Anna C-S Chang<\/strong> Research has shown that second language (L2) learners generally lack multiword expression knowledge, and L2 researchers and practitioners have tried various techniques to assist L2 learners to acquire it more efficiently. This study adopted an under-researched technique\u2014repeated oral reading\u2014to enhance the retention of high-frequency multiword items by 62 EFL college students divided into experimental (n =38) and control (n =24) groups. Fifteen unfamiliar multiword items comprising only known individual words were selected through a pre-test based on a theme-based text. All students received a formal instruction first, followed by the experimental group orally reading the text six times under a time constraint. A two-week delayed post-test was used to test students\u2019 retention of four dimensions of multiword knowledge: aural forms and aural meanings, and written meanings and use. Except for use, the experimental group significantly outperformed the control group in the other three dimensions. Four fixed factors (oral reading speed, prior vocabulary knowledge, dimensions of multiword knowledge, and the number of words per item) were analyzed via GLMM. Results showed three factors had significant effects on retaining multiword items except oral reading speed. Based on the results, pedagogical implications are discussed and suggestions are made.<\/p>\n Keywords<\/em><\/strong>: reading aloud, oral reading, repeated oral reading, multiword items, high-frequency words<\/p>\n Recent research found that L2 learners lack multiword item (MWI) knowledge, and they are familiar with far more high frequency individual words than MWIs that comprise high frequency words (e.g., Kim, 2016; Nguyen & Webb, 2017; Park & Chon, 2019); however, studies have also shown that approximately 50% to 70% of spoken and written discourse consists of fixed MWIs (Erman & Warren, 2000). An MWI is used in this study as an umbrella term for strings of language items containing more than a single word (Siyanova-Chanturia & Omidian, 2020), which may be seen as \u201cidioms,\u201d \u201ccollocations,\u201d \u201clexical bundles,\u201d or \u201clexical chunks\u201d in the literature. Many researchers have investigated how to increase learning effectiveness through various instructional methods, such as looking up the target items for meanings in dictionaries (Dziemianko, 2010; Laufer, 2011); deliberately learning unfamiliar collocations through their L1 translations or by doing various exercises (Webb & Kagimoto, 2009, 2011); using sound repetition (Lindstromberg & Boers, 2008); grouping idioms by themes or different verbs (Zyzik, 2011), to name only a few; however, one old-fashioned but commonly used technique\u2014reading aloud (used interchangeably with oral reading), is apparently under-researched or a missing piece in learning MWIs (Alali & Schmitt, 2012).<\/p>\n Oral reading is often considered boring, dated, or anxiety-provoking with learners barking out texts without comprehension (see Gibson, 2008 for the strengths and weaknesses of oral reading), but many psycholinguistic studies have shown the importance of vocalization in learning an L2. According to Baddeley (2013), the auditory memory traces decay very rapidly in the phonological loop of working memory; however, they can revive through articulatory rehearsal and be restored in the long-term memory system. Many autonomous learners also use this technique as part of private study to practice pronunciation and vocabulary, to build up confidence in speaking, and to improve fluency (Gibson, 2008). Moreover, teachers use it as a diagnostic tool to check students\u2019 pronunciation problems, graphemic-phonemic connections, etc. Reading aloud, when repeated many times, is called repeated oral reading. Many recent studies have demonstrated the importance of \u201crepetition\u201d or \u201cfrequency of encounter\u201d in L2 vocabulary learning (see Uchihara, Webb, & Yanagisawa, 2019), but few studies have examined how repeated oral reading affects MWI learning. Repeated oral reading offers at least three advantages over facilitating learning. Firstly, it is repetitive, which may increase the number of opportunities for encountering the same vocabulary words; secondly, vocalizing the newly learned linguistic elements may help learners to better retain what is learned (Gathercole & Conway, 1988). Thirdly, students enjoy the experience of bringing the language to life through their voice when they practice class content orally (Chang, 2019). For these encouraging reasons, the current study adopted repeated oral reading to examine the extent of EFL (English as a foreign language) students\u2019 MWI learning. If repeated oral reading can be demonstrated to improve retention of MWI learning for L2 learners, it will increase the practice varieties for vocabulary learning.<\/p>\n Repeated oral reading has been used extensively in L1 contexts for developing reading fluency and literacy, as well as for remedial purposes to assist struggling readers (Rasinski, Homan, & Biggs, 2009). In L2 contexts, repeated oral reading is normally conducted in chorus (see Alali & Schmitt, 2012) rather than individually; this is because individual oral reading is very time consuming and many language teachers simply cannot afford the time to do it. Though some researchers were able to employ individual oral reading, the number of participants was limited (e.g., Chang, 2019; Lin, 2016), and the main focuses were on improving oral reading rates, comprehension, and pronunciation rather than on vocabulary retention (Shimono, 2019). In the limited literature, three relevant studies into reading aloud on vocabulary learning, are discussed below:<\/p>\n To investigate whether doing oral practice or written practice makes a difference in acquiring vocabulary knowledge, Alali and Schmitt (2012) conducted a study of 35 Kuwaiti female junior high school students\u2019 acquisition of single words and idioms. In the study, oral practice involved the students reading target formulaic sequences 10 times in chorus for ten minutes; in the written review, students worked in groups for 10 minutes on a written recall task. Thirty idioms were selected; an unknown word from each idiom made up the target single words. The results showed that oral practice and written recall resulted in a similar pattern of learning of single words and idioms. However, written repetition was consistently more effective than oral reading for form and meaning recognition and recall. Students who did oral practice retained 0.50\/10 words for the idiom form recall task and 2.30\/10 for the meaning recall in the 12-day delayed post-test. Acquisition rates for idiom recall tasks were not very satisfactory, possibly because oral reading was done as a group; it may be possible to improve learning rate by conducting oral reading on an individual basis. Another significant reason could be that each idiom contained an unknown word, which might have increased the difficulty in acquisition (Kim, 2016; Zyzik, 2011). Could the acquisition rate be enhanced if all items comprised of familiar words? This study hence attempts to fill the two gaps.<\/p>\n Another study by Durrant and Schmitt (2010) examined the effects of learning 20 pairs of words with adjectives and nouns under three different conditions: presenting each collocation in a sentential context only once; orally repeating a collocation two times, and presenting a target collocation in different contexts. Participants were 84 non-native adult speakers of English from various countries. An uninformed cued recall test was given after the treatment. Results showed that the two experimental repetition conditions had an effect superior to the no repetition condition. The verbatim repetition of a single sentential context (Cohen\u2019s d<\/em> = .56; Mdn<\/em> = 4.5\/10) also revealed better learning outcomes over the exposure to varied contexts (Cohen\u2019s d<\/em> = .48; Mdn<\/em> = 5.0\/10). The authors concluded that reading identical sentences aloud twice might ease the cognitive load and hence improve collocational memory trace. Additionally, oral reading allowed participants to dedicate more attention to the language.<\/p>\n A more recent study by Chang and Chen (2022) investigated the effects of written and oral vocabulary exercises on EFL learners\u2019 retention of single-word items (SWI) and multi-word items (MWI). After receiving formal instruction of four texts, the written exercises group did gap-filling, meaning-matching, and rearranging scrambled sentences, whereas the oral vocabulary exercise group was given a list of the target items embedded in sentences and orally practiced the target items six times. It must be mentioned that students\u2019 oral reading speeds were not measured. Sixty unfamiliar vocabulary items were tested in four dimensions of vocabulary knowledge: aural form and meaning recall, L2 written form and meaning recall. Mixed findings were found in the one-week delayed post-tests. The results showed that students performed significantly better on the MWIs than SWIs regardless which types of vocabulary exercises were adopted. In terms of vocabulary dimensions, students performed best for L2 written meaning recall (66%), followed by aural form recall (59%), aural meaning recall (51%), and L2 form recall (50%). According to the authors, all SWIs were unfamiliar to students in all dimensions tested, which made it difficult for them to guess its meaning, but the MWIs were comprised mostly of known individual words, which might have made the tests easier for them to guess meanings from the other words in the MWI.<\/p>\n Taken altogether, although no conclusive findings were found, the studies suggest that repetition may play some role in vocabulary learning, and varying types of repetition also make some differences in learning outcome. One area that is not clear is whether a learner who reads more fluently can retain more vocabulary knowledge, because oral reading fluency has been used to measure a learner\u2019s reading competence or their familiarity with the content (e.g., Fuchs, Fuchs, Hosp, & Jenkins, 2001; Jiang, 2016). Furthermore, the better a learner is, the more fluently they can read with appropriate prosody (Grabe, 2009; Kuhn & Stahl, 2003). Having said this, no L2 research up to the present has investigated whether L2 learners who read a text more fluently can retain more MWIs. Therefore, in the present study, student participants\u2019 reading speed was included as a variable to examine its effect on learning multiword items.<\/p>\n Many studies that took into account learners\u2019 prior vocabulary knowledge have shown that learners with a larger vocabulary size tend to pick up more words from reading (e.g., Horst, Cobb, & Meara, 1998; Webb & Chang, 2015; Zahar, Cobb, & Spada, 2001) or viewing (Peters & Webb, 2018). For example, Horst, Cobb, and Meara (1998) examined the relationship between vocabulary knowledge and incidental learning gains through reading and listening to a graded reader. Students\u2019 vocabulary knowledge was measured through the Vocabulary Levels Test (Nation, 1983). A medium-sized relationship was found for the two variables. Similar results were shown in an extensive reading program by Webb and Chang (2015), whose students were divided into high-, intermediate-, and low- levels according to their scores on a pretest measuring knowledge of target vocabulary. They found that the more prior vocabulary knowledge students had, the greater their gains from extensive reading. Measuring students\u2019 meaning recall and meaning recognition in their two separate experiments, Peters and Webb (2018) also reported a positive relationship between learners\u2019 prior vocabulary knowledge and vocabulary learning through viewing TV.<\/p>\n Although the conclusions drawn above are specific to acquisition of single words, a study by Peters (2016) suggests that these findings can be generalized to MWI. Peters found that the larger a learner\u2019s vocabulary size, the higher the odds that they could recall a new collocation; specifically, she found that the odds to recall a new collocation in the first form recall test were 13 times higher for every 100 words more in a learner\u2019s vocabulary size. Taken together, these research findings suggest that prior vocabulary knowledge may affect the ability of L2 learners to acquire single word. The reason for this may be that students with greater vocabulary knowledge are likely to have greater text comprehension (Laufer, 2013; Webb & Chang, 2015), and this may allow them to pay greater attention to unknown words.<\/p>\n The characteristics of MWI may affect the extent to which an MWI can be learned. These characteristics involve the following: whether the items being semantically transparent or opaque (Gyllstad & Wolter, 2016; Macis & Schmitt, 2017), the presence of known or unknown words in an expression (Kim, 2016), the combination (verb-noun, adjective-noun) of collocations (Peters, 2016), number of node words (Webb & Kagimoto, 2011), and thematical groupings (Zyzik, 2011). However, few studies have taken into account how the numbers of words in a MWI affect learning outcome. In general, the longer a single word is, the more difficult it is to learn. Campoy (2008) found that shorter words (four-phonemes) were better recognized than strings of long words (six-phonemes). Peters (2016) also found an MWI that comprises of longer individual words, the lower chance for a correct response. In the same vein, an MWI containing three or more words is supposed to be more difficult to learn than one consisting of only two words. Hence, this study included this variable to determine how it affected MWI acquisition.<\/p>\n Word knowledge is not an all-or-nothing construct. According to Nation and Webb (2011), knowing a word involves 18 dimensions. Some learning tasks may be more effective in gaining receptive knowledge (e.g., how does the word look like?) whereas others in productive knowledge (e.g., how is the word spelled?). Therefore, it is important to measure different aspects of word knowledge in a test because it \u201cmay provide a much more clear evaluation of the relative efficacy of tasks\u201d (Webb, 2005, p. 50). Many recent studies involved several tests to measure students\u2019 learning efficacy. For example, Teng (2016) measured his students\u2019 gains from reading a graded reader in four dimensions: form, grammar, meaning, and collocation. Pigada and Schmitt (2006) measured a learner\u2019s spelling, meaning, and grammatical characteristics gains from reading four texts. The two studies reported that their students gained significantly more on spelling than on grammar. In a long-term narrow reading study, Chang (2019), however, found that L2 students made significantly more improvement in their knowledge of written and aural meanings than for spelling and use. In general, requiring students to produce answers is always more difficult than having them to select one from a few options, but to determine which aspects of knowledge should be measured must depend on the learning conditions and tasks.<\/p>\n Based on the limited number of studies reviewed above, the effects of repeated oral reading on L2 learners\u2019 vocabulary acquisition are mixed (Alali & Schmitt, 2012; Chang & Chen, 2022; Durant & Schmitt, 2010). The effects of reading aloud in chorus may be different from the effects of reading aloud individually. This study therefore adopted individual repeated oral reading. Fifteen unknown MWIs comprising only known individual words were selected from a theme-based story related to \u201ctime.\u201d Four dimensions of vocabulary knowledge were assessed: spelling, aural meaning, written meaning, and use. Two research questions were addressed. The first question focused on comparing the difference in retention for four dimensions of multiword items between the experimental group and the control group. The second research question explored mainly the learning effectiveness for the experimental group in terms of their vocabulary knowledge, vocabulary dimensions, reading rates, and the number of words in each item. The two research questions are as below:<\/p>\n RQ1: Were there any significant differences in the retention for four dimensions of multiword items between the experimental group and the control group? Sixty-two 18-19 year-old university freshmen majoring in hospitality and finance took part in the study. These students were randomly grouped by the university by their majors, but all of them were required to take Freshman English as one of their general education courses. Chinese was their native language. To measure students\u2019 sight and spoken vocabulary knowledge, 30 target words that are used in the first 1,000-word level of the New Vocabulary Levels Test (NVLT, Webb, Sasao, & Balance, 2017)[1]<\/a> were selected to test the students\u2019 productive spoken forms and meanings. The students were allowed 30 seconds to orally read the selected words, and another 10 minutes to write the meanings of the words. Each correct item was awarded one point. The test results were used to classify students\u2019 vocabulary knowledge into three levels. Those who answered an average of 25 or more items correctly on both tests were classified into the high group, between 18 and 24 items the intermediate group, and below 18 the low group. Results showed that 25 students were in the high group (n =13, 12), 19 in the intermediate (n = 13, 6), and 18 in the low (n = 12, 6). The numbers in parentheses were the number of students in the experimental and control groups respectively. Grouping idioms into thematic categories makes idioms easier to learn and memorize than unrelated ones (Boers, 2000; Cooper, 1998), so the target items for this study were all about the same theme. The reading text was selected from Idiom Magic<\/em> by John Ryan (1994). Idiom Magic<\/em> was written based on different themes using stories, and the theme\u2014time\u2014<\/em>was selected for the present study (see Appendix A<\/a> for the whole text). The text analyzed by BNC\/COCA, shows that there are 332 words (8 words are contracted forms, so it was 324 words in total), with 122 word families and 158 word types (see Table 1). Despite the text being short, it contains many multiword units, most of which are related to \u201ctime,\u201d such as once upon a time<\/em>, on time<\/em>, in time<\/em>, the time of one\u2019s life<\/em>, for decades<\/em>, and for ages<\/em>. The text was divided into three paragraphs for students to do oral reading practice, and each paragraph contained 117, 96, and 111 words respectively (see Appendix A<\/a>).<\/p>\n Table 1. Text Analysis<\/strong><\/p>\n To select the target unknown MWIs, a pre-test containing 24 MWIs embedded in sentences (different sentences from those in the delayed post-test) was administered to all the student participants. The pre-test contained two steps. The first step was listening for form (spelling) and aural meanings. Students were asked to spell the idioms they heard and give a corresponding Chinese meaning for each item. The second step was to translate the target items (presented in isolation) into Chinese and make a sentence if they could. They also had to circle the unfamiliar words. After marking, 9 items were excluded because some items contained unfamiliar words, such as for decades,<\/em> and some items were answered correctly by all or the majority of the students, e.g., have a hard time<\/em>, have a good time<\/em>, and on time<\/em>. Finally, 15 target multiword expressions were selected based on the following criteria: They were items for which no students gave full correct answers for aural-form meaning and written-form meanings; every item was related to \u201ctime;\u201d each unit contained no unfamiliar single words; every target item was listed in the Corpus of Contemporary American English (COCA); and all items appeared only once in the text. An analysis of all the words used in the 15 target items showed that all words were from the first 1,000 level. Among the 15 MWI, nine items were from the first paragraph, three from the second, and three from the third paragraph. The length (the number of words) of an MWI may also affect students\u2019 acquisition of it. Among the 15 selected MWIs, four items have 2 words, three items 3 words, five items 4 words, and three items 5 words. The complete profile of the target MWIs is presented in Table 2.<\/p>\n Table 2. Profile of the Target Multi-word Items<\/strong><\/p>\n
\nHsing Wu University
\n<annachang![\"atmark\"](\"http:\/\/www.tesl-ej.org\/atmark.png\")
livemail.tw><\/p>\nAbstract<\/h3>\n
Background<\/h3>\n
Repeated Oral Reading in L2 Vocabulary Learning<\/h4>\n
Learners\u2019 Prior Vocabulary Knowledge and Vocabulary Learning<\/h4>\n
Number of Words in Multiword Items<\/h4>\n
Measuring Different Dimensions of Vocabulary Knowledge<\/h4>\n
\n<\/a>
\nRQ2: To what extent was the learning effectiveness through repeated oral reading influenced by students\u2019 vocabulary knowledge, dimensions of vocabulary, reading speed, and number of words per multiword unit?<\/p>\nMethod<\/h3>\n
Participants<\/h4>\n
\n<\/a><\/p>\nStudy Materials<\/h4>\n
\n\n
\n Word levels<\/strong><\/td>\n Word counts<\/strong><\/td>\n Word types<\/strong><\/td>\n Word families<\/strong><\/td>\n<\/tr>\n \n 1,000<\/td>\n 310<\/td>\n 93.37%<\/td>\n 144<\/td>\n 91.14%<\/td>\n 114<\/td>\n<\/tr>\n \n 2,000<\/td>\n 5<\/td>\n 1.51%<\/td>\n 4<\/td>\n 2.53%<\/td>\n 3<\/td>\n<\/tr>\n \n 3,000<\/td>\n 2<\/td>\n 0.6%<\/td>\n 2<\/td>\n 1.27%<\/td>\n 2<\/td>\n<\/tr>\n \n 4,000<\/td>\n 1<\/td>\n 0.3%<\/td>\n 1<\/td>\n 0.63%<\/td>\n 1<\/td>\n<\/tr>\n \n 5,000<\/td>\n 2<\/td>\n 0.6%<\/td>\n 2<\/td>\n 1.27%<\/td>\n 2<\/td>\n<\/tr>\n \n Proper nouns<\/td>\n 0<\/td>\n 0%<\/td>\n 0<\/td>\n 0%<\/td>\n 0<\/td>\n<\/tr>\n \n Off list<\/td>\n 12<\/td>\n 3.61%<\/td>\n 5<\/td>\n 3.16%<\/td>\n <\/td>\n<\/tr>\n \n Total<\/td>\n 332<\/td>\n <\/td>\n 158<\/td>\n <\/td>\n 122<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n Target Multiword Items<\/h4>\n