Each distractor met at least one of the following criteria: (1) contextually proximate to the TW, (2) schematically appropriate, (3) orthographically or phonologically close to another known word in the L1 or L2 (Pulido, 2003). For instance, for the target word, denizen, presented in the following sentence: Denizens of Gangnam are mostly thought to be sophisticated and cultured, the choices provided were (1) denizen (correct), (2) hub (contextually proximate), (3) denial (orthographically or phonologically similar to another known L1 or L2 word), and (4) luxury (schematically appropriate). These four choices were written in participants’ first language (Korean) (Cronbach’s α = .86).

5 Procedure

This study was conducted in four sessions over a period of eight weeks. The experimental conditions were the same for each of the intact classes. Parental permission forms were obtained prior to the beginning of the sessions. During Session 1, students were first informed about the study and its procedures, given a chance to ask questions, and completed the informed consent. At the end of the first session, students completed a background questionnaire and the topic interest inventory. A week later during Session 2, participants completed the reading section of the National Academic Achievement Test of English, the vocabulary pretest, and the prior topic knowledge test. Two weeks later, at the beginning of Session 3, participants were assigned the reading tasks and informed that there would be comprehension questions after they read the passages; however, there was no mention of the vocabulary tests. To control for ordering effects, reading passages were assigned to the participants in a counterbalanced manner. Each group of participants read⁶ the first assigned passage (Psy or Medieval) for 12 minutes, then completed the interest rating scale, and answered multiple-choice comprehension questions written in their L1. Students then read the second assigned passage for 12 minutes, again followed by the interest rating scale and multiple-choice comprehension questions. Afterwards, participants performed an intervening number task to minimize the effect of immediate memory of the target words. Participants then completed the immediate vocabulary posttest, which was a combination of a word-form recognition test, a translation production test, and a translation recognition test. Four weeks later during Session 4, participants completed the same vocabulary posttests to measure vocabulary retention.

6 Analysis

A repeated measures MANCOVA was conducted to test the significance of the main effects of topic interest, L2 reading proficiency, gender, and time (immediately after reading vs. four weeks after reading) on the two dependent variables associated with vocabulary acquisition: word-form recognition and translation recognition. L2 reading proficiency served as a covariate. Two- and three-way interactions under investigation were also entered into the model. The data met the assumptions of the repeated measures MANCOVA.

The translation production score did not satisfy the normal distribution assumption, so it was transformed into a binary variable, either 0 = no gain or 1 = any sign of gain (Pulido, 2003). Generalized estimating equations (GEE) were used for the translation production because of their ability to analyze a binary outcome with repeated measurements (Zeger & Liang, 1986). Also entered into this model were the main effects, two-way and three-way interactions included in the MANCOVA. Because one MANCOVA and one GEE were completed, a Bonferroni adjustment required an alpha level of .025 (.05/2) for all statistical tests. If multivariate results were significant at the alpha level of .025, an alpha level of .05 was used for follow-up univariate analyses.

IV Results

Table 2 displays the descriptive statistics for the independent variable, covariate, and control variables. An independent-samples t test revealed that there was no significant difference in L2 reading proficiency between males and females, t (133) = −0.74, p = .460, suggesting that L2 reading proficiency was stable across gender.

Table 2.

Descriptive statistics for independent variables, covariate, and control variables.

Variables	‘Psy’s Gangnam Style’		‘Medieval life’
Variables	M	SD	M	SD
Pre-reading topic interest:	5.26	1.37	2.10	1.31
Male	5.44	1.13	2.65	1.57
Female	5.20	1.44	1.91	1.16
Post-reading topic interest:	4.78	1.39	2.50	1.40
Male	4.88	1.34	2.68	1.55
Female	4.74	1.41	2.45	1.35
L2 reading proficiency:	68.78	18.28	68.78	18.28
Male	66.77	19.05	66.77	19.05
Female	69.46	18.06	69.46	18.06
Prior topic knowledge:	6.70	1.01	6.61	1.40
Male	6.62	1.10	6.79	1.39
Female	6.73	0.98	6.55	1.40
Prior TW knowledge:	0.05	0.25	0.02	0.15
Male	0.09	0.38	0.06	0.24
Female	0.04	0.18	0.01	0.10

Note. TW = target word. Topic interest maximum = 7. L2 reading proficiency maximum = 100. Prior topic knowledge maximum per story = 8. Prior TW knowledge maximum per story = 15. N = 135 (Male = 34, Female = 101).

Table 3 illustrates the descriptive statistics for the dependent variables, which include gain and retention in word-form recognition, translation production, and translation recognition. The mean scores for gain in word-form recognition in the Psy and Medieval passages were 13.51 out of 15 words (90.10%) and 7.93 out of 15 words (52.87%), respectively, suggesting a high level of word gain in the initial step of vocabulary acquisition. Floor effects were observed for both gain and retention in translation production for the Psy and Medieval passages. For example, students accurately translated fewer than two words (9%) from each story. The mean scores for gain in translation recognition in the Psy and Medieval passages were 7.87 out of 15 (52.47%) and 4.27 out of 15 (28.47%). In the delayed posttests, these scores decreased to 5.81 (38.73%) and 3.42 (22.80%), respectively, suggesting a relatively low level of attrition. Overall, there were greater vocabulary gains in the Psy passage than the Medieval passage.

Table 3.

Descriptive statistics for dependent variables.

	Variables	Psy		Medieval
	Variables	M	SD	M	SD
Gain	Word-form recognition:	13.51	6.25	7.93	7.19
	Male	13.94	6.87	6.00	7.45
	Female	13.37	6.06	8.57	7.02
	Translation production:	1.27	2.08	0.35	0.67
	Male	1.52	1.92	0.32	0.60
	Female	1.20	2.14	0.36	0.70
	Translation recognition:	7.87	3.51	4.27	3.25
	Male	7.79	3.10	4.20	3.39
	Female	7.89	3.65	4.29	3.22
Retention	Word-form recognition:	4.47	6.54	1.43	6.01
	Male	4.70	6.86	−0.29	7.07
	Female	4.40	6.46	2.01	5.53
	Translation production:	0.47	0.98	0.16	0.40
	Male	0.43	0.91	0.18	0.39
	Female	0.48	1.00	0.15	0.41
	Translation recognition:	5.81	3.46	3.42	2.60
	Male	5.44	3.44	3.35	2.51
	Female	5.93	3.47	3.44	2.64

Note. Vocabulary gain and retention maximum per story = 15. Gain = immediately after reading. Retention = four weeks after reading. N = 135 (Male = 34, Female = 101).

The first research question asked whether topic interest affects L2 vocabulary acquisition through reading. MANCOVA results revealed a significant effect of topic interest with a moderate effect size, Wilks’s Ʌ = .94, F (2,131) = 3.96, p = .021, η_p² = .06. Follow-up univariate analyses revealed a significant effect of topic interest on word-form recognition, F (1,132) = 4.87, p = .029, η_p² = .04, and a significant effect of topic interest on translation recognition, F (1,132) = 5.03, p = .027, η_p² = .04. The results demonstrated that students learned significantly more words from the story they found more interesting.

The second research question asked whether L2 reading proficiency affects L2 vocabulary acquisition through reading. There was a significant effect of L2 reading proficiency with a large effect size, Wilks’s Ʌ = .62, F (2,131) = 40.13, p < .001, η_p² = .38. Follow-up univariate analyses revealed a significant effect of L2 reading proficiency on word-form recognition, F (1,132) = 61.85, p < .001, η_p² = .32, and a significant effect of L2 reading proficiency on translation recognition, F (1,132) = 63.19, p < .001, η_p² = .32. These findings demonstrate that higher proficiency students acquired significantly more words through reading than lower proficiency students.

The third research question considered whether L2 reading proficiency moderates the effect of topic interest on L2 vocabulary acquisition through reading. There was a nonsignificant interaction between topic interest and L2 reading proficiency, Wilks’s Ʌ = .97, F(2,131) = 2.06, p = .131, suggesting that the effect of topic interest was consistent for higher and lower proficiency students.

The fourth research question asked whether gender affects L2 vocabulary acquisition through reading. There was a nonsignificant effect of gender, Wilks’s Ʌ = 1.00, F (2,131) = 0.35, p = .705, indicating no difference in L2 vocabulary learning through reading between males and females.

The fifth research question considered whether gender moderates the effect of topic interest on L2 vocabulary acquisition through reading. There was a significant interaction between topic interest and gender with a moderate to large effect size, Wilks’s Ʌ = .89, F (2,131) = 7.87, p = .001, η_p² = .11. Follow-up univariate analyses revealed a significant interaction between topic interest and gender for word-form recognition, F (1, 132) = 14.20, p < .001, η_p² = .10, but not for translation recognition, F (1, 132) = 0.13, p = .722. There was a significant mean difference between males and females on the low-interest passage, t (133) = 2.01, p = .046, d = 0.35, demonstrating that females recognized significantly more word forms than males on the low-interest text. Although males recognized more word forms in the high-interest text than females, this difference did not reach significance.

The sixth research question considered whether the above effects were observed over time. The repeated measures MANCOVA revealed that the significant effects of topic interest, L2 reading proficiency, and topic interest by gender were maintained over time. Although there was a significant interaction between topic interest and time on both word-form recognition, F (1, 132) = 4.90, p = .029, η_p² = .04, and translation recognition, F (1, 132) = 8.34, p = .005, η_p² = .06, the follow-up mean comparisons revealed that the difference between the high- and low-interest texts in both word-form recognition and translation recognition was significant immediately after reading and four weeks later. For word-form recognition, on the immediate test, there was a significant mean difference between high and low interest passages, t (133) = 10.77, p < .001, d = 0.93. On the delayed test, there was a significant mean difference between high and low interest passages, t (133) = 8.24, p < .001, d = 0.71. For translation recognition, on the immediate test, there was a significant mean difference between high and low interest passages, t (133) = 13.14, p < .001, d = 1.15. On the delayed test, there was a significant mean difference between high and low interest passages, t (133) = 8.88, p < .001, d = 0.76. There was a nonsignificant interaction between L2 reading proficiency and time, Wilks’s Ʌ = .99, F (2,131) = 0.59, p = .554, indicating that the strong impact of L2 reading proficiency was consistent across time. There was a nonsignificant three-way interaction between topic interest, gender, and time, Wilks’s Ʌ = 1.00, F (2,131) = 0.25, p = .782, suggesting a similar interaction between topic interest and gender on word-form recognition in both gain and retention.

The impact of the independent variables and their interactions on the translation production was tested via GEE. The analysis revealed a significant effect of L2 reading proficiency, b = 0.05, p = .004, with an odds ratio of 1.05 (95% CI [1.02, 1.08]). The odds ratio suggests that each 1-point increase in L2 reading proficiency increased the odds of acquiring vocabulary through reading by 5%. None of the other variables were found to be statistically significant.

V Discussion

The findings support the first research hypothesis that topic interest significantly affects incidental vocabulary learning, showing that learners gained significantly more new vocabulary from the story they found more interesting. The results can be explained by Laufer and Hulstijn’s (2001) involvement load hypothesis, which posits that long-term retention of a word depends on level of involvement, where the construct of involvement includes both motivational and cognitive factors. For the motivational dimension of involvement, Laufer and Hulstijn (2001) argued that intrinsic motivation more strongly contributes to learners’ involvement in a task and affects their retention of unfamiliar vocabulary. Thus, students’ intrinsic interest in ‘Psy’s Gangnam Style’ may have resulted in engagement in the reading task and greater word gains. This finding supports inclusion of a motivational component in the involvement load hypothesis, which expands the cognition-centered depth-of-processing theory (Craik & Lockhart, 1972). This study affirms that both cognitive and motivational factors play an important role in L2 learning.

This theory-based interpretation is supported by two causal relationships observed in previous research: (1) topic interest facilitates reading comprehension (Ainley, Hidi, et al., 2002), and (2) reading comprehension influences incidental vocabulary acquisition (Pulido, 2004). Furthermore, de Bot, Paribakht, and Wesche (1997) explained that for L2 vocabulary acquisition to occur through reading, readers must first pay attention to a new word form. Then, readers must closely attend to the link between the lexical form and its referent in order to construct text meaning. Given this mechanism, learners interacting with texts depicting topics that are more interesting to them allocate more attention to those texts and the vocabulary contained within them, consequently learning more new words. In this study, the interest produced by ‘Psy’s Gangnam Style’ is assumed to have influenced the level of involvement with the text and the construction of text meaning, including meanings of any new words encountered within the text. This involvement may have spilled over into vocabulary gain and retention. In attending to the neglected affective variable of topic interest in L2 vocabulary research, this study expands prior L1 research (Elley, 1989) and refines the previous model of L2 lexical development through reading.

The second research hypothesis that L2 reading proficiency would be a significant predictor of vocabulary learning through reading was supported through its consistently strong impact across all dependent measures. These results provide further evidence for the Matthew Effect (Stanovich, 1986), a phenomenon whereby the higher an individual’s L2 reading proficiency, the more L2 words they gain through reading (Elgort & Warren, 2014; Webb & Chang, 2015).

The third hypothesis that the effect of topic interest on L2 vocabulary acquisition through reading would differ by students’ level of L2 reading proficiency was not supported. The results revealed a nonsignificant interaction of topic interest and L2 reading proficiency, indicating that the effect of topic interest is stable across all levels of L2 reading proficiency. This finding differs from Walker et al.’s (1979) L1 reading studies, which found that lower proficiency students were better facilitated by high topic interest. It also differs from Endo’s (2010) L1 vocabulary research, which revealed that topic interest significantly affected higher proficiency students’ vocabulary gain, but not lower proficiency students’ gain. This L1 research indicates the limited role for topic interest, but the present L2 study illustrates that learners at all proficiency levels learn a greater number of new words when reading more interesting stories in the L2.

For the fourth hypothesis, there was no significant main effect of gender in L2 vocabulary acquisition through reading. This finding aligns with Brantmeier (2003) and Young and Oxford (1997), who demonstrated no significant gender differences in L2 reading comprehension. This finding expands previous reading studies to the area of L2 incidental vocabulary learning research by showing that gender differences do not account for differences in lexical development when reading in an L2.

The fifth hypothesis that gender would moderate the effect of topic interest on L2 incidental vocabulary learning was partially supported. While gender did not have an overall significant main effect across the three vocabulary gain measures, there was a significant interaction between gender and topic interest in word-form recognition. For the lower interest text, females recognized significantly more word forms than males. In contrast, for the higher interest text, males recognized more word forms than females, although the difference did not reach statistical significance. One possible explanation for these results can be derived from Ainley, Hillman, et al. (2002) and Pulido (2004). Ainley, Hillman, et al.’s (2002) path analyses illustrated that female adolescents are more likely to continue reading lower interest texts, while male adolescents tend to discontinue reading. Graham et al. (2008) explains this phenomenon via gender differences in achievement orientations, considering gender as a ‘system of values that encompasses attitudes, approaches, and behaviors’ (p. 757). For instance, when faced with a choice of activities, girls select reading because of the high value that they assign to reading, whereas boys choose physical activities involving displays of masculinity. According to their preferential selection of certain activities, boys and girls learn the skills related to those particular activities, which influences gender differences in text engagement. Female participants in this study, in spite of their overall lower interest level, persistently processed the lower interest passage such that they were able to recognize significantly more lexical forms than the male participants. In contrast, male participants may have made less effort when they were less interested in the topic and were subsequently unable to recognize as many word forms as the female participants. This finding extends previous reading research reporting a significant interaction between interest and gender in L2 vocabulary research and highlights the role of gender overlooked in the literature on L2 vocabulary development.

As for the long-term effects, the sixth hypothesis was confirmed: significant effects were maintained over time. The results revealed that topic interest exerted a significant effect over time: learners retained significantly more vocabulary from more interesting texts. As the involvement load hypothesis suggests, a learner’s intrinsic interest may have influenced engagement in text processing to construct meaning and build stronger links to long-term memory and word retention. This explanation is consistent with Elley’s (1989) proposition that learners’ interest in stories produces intrinsic motivation to sustain attention and enhances long-term vocabulary retention. The results also showed that the positive impact of L2 reading proficiency remained four weeks after reading across the three vocabulary retention measures. These findings illustrate the substantial and consistent impact of L2 reading proficiency on the long-term retention of words. The significant interaction between topic interest and gender on word-form recognition also remained four weeks later. This finding is noteworthy because previous research has shown a significant interaction between interest and gender immediately after reading but not in long-term retention. By demonstrating that gender can influence the degree of word retention resulting from level of involvement, these findings expand on the involvement load hypothesis. Indeed, they point to another learner variable in SLA: gender.

In terms of limitations, only students from a Korean middle school participated in the study; replications of this study with other populations are needed to confirm our results. In addition, this study used only two expository passages, one higher and one lower interest. The use of multiple passages for each interest level could increase generalizability.

VI Pedagogical implications

This study offers implications for language educators in facilitating L2 vocabulary learning through reading. First, the significant positive impact of topic interest on L2 vocabulary acquisition for learners of all proficiency levels supports teachers’ and material developers’ efforts to provide topics of interest to maximize students’ learning from reading. Furthermore, the interaction of topic interest and gender in L2 text processing offers pedagogical insights on how gender differences affect reading and lexical development. Given the adverse influence of low-interest topic texts on males’ literacy development, language educators should address the scarcity of topics related to conventionally male areas of interest (e.g. science and sports) compared to topics typically preferred by females (e.g. humanities-oriented topics). Educators can try to balance between these topics in their curriculum.

VII Conclusions

This study investigated the impact of topic interest on L2 incidental vocabulary acquisition as well as how topic interest interacts with L2 proficiency and gender. This research contributes to the literature in several ways. First, the study provide further empirical support for the involvement load hypothesis (Laufer & Hulstijn, 2001) by demonstrating the role of motivational and cognitive factors in incidental vocabulary learning. It also confirms the interaction between topic interest and gender, suggesting that the degree of word retention resulting from the level of involvement can be influenced by gender, which has received inadequate attention in L2 vocabulary research.

Given that maximizing engagement underlies all effective vocabulary teaching and learning (Schmitt, 2010), topic interest is worth considering in both theoretical research and in classroom practices. Such engagement leads to students’ enjoyment of and immersion in reading-based vocabulary building, an essential process in second/foreign language development.

Footnotes

Appendix

Appendix 2.

Target words.

	‘Psy’s Gangnam Style’	‘Medieval life’
Nouns	hub	toil
	throng	fealty
	fanatic	cathedral
	denizen	allocation
	component	obligation
Verbs	reel	grant
	portray	pledge
	perceive	derive
	attribute	comprise
	contradict	emancipate
Adjectives	viral	meager
	absurd	arduous
	eccentric	majestic
	whimsical	obsolete
	contagious	adequate

Declaration of conflicting interest

The authors declare that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

References

Ainley

Hidi

Berndorff

(2002). Interest, learning, and the psychological processes that mediate their relationship. Journal of Educational Psychology, 94, 545–561.

Ainley

Hillman

Hidi

(2002). Gender and interest processes in response to literary texts: Situational and individual interest. Learning and Instruction, 12, 411–428.

Brantmeier

(2003). Does gender make a difference? Passage content and comprehension in second language reading. Reading in a Foreign Language, 15, 1–27.

Bray

G.B.

Barron

(2004). Assessing reading comprehension: The effects of text-based interest, gender, and ability. Educational Assessment, 9, 107–128.

Brysbaert

Warriner

A.B.

Kuperman

(2014). Concreteness ratings for 40 thousand generally known English word lemmas. Behavior Research Methods, 46, 904–911.

Carrell

C.L.

Wise

T.E.

(1998). The relationship between prior knowledge and topic interest on second language reading. Studies in Second Language Acquisition, 20, 285–309.

Craik

F.I.M.

Lockhart

R.S.

(1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671–684.

Davies

Gardener

(2010). Word frequency list of American English: 20,000 Word List. Available at: http://www.wordfrequency.info (accessed February 2016).

de Bot

Paribakht

T.S.

Wesche

M.B.

(1997). Toward a lexical processing model for the study of second language vocabulary acquisition: Evidence from ESL reading. Studies in Second Language Acquisition, 19, 309–329.

10.

Elgort

Warren

(2014). L2 vocabulary learning from reading: Explicit and tacit lexical knowledge and the role of learner and item variables. Language Learning, 64, 365–414.

11.

Elley

W.B.

(1989). Vocabulary acquisition from listening to stories. Reading Research Quarterly, 24, 174–187.

12.

Endo

Y.A.

(2010). The effects of topic interest on the vocabulary retention of third grade students with and without learning disabilities. Unpublished doctoral dissertation, Columbia University, New York, USA.

13.

Erçetin

(2010). Effects of topic interest and prior knowledge on text recall and annotation use in reading a hypermedia text in the L2. ReCALL, 22, 228–246.

14.

Graham

Tisher

Ainley

Kennedy

(2008). Staying with the text: The contribution of gender, achievement orientation, and interest to students’ performance on a literary task. Educational Psychology, 28, 757–776.

15.

Hidi

Renninger

K.A.

Krapp

(2004). Interest, a motivational variable that combines affective and cognitive functioning. In: Dai

D.Y

Sternberg

R.J.

(Eds.), Motivation, emotion, and cognition: Integrative perspectives on intellectual functioning and development (pp. 89–115). Mahwah, NJ: Erlbaum.

16.

Joh

(2006). What happens when L2 readers recall? Language Research, 42, 205–238.

17.

Laufer

Hulstijn

J.H.

(2001). Incidental vocabulary acquisition in a second language: The construct of task-induced involvement. Applied Linguistics, 22, 1–26.

18.

Langley

(1996). Medieval life. New York: DK Publishing.

19.

LeLoup

J.W.

(1993). The effect of interest level in selected text topics on second language reading comprehension. Unpublished doctoral dissertation, The Ohio State University, Columbus, OH, USA.

20.

Mochida

Harrington

(2006). The yes/no test as a measure of receptive vocabulary knowledge. Language Testing, 23, 73–98.

21.

Nation

I.S.P.

(2001). Learning vocabulary in another language. Cambridge: Cambridge University Press.

22.

Ortega

(2014). Understanding second language acquisition. New York: Routledge

23.

Pulido

(2003). Modeling the role of second language proficiency and topic familiarity in second language incidental vocabulary acquisition through reading. Language Learning, 53, 233–284.

24.

Pulido

(2004). The relationship between text comprehension and second language incidental vocabulary acquisition: A matter of topic familiarity? Language Learning, 54, 469–523.

25.

Schmitt

(2008). Instructed second language vocabulary learning. Language Teaching Research, 12, 329–363.

26.

Schmitt

(2010). Researching vocabulary: A vocabulary research manual. New York: Palgrave Macmillan.

27.

Stanovich

K.E.

(1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360–407.

28.

Stevens

(1980). The effect of topic interest on the reading comprehension of higher ability students. Journal of Educational Research, 73, 365–368.

29.

Walker

S.M.

Noland

R.G.

Greenshields

C.M.

(1979). The effect of high and low interest content on instructional levels in informal reading inventories. Reading Improvement, 16, 297–300.

30.

Waring

Takaki

(2003). At what rate do learners learn and retain new vocabulary from reading a graded reader? Reading in a Foreign Language, 15, 130–163.

31.

Webb

Chang

C.-S.

(2015). How does prior word knowledge affect vocabulary learning progress in an extensive reading program? Studies in Second Language Acquisition, 37, 651–675.

32.

Young

D.J.

Oxford

(1997). A gender-related analysis of strategies used to process written input in the native language and foreign language. Applied Language Learning, 8, 43–73.

33.

Zeger

S.L.

Liang

K.Y.

(1986). Longitudinal data analysis for discrete and continuous outcomes. Biometrics, 42, 121–130.