51

English Teaching, Vol. 67, No.2, Summer 2012

Paraphrase Practices for Using Sources

in L2 Academic Writing*

Yeon Ree Choi

(Ewha Womans University)

Choi, Yeon Ree. (2012). Paraphrase practices for using sources in L2 academic
writing. English Teaching, 67(2),51-79.

The present study explores the effects of two types of paraphrase practices (teacher-led
and Web-based) on Korean university and graduate students' paraphrase awareness,
performance, and types. The teacher-led and Web-based group both had four-week
intensive practices on paraphrasing including instruction on plagiarism, citation, and
paraphrasing strategies with teacher modeling of paraphrasing. Teacher-led practices
placed a focus on expl.icit instruction of grammar and vocabulary, whereas Web-based
practices triggered learner-d irected practices. The Web-based group was informed of
Internet sites and trained to use them for paraphrasing. The resu lts of the study show
positive effects of both practice types on paraphrase awareness and performance, but not
on paraphrase types. More beneficial effects of teacher-led practices were noted over
that of Web-based practices, though the latter were more positively evaluated by the
paJticipants. The fmdings suggest the positive effects of short-term explicit instruction
on paraphrasing to prevent plagiarism; however, they also imply positive impact of
Web-based practices for long-term learning.

I. INTRODUCTION

With an increasing number of English-mediated courses in L2 context (ESL and EFL

contexts), academic writing in English has become a day-to-day task. Academic writing
from sources has been a central issue in L2 writing research and instruction over the past
era (Carson, 200 1; Choi, 2009; Johns, 1993; Keck, 2006). Before they write on a given
topic using sources, L2 student writers are expected to read and think critically about the

• This work was supported by the National Research Foundation of Korea Grant funded by the
Korean Government (NRF-20 11-327-A00656).
52 Choi, Yeon Hee

topic (Weigle, 2002). Such a writing task requires students to make an incorporation of
their own opinions "textually with existing knowledge from sources" (Shi, 2004, p. 171).
As academic writing tasks using sources have become prevalent, more L2 writing
teachers and researchers have noticed cases of plagiarism in students' written texts.
Plagiarism is "the representation of a source's words or ideas as one's own" including
"cheating, non-attribution of sources, and patchwriting" (Howard, 1995, p . 799).1 When
applied in academic settings, plagiarism tends to be viewed as an act of academic
dishonesty (Howard, 1995). Therefore, both ESL and EFL writing studies have attempted
to reveal the main causes of plagiarism and students' perceptions of it in order to prevent
the illegitimate act and help L2 students write fully in their own words (Currie, 1998;
Deckert, 1993; Howard, 1995; Hsu, 2003; Hu & Lei, 2011; Pecorari, 2003; Pennycook,
1996; Prochaska & Moon, 2004; Suh, 2008; Sutherland-Smith & Carr, 2005; Wheeler,
2009).
Negative interpretations of the motivations of plagiarism such as an absence of ethics or
an ignorance of citation acknowledgements were aroused in early studies (Howard, 1995).
Nevertheless, attempts have been made lately to comprehend the causes of plagiarism,
especially within the frameworks of EFL academic composition settings, including the
incongruities between students' academic workloads and their still developing linguistic
and cognitive resources (Penny cook, 1994), differences of culturaJ perceptions (Prochaska
& Moon, 2004; Wheeler, 2009), and language proficiency (Keck, 2006; Thomas, 2010).
These factors have been understood as not an intention to deliberately deceive on behalf of
the writers, but rather as a demonstration of their consistently developing abilities in text-
accountable writing (Chandrasoma, Thompson, & Pennycook, 2004; Currie, 1998;
Howard, 1995; Pecorari, 2003). These elements, in turn, have also influenced subsequent
studies to investigate proper ways to help students overcome with the very problems. The
EFL studies conducted to examine students' perceptions of plagiarism in academic writing
have correspondingly furthered more research to notice the significance of teaching aptly
about what plagiarism is and training EFL student writers to prevent it effectively
including inter-textual practices (Lee, 2007; Park & Lee, 2010; Sung, 2011; Wette, 2010).
As the rise of the importance of raising awareness of what plagiarism is and of the
significance of training on avoiding it in L2 academic context, paraphrasing practices
when integrating source texts in academic writing have been suggested as one of the
pedagogical interventions (Campbell, 1990; Currie, 1998; Howard, 1995; Hyland, 2001;
Johns & Mayes, 1990). Parapbrasing2 can provide students to practice combining the

1 Patchwriting refers to "copying from a source text and then deleting some words, altering
grammatical structures, or plugging in one-far-one synonym-substitutes" (Howard, 1993, p. 233).
2 Paraphrasing refers to rewriting the source text in the writer's own words by using different words
and sentence structures from the source. It demonstrates the writer's understanding of the source
 Paraphrase Practices for Using Sources in L2 Academi c Writing 53

ideas of a given source text with their own point of view being freed from making exact
copies of the original sentences (Keck, 2006). Likewise, it is suggested as a strategy to
avoid plagiarism in university-level writing instruction including on line programs such as
Purdue University's Online Writing Lab. Containing the essence of the original source text
and expressing this into one's own words and sentences is what makes the paraphrasing
valuable as an academic writing strategy.
Mostly used strategies noted in a number of studies when attempting to paraphrase a
given source text in academic writing include replacement of original words with
synonyms and modification of sentence structures (Swales & Feak, 2004). in order to
create an acceptable paraphrase, however, these simple strategies are not sufficient; the
provision of proper paraphrasing practices is unquestionably required in which it could
offer students to capture the true essence of what paraphrasing is and to break away from
close reliance of simple rearranging strategies (Keck, 2006). Then, the key issue is how to
train student writers effectively to paraphrase source texts in their academic writing
without committing plagiarism. A few studies on the effect of instruction on paraphrasing
or using sources in L2 academic writing (e.g., Sung, 2011 ; Wette, 2010) have not
specifically explored whether the effects would vary with practice techniques or
procedures. Rather, they mainly have dealt with the influence of intervention or practice.
To see whether L2 student writers would have different benefit from different types of
intervention, the present study aimed at exploring two types of paraphrasing practices:
teacher-led practices (TL) which place a main focus on explicit instruction of grammar and
vocabulary with step-wise teacher demonstration of paraphrasing; and Web-based
practices (WB) emphasizing students to exploit various online resources and programs for
constructing a paraphrase. 3 The previous studies on L2 academic Wliting (e.g., Currie,
1998; Hsu, 2003; Qian & Krugly-Smolsk, 2008; Sung, 2011; Wette, 2010) noted lexical
and syntactic reformulation as a key problem in paraphrasing due to lack of vocabulary
and grammatical knowledge. Thus, both practice types put a focus on how to search or
choose synonyms and syntactic transformation. However, the two types of practice were
differentiated by the degree of teacher control and learner autonomy, and the availability
of resources. Unlike the teacher-led group, the Web-based group was allowed to
paraphrase the given text using Internet resources and an online plagiarism check tool. TL
practices were selected since L2 students often favor explicit instruction over implicit
instruction or self-learning. WB practices were chosen as learner-directed learning. L2

text. " Paraphrased material is usually shorter than the origina l passage, taking a somewhat broader
segment of the source a nd condensing it s lightly" (retrieved on September. 26, 20 II , from
http://owl.english.purdue.edulowllresource/564/0 I).
3 In the study, the tel111 S ' practice,' ' training,' and ' interventi on ' a re exchangeab ly used to refer to
the in-c lass interventi on and practices.
54 Choi, Yeon Hee

students are usually involved in their academic writing including paraphrasing written
sources by themselves without their teachers' presence and online sources such as online
dictionaries or thesaurus can be good sources for their paraphrasing. WB practices were
assumed to promote learner autonomy or learner-directed paraphrasing process.
With an aim at offering empirical evidence on effectiveness of training students with
appropriate paraphrasing strategies by practice types, two groups of Korean university and
graduate students were trained for four weeks (14 hours and 7 sessions). Their paraphrase
awareness (including the identification of paraphrase types) and perfonnance were
compared before and after the practices, and their perceptions on paraphrase and practices
including instructional techniques were also analyzed in order to find an answer for the
following questions.

1. What are the effects of paraphrase practices on paraphrase awareness?

2. What are the effects of paraphrase practices on paraphrasing perfonnance?
3. What are the effects of paraphrase practices on paraphrase types (their frequency and
ratio)?
4. What are the effects of paraphrase practices on L2 student writers' perception on
paraphrase awareness and practices?

II. PREVIOUS STUDIES ON PARAPHRASING

As attention has increasingly been given to the issue of plagiarism, a number of studies
have investigated students' use of source texts in composing academic texts (Hyland, 2009;
Keck, 2006; Qin, 2008; Rhee, 2010; Shi, 2004; Suh, 2008). These empirical studies have
generally paid more attention to analyzing and identifYing the way students use source
texts and paraphrase, especially in tenns of comparing Ll and L2 writers. Campbell
(1990), for instance, compared textual borrowing in the expository writing of Ll and L2
students, including their use of quotations, exact copies, near copies, paraphrases,
summaries and original explanations. The results of the study showed that copying was a
major strategy, regardless of the students' native language, which led to a conclusion that
writing by synthesizing source infonnation is difficult for both L 1 and L2 student writers.
However, a contradictory fmding was reported in Shi's study (2004), which was set to
examine the effects of students' Ll (language proficiency) and different types of writing
tasks (a summary and an opinion task by using two pre-selected source texts) on amount
and types of textual borrowing in L2 (English) writing. The comparison of Ll and L2
writers' use of textual borrowing strategies illustrated that the summary task group
borrowed more words than the opinion task group, and Chinese students borrowed
 Paraphrase Practices for Us ing Sources in L2 Academic Writing 55

significantly more words from the source texts without citing references for both tasks.
Studies on Ll and L2 students' use of source texts have led researchers to consider the
ways to avoid committing plagiarism. A number of researchers have agreed on teaching
paraphrasing as one of the effective pedagogical interventions (Campbell, 1990; Currie,
1998; Howard, 1995; Hyland, 2001). According to Campbell (1990), paraphrasing is a
prerequisite strategy for students in order to integrate source texts into their writing
without copying the texts. Despite of the recognition of paraphrasing practice being
effective, little research has attempted to observe accurately students' paraphrasing
strategies and to suggest ways teach them effectively. The initial studies placed a focus on
students' realization of plagiarism instances and paraphrasing examples after instruction.
Roig (2001) examined the effects of instruction about plagiarism identification and proper
paraphrasing skills on Ll college students' ability to detect plagiarism. Participants of the
study were divided into four groups: a control group (no treatment) and three experimental
groups (feedback only group, plagiarism examples group, and a combination of feedback
and examples group). The findings from the study indicated that the three experimental
groups outperformed the control group. Additionally, the participants in the example
groups plagiarized less than the other groups in paraphrasing exercise. Moon (2002) also
reported positive effects of instruction on plagiarism for L2 lmiversity students, as the
incidents of direct copying from the source texts decreased. In Barry's study (2006), L2
students wrote definitions of plagiarism before and after a six-week practice of
paraphrasing and citing original sources. A significant increase was found in post-
paraphrasing plagiarism definition scores. After paraphrasing practice, the students were
more likely to include two additional specific elements of plagiarism in their definitions
("taking someone else's ideas is plagiarism and not giving credit is plagiarism") besides
the notion "taking someone else's words is plagiarism" (p. 377). Barry discussed the
importance of having students practice paraphrasing techniques, rather than merely
teaching them definitions of plagiarism.
While the above-mentioned studies have acknowledged paraphrasing as an effective
skill of helping students avoid apparent text copying, the majority of them have not set out
to actually deal with teaching and training students appropriately to paraphrase source
texts in academic writing. This issue has been explored by Wette (2010), who carried out
an action research to observe the effects of instructional interventions with an aim at
improving L2 undergraduate students' writing skills in using sources. After an 8-hour unit
of instruction on appropriate ways of acknowledging sources, paraphrasing, summarizing
and citing, the post-unit scores revealed improvement in L2 students' declarative
knowledge and in the rule-governed aspects of the skill, with a decrease in a number of
direct copying from the source texts. Park and Lee (2010) investigated the effects of one-
session plagiarism-prevention practices on Korean university students' essay writing test.
56 Choi, Yeon Hee

The nwnber of the sentence-level and word-level plagiarisms declined after the practices.
Sung (2011) explored the effect of paraphrasing and quotation practices by writing task
types: summary and opinion writing with source texts. The results of the study showed that
Korean undergraduate students benefited from the 5-session intervention on the use of
quotation and paraphrasing in English. The comparison of pre- and post-paraphrasing test
results indicated the positive effects of intervention on developing L2 students' ability to
paraphrase appropriately. Additionally, the post-survey responses illustrated that the
practices were helpful to enhance paraphrasing skills, regardless of the task types.
These enlightening and informative previous studies confirm that practices or
intervention can have a positive influence on raising L2 writers' awareness of plagiarism
or paraphrasing and developing paraphrasing strategies. However, the majority of studies
did not suggest an effective way of teaching paragraphing. In fact, it was suggested that
incidental plagiarism could attribute to students' inadequate knowledge of paraphrasing
and citation (Moon, 2002; Rhee & Nam, 2010; Roig, 1997) and insufficient linguistic
ability such as vocabulary and grammar knowledge (Hsu, 2003; Sung, 2011; Wette, 2010).
Therefore, further studies seem to be needed to provide clearer empirical evidence of an
effective way to teach paraphrasing strategies to prevent plagiarism. Moreover, none of the
previous studies has explored whether the effects of paraphrasing intervention would vary
with instructional techniques or procedures. By recognizing the room for further research,
the present study was designed with an ultimate goal to propose an effective way to train
L2 student writers to employ appropriate paraphrasing strategies in writing using source
texts. Additionally, the need of paraphrasing practice was recognized from a theoretical
stance that emphasizes development from declarative knowledge to procedural knowledge
through extensive practices (Wette, 2010).

III. RESEARCH METHOD

1. Participants

Twenty-four Korean university students were recruited for the study, including 19
undergraduate (one freshman, eight sophomores, six juniors, and four seniors) and five
graduate students in the age of20 to 30. Their major and L2 proficiency were controlled to
use source texts in hwnanities and social science, and make sure that they have relatively
sufficient L2 proficiency to paraphrase L2 source texts. Thirteen subjects were majoring in
education; seven, social science; three, language; and one, business. All the participants
had a TOEIC or iBT TOEFL score, except for one. The average TOEIC score of fifteen
subjects was 857.9 and the average iBT TOEFL score of ten subjects was 102. Most of
 Paraphrase Practices for Using Sources in L2 Academic Writing 57

them had some experience of learning composition in English or Korean. All of them
stated in the pre-questionnaire that they had some experience of revising their own writing;
however, only about a half of them received formal instruction on revision.
Based on the scores of an English argumentative writing test adm inistered before the
practices, the participants were randomly divided into two groups: 13 participants for the
teacher-led practice (TL) (11 undergraduates and 2 graduate students) and II for the Web-
based practice (WB) (8 undergraduates and 3 graduate students). The result of t-test
illustrates no significant difference between the two groups (TL, M = 3.72, SD = .35; WB,
M = 3.70, SD = .35; t = .108,p = .915).

2. Paraphrase Practices

The in-class procedures and techniques of paraphrasing practice were devised for two
types, that is, TL and WB groups (see Table 1). Regardless of the types, a practice cycle
was designed for each 90-minute practice session: 30-minute in-class individual
paraphrasing or revising using Microsoft Word without any interaction with the instructor;
20-minute in-class paraphrasing with teacher help if requested, followed by the
presentation of a student's paraphrase with class discussion and teacher comments. Next
30 minutes were assigned to teacher lecture on paraphrasing strategies including a
demonstration of step-wise paraphrasing for the TL group, or teacher demonstration
of paraphrasing using online sites and resources for the WB group, as shown in Table 1.
This cycle was repeated six times over three weeks, twice per week.

TABLE]
In-class Practices Cycles
Teacher-led practice (TL) Web-based practice (WB)
Step I: Indi vidual Individual paraphrasing or revising Individual paraphrasing using online
paraphrasing or the first draft of paraphrasing with sites, or revising the first draft of
revising (30 mins.) teacher feedback paraphrasing with teacher feedback
using online sites
Step 2: Individual Individual paraphras ing using the Individual paraphrasing using online
paraphrasing and dictionary, and one student' s sites, and one student's presenting
presentation (20 presenting the paraphrasing to the the paraphrasing to the class,
mins.) class, followed by teacher comments followed by teacher comments
Step 3: Teacher Teacher's step-wise lecture on useful Teacher's modeling of how to use
lecture or modeling paraphrasing strategies various online sites to paraphrase a
(30 mins.) given passage
Step 4: Whole-class Whole-class paraphras ing with the Whole-class paraphrasing with the
paraphrasing (10 teacher teacher, using online sites
mins.)
58 Choi, Yeon Hee

There was a difference in the initial step of the two sessions of each week. Both groups
were asked to paraphrase a given text on the first session and revised their drafts of
individual paraphrasing with the teacher feedback provided on the second session. They
were allowed to ask questions to the instructor or get help from him, if necessary, and to
use dictionaries. Additionally, the WB group utilized online tools and resources.
In the second step of the cycle, one of the participants was requested to project his/her
paraphrase to the classroom screen to share it with the other participants in each practice
group and the instructor. Then, the instructor made comments on the parts that were wrong
or could be revised.
As for the third step, the TL group received an explicit instruction on paraphrasing
strategies, such as replacing original words with synonyms, connecting sentences, and
using different sentence structures (see Figure 1). The instructor also explained the implicit
meaning (nuance) of the words to show the participants why a certain synonym can be
used while others cannot be for paraphrasing the source text. Moreover, he demonstrated
techniques for paraphrasing concisely by deleting redundant words and expressions, with
the text given for the individual paraphrasing activity in the second step. At the end, the
instructor compared the source text with a model paraphrased text. The WE group had the
sanle procedures. However, its third step was distinctive in that the instructor provided a
list of online sites and resources useful for paraphrasing and a step-wise demonstration of
paraphrasing using online sites (e.g., thesaurus.com) to search for the meaning of the
words or synonyms. Additionally, the group participants utilized paraphrasing self-test
sites (e.g., plagiarismchecker.com) to examine how many words in their own
paraphrased text were overlapping with the words used in the source text (see Figure 2).
Such sites led the WE group participants to pay more attention on their own paraphrasing
processes from word-level to sentence-level paraphrases.
The final step was whole-class paraphrasing. The instructor attempted to paraphrase a
given passage with the whole class through interaction. He encouraged the students to
partake actively in the activity by asking questions on synonyms and sentence structures.
The same procedure was implemented for the WB group, except for using online sites.
Each participant completed three drafts of individual paraphrasing and revision, another
three individual paraphrasing tasks, and three whole-class paraphrasing tasks. Thus, they
practiced paraphrasing including the revision task for 12 times, in total. The WB group
used online sites for each paraphrasing, unlike the TL group.

3. Materials

The research materials included an English writing test, a pre- and post-practice test of
paraphrasing, in-class paraphrasing tasks, source texts, a pre- and post-questionnaire, and
 Paraphrase Practices for Using Sources in L2 Academic Writing 59

FIGUREl
Sample Handout for the Teacher-led Practice Group

Original Passage (Individual Paraphrasing 3)

[The] top-down [facet of] listening involves the [listener's] abili ty to bring pri or information to
[bear on the task of] understanding the "heard" language.

Paraphrasing through deleting words

o top-down [c=JJ listening involves the re=J] ability to bring prior informat ion to
"-___---'h understanding the "heard" language.
I. The word "facet" adds almost nothing to the meaning of "top-down .. . listening."
2. This deletion takes along "The" and "ill" since "ill" belongs to "facet" and "The" is required
by "Qf."
3.The possessive " listener's" is implied in "top-down . .. listenin g."
4. "bear on the task of'= " use in"
- to bring prior information to bear on the task of understanding
- to bring prior information to use in understanding
- to bring prior information to understanding

FIGURE2
Sample Handout for the Web-based Practice Group
Paraphrase Self-test
T ype som e th ing in as your base text :

At rnp In ri ; ... I " " ,,' I fl'VR I , "'0 ..... pprfnr.Anr.R t;.",n 1:",> Rftl'lr;t;;;i ~y !'lIH- .-. prc;t;IRMI'I ::>:c::
lo ss 0 1 sleep, I lln ess. a nd drugs . Ot h er pe r son a l f act ors could I n c lud e f aMi Iy
prob l eMS . l ack of socia l acceptan ce . l ack o f I n te re st in t he work . a nd so o n .
I n such cirCUMs t ances. th e s o l utions t o the prab l ellls do n o t l ie IUll y w i th in
the organi zation ' s purview . Easing t he per s o n ' s work l oad 1Ia,y b e ap p r opriate
a~ III tG .. porQry 9><podlont. b u t It Ie unllko l y to bo a ",I(lb l o l ong-tG .... C"o l utlo n .
Fu rthe r Mo r e , t el l i n s an e.plo),ee h ow to 90 l Ye h is o r h er pers onal p r o bl e_ Is
unllke l )' to be e ff ecti v e . A so l ution app r op ri ate l or t h e person giv ing the
a d vice May be totally I n approp r iate o r even u n acc eptab l e for t he pe r son "l t M
t h e p r obl e_

N ow try to paraphrase that text u s ing your own words :

Per s onal problellts s uc h as Il l ne ss o r falli Iy prob l e", can influence on wo r k. The
probll;l_$ ca n ' t bQ ... olvQd by or!;Janlzat lon ' II p u r view. Rodu c ln!:l .ork l oad I ...
on l y a l fectl ... e as a s h or t - t el"_ solu t ion . Moreo ... er, giv i ng a n eMployee so lut ion s
Is not ef f ee t Ive a nd Is not acc eptab l e. 1

Now s ee if you used any overlappin g language:

I WO di'ttere nt ways O f l ooki ng a t the diff e r e n ces lcolor-coded) ;

Ann ~ nl;::\in- t pxt lonk :

60 Choi, Yeon Hee

in-class materials including source texts.

1) Tests, In-class Paraphrasing Tasks and Source Texts

An English writing test was constructed to examine the homogeneity of two groups of
the participants in L2 writing proficiency before the practices. It was a 30-minute
argumentative writing task on the topic of cyber-Iearning (online learning) and traditional
classroom teaching chosen from the Test of Written Examination (TWE).
A pre- and post-practice 40-minute paraphrasing test were also developed which
contained two sections. The first section was composed of objective items to measure the
participants' ability to judge appropriateness of paraphrasing as their paraphrasing
awareness. The items were adopted from Oshima and Hogue (2006, pp. 129-131). This
section was repeated in both pre- and post-practice test to measure any changes in
paraphrasing awareness. The second section was a 30-minute paraphrasing task of a given
source text which was designed to measure the participants' ability of paraphrasing the
text. The pre-practice topic was 'international language' (117 words) from an academic
writing textbook written by Gates (1977) (cited from Oshima & Hogue, 2006); the post-
practice topic was 'personal problems' (116 words) from an academic textbook on
leadership (Wright & Taylor, 1994). The prompt included the following task directions: 1)
Do not change the meaning of the original passage; and 2) Do not summarize and
paraphrase sentence by sentence. The use of dictionary was not allowed; there was no limit
on the number of words.
Additionally, three sets of paraphrasing tasks were devised: three 30-minute in-class
individual paraphrasing tasks with 130-140 word source texts; three IS-minute in-class
individual paraphrasing tasks with 90-110 word texts; and three 10-minute whole-class
paraphrasing tasks with 90-110 word texts. There was no limit on the number of words,
and dictionaries were allowed to use. The nine source texts were selectively chosen from
academic textbooks varying in topics such as humanities, social studies, education, and
language learning and teaching (e.g., goal clarity, active play, and washback).
The researcher, the research assistant, and the native instructor carefully reviewed the
number of words, topics, and content as well as linguistic difficulty of the source texts in
order to maintain the consistency across the source texts used for the pre- and post-practice
tests, and for each type of in-class activities. The topics were controlled in consideration of
the majors of the participating students.

2) Pre- and Post-Questionnaire

A pre-questionnaire and a post-questionnaire were constructed for the study. The pre-
 Paraphrase Practices for Using Sources in L2 Academi c Writing 61

questionnaire contained two sections. The first section was the questions of background
information such as major, previous experience of learning English writing, and
standardized English test score (five yes-no questions, two choice questions, and 13 open-
ended questions). The second part contained the questionnaire of participants' perceptions
of paraphrasing awareness (including plagiarism recognition), ski lls, and practices
including in-class procedures and practice-specific techniques (20 six-Likert- scale items
and eight open-ended questions). The post-questionnaire included the second part of the
pre-questionnaire to measure any changes in the participants' perceptions and further had
new questions to investigate how each practice group perceived its own paraphrasing
practices.

3) In-class Practice Materials

Materials for in-class lectures, tasks, and demonstrations were developed by using
online materials, including those from the Purdue University Online Writing Lab
(http ://owl.english.purdue.edu/), Indiana University Writing Resources Web page
(http://www.indiana.edu/~wts/pamphlets . shtml), and AP A In-Text Docwnentation. They

were differentiated according to the practice type. For the TL group, handouts and PPT
slides were constructed to provide the list of useful expressions and vocabulary, the ways
of transfonning sentence structures for paraphrasing, a step-wise demonstration of
paraphrasing, and model paraphrases (see Figure 1). Conversely, materials for the WB
group mainly contained guidelines of utilizing online sites and programs for paraphrasing,
besides model paraphrases (see Figure 2).

4. Data Collection Procedure

For four weeks in the summer of2011 , the paraphrasing practices were conducted by a
native English speaking writing teacher who was a university English program professor
with a doctoral degree in history and had twenty-five years of teaching experience in
academic writing in Korea as well as in the United States. On the first session, the
participants' English writing proficiency was measured through a 30-minute test of
English argumentative writing. Their background information including L2 academic
writing was collected through the pre-questionnaire. Their paraphrase awareness and
performance were measured. Furthermore, their perceptions on plagiarism, paraphrasing,
and paraphrasing practice technjques were surveyed. In the second part of the first session,
the instructor gave a 90-minute lecture on the definition and examples of plagiarism,
appropriate ways of citation, and paraphrasing strategies.
On the subsequent six 90-minute practice sessions, the TL and WB practices were
62 Choi, Yeon Hee

scheduled differently. The participants had a practice cycle, as mentioned before:

individual in-class paraphrasing or its revision, another individual in-class paraphrasing
followed by the presentation of a student's paraphrase, teacher's lecture or demo, and
whole-class paraphrasing. Each practice session was videotaped to supplement two
research assistants' observation notes of the contents of the teacher's instruction and
feedback, the participants' participation, interactions between the teacher and the
participants, and the online sites used. As for the Web-based practice, another video
camera was set in order to track the participants' use of online sites and programs.
After the practices, the participants' paraphrase awareness and performance were
measured again. Their perceptions on plagiarism, paraphrasing, and paraphrasing practice
techniques were resurveyed and their evaluation of or satisfaction with group-specific
practice techniques was also surveyed.

5. Data Analysis

1) Scoring of the Tests

The English argumentative writing test administered prior to the practices was assessed
by two raters, using the Test of Written Examination (TWE) scoring criteria (the 0 to 5
scale holistic rubric). They were the native English instructor for the training and a
research assistant who was a doctoral student in English Education with a MA in TESOL.
The inter-rater reliability was .868.
The first part of the pre- and post-paraphrase objective test items of paraphrase
awareness asked the participants to determine the most appropriate paraphrase between the
two sample paraphrases of the original text and provide reasons. Each correct answer was
given one point, resulting in the total of two points. For the second part of the objective
items, the participants were required to classify the four samples of paraphrase into 'best,'
'too similar,' 'no citation,' and 'inappropriate. ' One point was given for each correct
answer, which tallies up to four points in total.
The participants' pre- and post-practice, and in-class paraphrases were analytically
scored using the rubric that was specifically constructed for the present study by the
researcher, the native English instructor, and the two research assistants (see Appendix A).
Since no adoptable paraphrasing rubric was available in the previous studies, a mbric was
developed, incorporating the defmitions of the paraphrase provided by the writing center
of the University of Texas at Austin (retrieved on September 4, 2011, from http://www.
uwc.utexas.edulnode179). It included five dimensions: content in terms of quality and
conciseness, a degree of paraphrasing (linguistic reformulation), language use, and citation.
Based on the importance of each dimension, different scores were designated: 10 for
 Paraphrase Practices for Using Sources in L2 Academic Writing 63

content (five for quality and conciseness, respectively) and for linguistic reformulation,
five for language use, and two for citation (27 points in total). As any demonstration of
native-like fluency was not the main concern of the study, a relatively small number of
scores were assigned to language use, compared to the first two dimensions. In terms of
citation, which was the matter of using either accurate or inaccurate citation forms, a full
score of two was given if the participants used an accurate form. Score of one was given to
partially accurate forms and zero to no citation and no reference. The final version of the
rubric was carefully reviewed by the native English-speaking instructor. All the participants'
pre- and post-practice paraphrasing performance was rated by three research raters (the
native instructor and two research assistants), and any discrepancies between them were
solved through discussion. Additionally, the two research assistants also scored the in-
class paraphrasing. The inter-rater reliability of pre- and post-practice paraphrasing tests
and in-class paraphrasing tasks was .738 to .840.

2) Analysis of Questionnaire

The participants' responses for the Likert-scale items of the pre- and post-questionnaire
were analyzed by encoding them for statistical analyses. Their answers for the open-ended
questions were carefully analyzed and categorized, and then the frequency of each category
was counted.

3) Analysis of Paraphrase Types

The taxonomy of five paraphrase types was recreated by adapting Keck 's taxonomy
(2006) to measure the degree of paraphrasing in the study: exact copy, near copy, minimal
modification, moderate modification, and substantial modification (see Table 2).
The participants' texts were carefully examined in order to distinguish attempted
paraphrases from exact copy. An attempted paraphrase refers to a sentence consisting of a
single word replaced with synonyms without distorting original meaning (Keck, 2006). To
assign each attempted paraphrase into one of the five categories, the unique links were
identified and their number was counted on the basis of the total number of words in each
attempted paraphrase. Unique links refer to "individual lexical words (i.e., nouns, verbs,
adjectives, or adverbs), or exactly copied strings of words used in the paraphrase that (a)
also occurred in the original excerpt but, (b) occurred in no other place in the original text"
(Keck, 2006, p. 266). They are differentiated from general links, which are " lexical words
used in the paraphrase that occurred in the original excerpt but that also occurred
elsewhere in the original text" (Keck, 2006, p. 267). The criteria used to define each
paraphrase type and its examples are displayed in Table 2.
64 Choi, Yeon Hee

TABLE 2
Definition of Paraphrase Types and Examples of Attempted Paraphrases
Category Definition Original sentence Attempted paraphrase
Exact copy 100% of copy ing from The question is - which The question is - which
the original text language? language? (100%)
Near copy 50% or more of the In such circumstances, the In such conditions, the
paraphrase is made up solutions to the problems solutions to the problems do
of words contained do not lie fu ll y within the not solve perfectly within the
with in unique links. organization ' s purview. organ ization ' s purview.
(68 .75%)
Minimal 20-49% of the Because the Internet The impOitance of common
modification paraphrase is made up makes the world a smaller language has increased along
of words contained place, the va lue of having with the introduction of the
within unique links. a common language is Internet that makes the world a
greatly increased. smaller place. (33 .3%)
Moderate 1-1 9% of the At the individual level, At the personal level, work
mod ification paraphrase is made up work performance can be achievement can be resu lted
of words contained affected by such problems from such problems as sholt of
within uniq ue links. as loss of sleep, illness, sleep, ailing, and drugs.
and drugs. (15 .8%)
Substantial 0% of the paraphrase Easing the person's work Even lessening the burden of
mod ification is made up of words load may be appropriate work can't be the ultimate
contained within as a temporary expedient, solution. (0%)
unique links. but it is unlike ly to be a
viable long-term solution.
Note: italicized expressions are unique li nks; underlined expressions refer to general links.

One of the two research assistants identified the five types by the percentage of words in
unique links out of the total number of words per attempted paraphrase. A moderate
modification, for example, is identified if an attempted paraphrase contains I to 19
percentage of words in unique links (e.g., when the total number of words is 19 and three
words are unique links; 3 out of 19 words is 15.8 %, as in Table 2). The researcher
randomly examined the accuracy of the identification; the other assistant reviewed all of
the identification. Any discrepancies were solved through discussion. The inter-coder
reliability was not measured due to the lack of independent coding. After coding each type
of meaning-preserving attempted paraphrases, the frequency was counted and then the
ratio was calculated for statistical analyses.
 Paraphrase Practices for Using Sources in L2 Academic Writing 65

4) Statistical Analyses

In order to examine the effects of paraphrasing practices, pre- and post-practice

paraphrase test scores, ratios of paraphrase types, and questiOlmaire responses were
statistically analyzed by independent and paired samples t-test. The samp le size was small;
however, the results of the Shapiro-Wilk test illustrated that each of the two practice
groups followed a normal distribution and the results of Levene 's test confirmed that the
two groups had the same variance. Additionally, the ANCOVA was emp loyed to see the
practice group differences after the practices.

IV. RESULTS AND DISCUSSION

1. Pre- and Post-Practice Paraphrase Awareness

For the first research question, the participants' paraphrase awareness including the
identification of paraphrase types was measured before and after the practices. As shown
in Table 3, both practice groups illustrated higher awareness after the practices. Such a
positive effect was also noted in Sung (2011) and Park and Lee (2010). This implies that
the knowledge on the definition and nature of paraphrase can be acquired through a short
period of intervention, though it is not necessarily connected to the development of
paraphrasing skills. In the open-ended questions, some participants expressed that they
realized what paraphrasing or plagiarism is, as a strong point of the practice (3 TL and 1
WE students). This finding is also consistent with Barry (2006) and Wette (2010), which
illustrate that an intensive intervention for paraphrasing or using academic texts leads Ll
or L2 student writers to enhance their knowledge of plagiarism or academic conventions
including citation.

TABLE 3
ANCOVA Results of Post-practice Paraphrase Awareness Scores by Practice Type
Pre-practice Post-practice Adjusted
Group F p
M SO M SO M SE
TLgroup 2.23 1.54 3.62 2.26 3.63 .50 1.239 .278
WB group 2.27 1.19 2.82 1.60 2.80 .55

No statistically significant difference was noted between the two practice types in the
results of ANCOVA (see Table 3). Nonetheless, the t-test results of the pre- and post-
practice paraphrase awareness scores in each group indicate a significant difference in the
66 Choi, Yeon Hee

TL group (t = -2.374, P = .035), but not in the WB group (t = -1.257, P = .237). This
suggests a more positive effect of teacher-led practices, that is, explicit instruction, on the
enhancement of paraphrase awareness.

2. Pre- and Post-practice and In-class Paraphrasing Performance

The participants' paraphrasing performance was measured before and after the practices
as well as during the practices (in-class) to explore the second research question (see
Table 4). The quality of their performance gradually enhanced over four weeks, which
suggests the positive effects of practice on the development of paraphrasing skills .
Interestingly, however, the scores of the post-practice performance were lower than those
of the third in-class performance. This may be attributed to the resources allowed for the
in-class paraphrasing: the TL group was allowed to use dictionaries and the WB group, to
use online resources and programs.
No statistically significant differences between the two practice types were noted in the
results of t-test in the in-class paraphrasing perfonnance as well as in the pre- and post-
practice paraphrasing performance (see Table 4). The results of ANCOVA also reveal no
statistically significant differences between the two practice types (total, F = 1.58 7, P
= .222; content quality, F = .085, p = .773; content conciseness, F = .672, P = .422;
linguistic fonnulation, F = 1.113, P = .303; language use, F = .602, P = .446; citation, F =
3.384, P = .080) .
In line with the results of paraphrase awareness, the results of t-test illustrate more
significant differences between the pre- and post-practice paraphrasing scores of the TL
group than between those of the WB group, as shown in Table 5. Significantly higher post-
practice scores of linguistic reformulation, language use, and citation noted from the TL
group suggest that step-by-step teacher modeling of paraphrasing and explicit instruction
on lexical and syntactic variations for paraphrasing can enhance L2 student writers '
linguistic perfonnance and citation for paraphrasing even after a short period of intensive
practice, compared to leamer-autonomous practices.
Table 5 also illustrates a variation in different dimensions of paraphrasing performance:
paraphrasing concisely and citation skills seem easier to acquire than the ability for the
content quality of paraphrase (see Table 4). No significant difference was noted in the
latter, regardless of practice types. These findings show that acquiring the knowledge of
citation and concise paraphrasing can lead to the development of the participants' paraphrasing
skills through a short period of intensive practice, compared to other sub-paraphrasing skills.
Wette's study (2010) also illustrated that mechanical and rule-directed elements of
paraphrasing seem to be capable of being acquired, even after a short period.
 Paraphrase Practices for Using Sources in L2 Academi c Writing 67

TABLE 4
I-Test Results ofPre-, During-, and Post-practice Paraphrase Scores by Practice Type
TL group WB group
P
M SD M SO
Pre-practice 16.46 2.05 16. 15 3.40 .277 .785
Content quality 3.82 0.73 3.73 0.97 .272 .788
Content conciseness 3.36 0.35 3.09 0.49 1.56 1 .1 33
Linguistic refo rmul ation 6. 10 0.97 5.88 1.46 .448 .659
Language use 3.08 0. 56 3.30 0.7 1 -.875 .391
Citation 0.10 0.21 0. 15 0.23 -. 542 .594
In-class 1 19.46 1.73 \7 .95 3.45 \. 386 .180
Co ntent quality 4.62 0.42 4.23 0.96 1.324 .199
Co ntent conciseness 3.8 1 0.33 3.50 0.87 1.190 .247
Linguistic reformul ation 7.3 1 1.38 6.45 1.37 1.5 16 . 144
Language use 3.46 0.52 3.64 0.64 -. 742 .466
C itation 0.27 0.33 0.14 0.23 1.117 .276
In-c lass 2 21.04 2.55 20.59 1.58 .506 .6 18
Content quality 4.38 0.68 4.73 0.52 -1 .365 .1 86
Content co nc iseness 4.62 0.46 4.36 0.32 1.5 14 .144
Linguistic reformul ation 7.54 0.78 7.27 0.90 .775 .447
Language use 3.8 1 0.48 3.86 0 .39 -.308 .76 1
Citation 0.69 0.63 0.36 0.50 1.39 1 .1 78
ill-class 3 2 1.73 1.39 20.59 1.69 1.8 15 .083
Content quality 4.69 0.43 4.77 0.41 -.463 .648
Content co nciseness 4.62 0.36 4.27 0.47 2.024 .055
L inguistic reformul atio n 7.54 0.66 7.09 0.94 1.363 . 187
Language use 3.81 0.38 3.77 0.34 .233 .818
Citation 1.08 0.49 0.68 0.64 1.703 . 103
Post -practice 20. 95 1.83 19.76 2.67 1.290 .2 10
Co ntent qu ality 4.15 0.46 4.21 0.54 -.282 .780
Co ntent conciseness 4.33 0.41 4.15 0.40 1.092 .287
Linguistic reformulation 7.3 1 0.97 6.73 1.44 1.135 .272
Language use 3.64 0.53 3.58 0.54 .296 .770
C itation 1.5 1 0.48 1.09 0.70 1.737 .096

TABLE 5
t- Test Results of Pre- and Post-practice Paraphrase Scores
TLgroup WB group
t p P
Total -9.047 .000 -3.371 .007
Content quality -1.364 .198 -1. 543 .154
Content conciseness -6.395 .000 -6.627 .000
Linguistic reformul ation -4 .364 .00 1 -1.717 .11 7
Language use -3.582 .004 - 1.400 .192
C itation -I 1. 180 .000 -4.376 .00 1
68 Choi, Yeon Hee

The score differences between the pre- and post-practice test of paraphrase (post-
practice changes) illustrate further dimension and practice group variations, as shown in
Figure 3. As discussed before, the TL group gained more abilities in linguistic
reformulation, language use, and citation. On the other hand, the WB group participants
improved their paraphrasing skills in the aspects of content quality and conciseness. This
result can also be observed in the participants' sample paraphrases, as shown in Figure 4.

FIGURE 3
Post-practice Changes in Paraphrase Scores by Practice Type
5.00
4. 00
3.00
2.00
1.00
0.00
Total Content quality Content Lingui stic La nguage use C itation
conciseness refonTIul ation

FIGURE 4
Pre- and Post-practice Paraphrasing Samples
.---------------------------~~
Pre-practice = Post-practice
[Original Passage] [Original Passage]
Someday, software may be available to Furthennore, tell ing an employee how to solve his or
instantly translate both written and spoken her personal problem is wllikely to be effective. A
language so well that the need for any solution appropriate for the person giving the advice
common language could decline . .... (Oshima may be totally inappropriate or even unacceptable for
& Hogue, 2006, p. 148) the person with the problem . .. .. (Wright & Taylor,
Oshima, A., & Hogue, A. (2006). Writing academic 1994, p. 50)
English (4th ed). White Plains, NY: Pearson Wright, P. L. & Taylor, D. S. (1994). Improving leadership
Longman. performance: Interpersonal skills for effective leadership.
~r1g!~o..9(L<::Iilf.~,_NJ : !:t:~!i~e.!:la,II}ll.!.er:t1~on a l .
__. __._._._
[paraphrase ofTL-S8]
Someday, a kind of software that can instantly
= Moreover,
[paraphrase ofTL-S8)
advising an employee can be ineffective as
translate both written and spoken language solutions can be different for people depending on
perfectly might become available, reducing their situations (Wright & Taylor, 1994, p. 50).
the need for any common language.
[Original Passage] [Original Passage]
Because the Internet makes the world a At the individual level, work perfonnance can be
smaller place, the value of having a common affected by such problems as loss of sleep, illness, and
_l<ll1gu<lge_ i~. grc;:<ltly. il1~rc;:<l~c;:t1:. __ . . Ichugs.
[paraphrase ofWB-S6]
We need to have a common language on the
= [paraphrase ofWB-S6)
At the personal level, problems in lack of sleep,
Internet that makes the earth into "A village." sickness and drugs can have an impact on work
achievement.
*Notes: S stands for student; the following number refers to the student number.
 Paraphrase Practices for Using Sources in L2 Academic Writing 69

Figure 4 illustrates that the sample sentences of a TL participant demonstrated the

improvement of citation skills, while the WE participant's samp le sentences revealed the
enhancement of content quality, with the post-practice sentence displaying more
clear comprehension of the original passage's meaning. Likewise, the TL group did not
outperform the WE group in the content quality of paraphrase. It may be accounted by the
focus of the practices and the reading abilities required for the content quality. The
instructor did not place an emphasis on understanding the source text throughout the
practices in both groups, since he assun1ed that the participants had comprehended the text.
Rather, he highlighted how to paraphrase concisely the source text by changing vocabulary
and sentence structures, and deleting unnecessary parts. As Wette (2010) noted, thus,
further studies need to deal with comprehending source texts to enhance L2 student writers'
quality of paraphrasing.

3. Frequency and Ratio of Pre- and Post-practice Paraphrase Types

For the third research question, attempted paraphrases were counted before and after the
practices. Before the practices, the participants made 149 attempted paraphrases, 26 of
which were ones distorting the meaning. The average was 5.2l. The most freq uent type
was minimal modification (M = 2.25), followed by near copy (M = 1.29) and moderate
modification (M = 1.13). After the practices, the number of attempted paraphrases
decreased to 110 including 18 attempted paraphrases distorting the meaning (M = 3.83),
because the participants attempted to paraphrase as concisely as possible since the
instructor placed a focus on conciseness. As a result, their post-practice paraphrase was
much shorter than their pre-practice paraphrase. The average number of words and
sentences were 66.04 and 4.54, respectively, in their post-practice paraphrase of the 116-
word and 6-sentence source passage, whereas their pre-practice paraphrase of the 120-
word and 7-sentence source passage was 107.75 words and 6.25 sentences long. The most
frequent type after the practices was also minimal modification (M = 1.79), followed by
moderate modification (M = 1.04) and near copy (M = .71). The reduction of the
frequency of near copy was more noticeable than the other types of modification.
The pre- and post-practice ratio of paraphrase types was calculated, as shown in Table 6.
In the TL group, no case of exact copy was identified before and after the practices.
The ratio of minimal modification increased and that of moderate and substantial
modification also slightly increased, while that of near copy sharply declined. These
findings imply that the explicit instruction including step-wise paraphrasing (e.g.,
substituting words with their synonyms) led the TL group participants avoid copying the
source sentences. However, it did not enhance their ability to transform sentence structures
and rewrite the sentences in their own words, though their post-practice paraphrasing
70 Choi, Yeon Hee

scores illustrated a statistically significant improvement in the quality of paraphrase (see

Tables 4 and 5, and Figure 3). As for the WB group, the ratio of near copy, and minimal
and moderate modification marginally increased, whereas that of substantial modification
declined. Prior to the practices, two cases of exact copy were found; however, none of
the WB group participants made an exact copy after the practices. This resulted in a
slightly higher ratio of the other types including near copy and minimal modification.

TABLE 6
ANCOVAResults of Post-practice Ratio of Paraphrase Types by Practice Type
Pre-practice Post-practice Adjusted
F P
M SD M SD M SE
Exact copy
TLgroup 0.00 0.00 0.00 0.00 0.00 0.00
WBgroup 3.79 8.63 0.00 0.00 0.00 0.00
Near copy
TLgroup 26.74 27.96 16.54 16.46 17.40 6.44 .060 .809
WB group 19.33 19.28 20.76 29.64 19.74 7.01
Minimal modification
TLgroup 33.70 25.12 42.44 31.16 43.86 9.11 .440 .514
WBgroup 51 .80 20.32 54.85 31.26 53.16 9.97
Moderate modification
TLgroup 32.37 33.44 33.33 27.43 34.77 7.27 2.603 .122
WBgroup 16.97 20.73 18.79 23 .74 17.09 7.93
Substantial modification
TLgroup 7.19 12.62 7.69 16.12 7.92 3.37 .269 .609
WBgroup 8.12 12.22 5.61 9.78 5.34 3.66

TABLE 7
I-Test Results of Pre- and Post-practice Ratio of Paraphrase Types
TLgroup WB group
p t P
Exact copy 1.456 .176
Near copy 1.062 .309 -.114 .912
Minimal modification -.748 .469 -.362 .725
Moderate modification -.072 .944 -.178 .862
Substantial modification -.135 .895 .649 .531

The results of ANCOVA reveal no statistically significant differences between the two
practice types (see Table 6). In a comparison of the pre and post-practice ratios in each
group, moreover, no statistically significant differences were noted in all the paraphrase
types (see Table 7).
 Paraphrase Practices for Using Sources in L2 Academic Writing 71

4. Participants' Perception on Paraphrase Practice

The participants' perceptions on paraphrase practice were investigated through their pre-
and post-practice questionnaire survey to explore the fourth research question. All of the
participants demonstrated more confidence on their paraphrasing skills after the practices
(see Table 8). Interestingly, the independent and paired samples t-test results reveal that the
TL group had statistically higher confidence prior to the practices than the WE group, but
not after the practices (t = 2.294, P = .032); the WE group gained statistically more
confidence after the practices (t = -7.111, P = .000). A statistically higher post-practice
confidence of the WE group over the TL group was also noted in the ANCOVA results
(see Table 8). After the practices, all of the participants also illustrated more need of
learning paraphrasing skills as well as more positive perception on the usefulness of
paraphrasing practice. In the open-ended questions, the participants expressed the
usefulness of paraphrasing practice (3 TL and 3 WE students); learning paraphrase
strategies including word deletion (2 TL and 2 WE students); recognition of the
importance of paraphrasing to avoid plagiarism (1 TL student and 2 WB students);
improvement in vocabulary and grammar (2 TL students); and usefulness of class
contents for student needs (1 TL student). The t-test results between the pre- and post-
practice perception showed a statistically significant difference in the need of learning
paraphrasing in both groups (IL, t = -2.739, p = .018; WB, t = -4.224, P = .002).

TABLE 8
ANCOVA Results of Pre - and Post-practice Perception on Confidence, Need and Useful ness of
Paraphrase Practice, and the Role of Native Teachers and Using Internet by Practice Type
Pre-practice Post-practice Adjusted
F P
M SD M SD M SE
Confidence for paraphrasing
TLgroup 3.77 0.83 4.00 0.35 3.92 0.11 8.876 .007
WB group 2.95 0.91 4.32 0.46 4.42 0.12
Need to learn paraphrasing
TLgroup 4.35 0.52 4.73 0.44 4.70 0.10 1.723 .203
WB group 4.09 0.63 4.86 0.23 4.89 0.11
Usefu lness of paraphrase practice
TLgroup 4.38 0.65 4.62 0.51 4.64 0.16 .276 .605
WB group 4.55 0.52 4.55 0.69 4.52 0.17
Usefulness of learning from native teachers
TLgroup 3.92 0.73 4.23 0.83 4.23 0.21 .352 .560
WB group 3.86 1.27 4.05 0.6 1 4.05 0.23
Usefi.llness of leaming by using the lnternet
TLgroup 3.73 0.67 4.00 0.71 3.98 0. 18 3.048 .097
WB group 3.61 0.99 4.11 0.65 4.13 0.22
72 Choi, Yeon Hee

After the practices, the TL group perceived learning paraphrasing from a teacher more
positively, whereas the WE group viewed learning paraphrasing using the online sites
more positively (see Table 8). These results suggest that each group was satisfied with
group-specific (or practice-specific) practice methods. This is also noted in both groups'
positive responses about the practice techniques and methods, as shown in Table 9. In the
open-ended questions of the post-practice survey, the participants expressed effectiveness
of learning synonyms (4 TL and 3 WE students); whole-class paraphrasing (2 TL and 4
WE students); step-wise paraphrasing demonstration (2 TL and 2 WE students);
comparison of a model paraphrase with the original passage (1 TL student); and teacher
feedback on paraphrasing (3 TL and 3 WE students).

TABLE 9
t- Test Results of Post-practice Perception on Usefulness of Practice Techniques and
Methods by Practice Type
TLgroup WBgroup
p
M SO M SO
Usefulness of practice techniques and procedures 4.26 0.39 4.33 0.34 -.439 .665
Usefulness of practice-specific techniques 4.10 0.39 4.30 0.25 -1.534 . 139

Though no statistically significant difference was found in the results of t-test between
the groups, the WE group participants more positively evaluated the usefulness of their
group-specific in-class techniques and procedures, and practice-specific techniques. In the
post-practice open-ended survey, they expressed effectiveness of getting aware of the
online sites (3 students); and utilizing online sites and tools for paraphrasing (11 students),
especially for searching synonyms (e.g., www.thesaurus.com and free dictionary online
sites) and detecting plagiarism (see Figure 2). The TL group also expressed effectiveness
of the teacher 's demonstration on step-wise paraphrasing (4 students) and explanation on
synonyms, different nuances of synonyms, vocabulary, and deletable words (7 students).
The whole practices placed a focus on helping the participants develop paraphrasing
skills including how to use synonyms and transform sentence structures to write a compact
paraphrase, regardless of practice types. Thus, the participants' perception on effectiveness
of the practices in the aspects of vocabulary and grammar was surveyed after the practices
(see Table 10).

TABLE 10
t-Test Results of Post-practice Perception on Usefulness of Practice Techniques
on Grammar and Vocabulary by Practice Type
TL group WB group
p
M SO M SO
Usefulness of practice techniques on vocabulruy 4.15 0.80 4.73 0.47 -2.089 .048
Usefulness of practice methods on grammar 3.54 0.78 3.18 0.60 1.239 .229
 Paraphrase Practices for Usi ng Sources in L2 Academic Writing 73

Both practice groups viewed their practices were more effective in vocabulary than in
grammar. Nonetheless, the WB group participants statistically evaluated more higWy the
effectiveness of their group-specific practice in the aspect of vocabulary, though one WB
student pointed out that it was difficult to differentiate nuances of synonyms listed on the
Internet sites. In the aspect of grammar, no statistically significant difference was found
between the groups. However, the TL group evaluated its practice more positively than the
WB group.

V. CONCLUSION

The present study has developed two types of paraphrase practice methods (teacher-led
and Web-based) with a practice cycle, as an attempt to propose an interventional technique
to train L2 student writers to paraphrase academic source texts effectively without
committing plagiarism. Overall, the results of pre- and post-practice paraphrase awareness
and performance scores and questionnaire responses reveal the effectiveness of both short-
term intensive practice methods in acquiring the knowledge on the defmition and nature of
paraphrase and in developing paraphrasing skills, specifically how to cite source texts and
paraphrase them concisely, as found in Wette (2010).
Statistically, not many significant differences were found between the two practice types.
Nonetheless, the study reveals variations in their effects. The higher awareness and
performance scores of the TL group suggest more direct impact of explicit instruction
when the intervention is a short intensive program. Teacher-led practices are found to be
useful to enhance paraphrasing ability in linguistic reformulation, language use, and
citation, while Web-based practices are effective in content quality and conciseness. The
responses from the post-questionnaire also indicate different effects of the practice
techniques, though each practice group highlights the group-specific practice method
preferences and satisfaction with the overall practice cycle. Teacher-led practices appear to
be helpful to learn how to restructure sentences; Web-based practices are effective in
searching vocabulary and synonyms. These findings thus suggest that L2 academic writing
instruction should be designed differently to tailor the needs of L2 student writers
including sub-skills of paraphrasing.
The results of the study reveal the pedagogical values of the practice models; however,
the limitations of the study should be noticed for further research. Firstly, the fmdings from
the study with a small number of participants should be only suggestive. To generalize the
results of the study, inclusion of more participants seems to be preferable. Secondly, the
paraphrasing practice was provided intensively for a short period of time. No significant
effects of paraphrase practices on paraphrase types were noted in tenns of their frequency
74 Choi , Yeon Hee

and ratio. The participants' significant gains of declarative knowledge of paraphrase and
their positive reactions, especially those of the Web-based group, did not lead to greater
paraphrasing performance. If future studies explore more longitudinal effects of practices
(e.g., one semester), more significant effects of practice would be revealed, as Campbell
(1990) stated that it seems to be difficult for L2 student writers to acquire essential
paraphrasing skills and strategies within a short period of practice due to lack of L2
linguistic competence. Specifically, elements such as content quality and linguistic
reformulation or Web-based self-learning or learner-autonomous practices may require
long-term learning for more beneficial effects.
Reading comprehension ability of the participants was not considered crucially during
the practice sessions since the instructor was inattentive to the participants' content
comprehension. However, Wette (2010) emphasizes the importance of comprehending
source text in order to improve the content quality in paraphrasing. In this regard, the
question of whether the participants of the study fully grasped the main idea of the source
text and incorporated this understanding to their writings remains. Thus, further studies
need to consider carefully the role of reading ability when paraphrasing source texts.
Moreover, it is suggested to explore the influence of content schema or academic
background knowledge on paraphrasing by involving L2 student writers from one
discipline.
More effective ways of training paraphrasing in tenns of lexical and syntactic aspects
could also be considered in the future studies to draw on in-depth effects of practices. The
present study placed a focus on grammar and vocabulary since these elements had been
noted as problematic aspects in L2 student writers' paraphrasing (Currie, 1998; Howard,
1995; Johns & Mayes, 1990; Keck, 2006; Shi, 2004; Sung, 2011). Still, the participants
struggled with these matters throughout the practices. Thus, further research should
explore how to design paraphrasing interventions to overcome lexical and syntactic
obstacles. In addition, L2 proficiency levels could be one of the main concerns in
conducting future studies. There would be a possibility that differential effects could be
reported if more advanced or lower level students were trained through the paraphrasing
practice types suggested in the study.

REFERENCES

Barry, S. (2006). Can paraphrasing practice help students define plagiarism? College
Student Journal, 40, 377-385.
Campbell, C. (1990). Writing with others' words: Using background reading text in
academic compositions. In B. Kroll (Ed.), Second language writing: Research
 Paraphrase Practices for Us ing Sources in L2 Academic Writing 75

insights for the classroom (pp . 211 -230). Cambridge: Cambridge University Press.
Carson, J. G. (2001). A task analysis of reading and writing in academic contexts. In D.
Belcher & A. Hirvela (Eds.), Linking literacies: Perspectives on L2 reading-
writing connections (pp. 48-83). Ann Arbor, MI: The University of Michigan Press.
Chandrasoma, R., Thompson, C., & Pennycook, A. (2004). Beyond plagiarism:
Transgressive and nontransgressive intertextuality. Journal of Language, Identity,
and Education, 3, 171- 194.
Choi, S. K. (2009). Daehaksaeng geulsseuygi-ey nathanan oryubunseok: Inyoung-
bangshik-euy oryu-leuljungshim-euyro (Analysi s of en-ors in academic writing of
university students: Focusing on the style of citation). Saegukeo Gyoy uk (Journal
ofKorean Language Education), 81, 299-324.
Cunie, P. (1998). Staying out of trouble: Apparent plagiarism and academic survival.
Journal ofSecond Language Writing, 7, 1-18.
Deckert, G. (1993) . Perspectives on plagiarism from ESL students in Hong Kong. Journal
ofSecond Language Writing, 2, l31-148.
Howard, R . M . (1995) . Plagiarisms, authorships, and the academic death penalty.
College English, 5 7, 788-806 .
Hsu, A. Y-P. (2003). Patterns of plagiarism behavior in the ESL classroom and the
effectiveness of instruction in appropriate use of sources. Unpublished doctoral
dissertation, University of Illinois, Urbana-Champaign, IL.
Hu, G., & Lei, J. (2011). Investigating Chinese university students ' knowledge of and
attitudes toward plagiarism from an integrated perspective. Language Learning, 61,
1-38.
Hyland, F. (2001). Dealing with plagiarism when giving feedback . ELT Journal, 55, 375-
38l.
Hyland, T. A. (2009). Drawing a line in the sand: Identifying the borderzone between self
and other in ELl and EL2 citation practices. Assessing Writing, 14,62-74.
Johns, A. M. (1993). Reading and writing tasks in English for academic purposes classes :
Products, processes and resources. In J. G. Carson & 1. Leki (Eds.), Reading in the
composition classroom: Second language p erspectives (pp . 274-289). Boston, MA:
Heinle & Heinle.
Johns, A. M ., & Mayes, P. (1990) . An analysis of summary protocols of university ESL
students. Applied Linguistics, 11 , 25 3-27l.
Keck, C. (2006). The use of paraphrase in summary writing: A comparison of Ll and L2
writers . Journal ofSecond Language Writing, 15, 261-178.
Lee, 1. J. (2007). Yeongu yunri hwakrip-eul wiyhan inyoung-gwa phyojeol-euy ihae (The
understanding of citation and plagiarism for establishment of the research ethics).
Yunriyeongu (Journal ofEthical Studies), 66, 1-25.
76 Chai, Yean Hee

Moon, Y. 1. (2002). Korean university students' awareness of plagiarism in summary

writings. Language Research, 38(4), 1349-1365.
Oshima, A., & Hogue, A. (2006). Writing academic English (4th ed). White Plains, NY:
Pearson Longman.
Park, M, H., & Lee, H. (2010). Yeongeo nonsulsiheom-eyseo phyojeolbangjjigyoyuk-euy
hyogwa-wa phyojeolyuhyeong-ey daehan yeongu (Exploring the effect of
plagiarism-prevention training and the type of plagiarism in an L2 essay writing
test). Yeongeohak (Korean Journal of English Language and Linguistics), 10(4),
798-805.
Pecorari, D. (2003). Good and original: Plagiarism and patchwriting in academic
second-language writing. Journal of Second Language Writing, 12,317-345.
Pennycook, A. (1994). The complex contexts of plagiarism: A reply to Deckert. Journal of
Second Language Writing, 3,277-284.
Pennycook, A. (1996). Borrowing others' words: Text, ownership, memory, and
plagiarism. TESOL Quarterly, 30, 201 -230.
Prochaska, E., & Moon, y'-1. (2004). The relation of writing ability to instances of
"plagiarism" in Korean EFL university students' summary writings. Foreign
Languages Education, 11(1),1-26.
Qian, J., & Krugly-Smolsk, E. (2008). Chinese graduate students' experiences with
writing a literature review. TESL Canada Journal, 26, 68-86.
Qin, Y. (2008). Challenges in source use for Chinese graduate students in the United
States. Unpublished doctoral dissertation, Syracuse University, Syracuse, NY.
Rhee, E. H. (2010). Complexities and dynamics of Korean graduate students' textual
borrowing in academic writing. Unpublished doctoral dissertation, The Temple
University, Philadelphia, PA.
Rhee, H. Y. , & Nam, T. W. (2010). Daehaksaeng-deyl-euy inyong mit phyojeol-ey
gwanhan inshik yeongu (A study on undergraduates ' citation use and plagiarism
awareness). Hangukmunheonjeongbohakhwoyji (Journal of Korea Library &
Information Science), 44(3), 175-198.
Roig, M. (1997). Can undergraduate students determine whether text has been plagiarized?
The Psychological Record, 47, 113-122.
Roig, M. (2001). Plagiarism and paraphrasing criteria of college and university professors.
Ethics & Behavior, 11,307-323.
Shi,1. (2004). Textual borrowing in second-language writing. Written Communication, 21,
171-200.
Suh, S. 1. (2008). Plagiarism, textual borrowing or something else?: An L2 student 's
writing-from-source tasks. Unpublished doctoral dissertation, University of
Maryland, College Park, MD.
 Paraphrase Practices for Us ing Sources in L2 Academ ic W riting 77

Sung, H. E. (2011). Daehaksaeng yeongeo sseuygi-eyseo phyojeol yeybang-eul wuyhan

inyoung-gwa hwaneonhagi hyullyeon hyogwa bunseok (A study on the effects of
quotation and paraphrase practices for avoidance of plagiarism for Korean
university students). Unpublished MA thesis, Ewha Womans University, Seoul,
Korea.
Sutherland-Smith, W., & Carr, R. (2005). Turnitin.com: Teachers' perspectives of anti-
plagiarism software in raising issues of educational integrity. Journal of University
Teaching & Learning Practice, 2, 95 -1 01.
Swales,1. M., & Feak, C. B. (2004). Academic writingfor graduate students: A course for
nonnative speakers of English (2nd ed.). Ann Arbor, Ml: The University of
Michigan Press.
Thomas, Z. (2010). Addressing pedagogy on textual borrowing: Focus on instructional
resources. Writing & Pedagogy, 2,223 -250.
Weigle, S. C. (2002). Assessing writing. Cambridge: Cambridge University Press.
Wette, R. (2010). Evaluating student learning in a university-level EAP unit on writing
using sources. Journal of Second Language Writing, 19, 158-1 77.
Wheeler, G. (2009). Plagiatism in the Japanese universities: Truly a cultural matter?
Journal ofSecond Language Writing, 18, 17-29.
Wright, P. L., & Taylor, D. S. (1994). Improving leadership peljormance: Interpersonal
skills fo r effective leadership. Englewood Cliffs, NJ: Prentice Hall.

APPENDIX A
Scoring Rubrics of Paraphrase
Content
Ling uistic Reformulation Language Use C itation
Quality Conciseness

5 Paraphrase includes all 5 Paraplu'ase is IeParaphrase d isplays writer's use 5 !Native-like flu ency in 2 Form of
key points of orig inal well written of own sentence structures or lexical usage and citation
passage and displays the WiUl key expressions, which have not language usage, such and
writer's clear points in been used in orig inal passage. as correct use of reference
understanding of content concise relative clauses, is
of original passage. manner. prepositions, modals, accurate.
aIticles, verb forms,
and tense sequencing,
and no fi'agments or
run-on sentences.

4 PaI'aphrase includes 4 Paraplu'ase 8 PaI'aphrase displays substantial 4 Advanced proficiency I Fonn of

almost all key po ints of occasionally use of writer's own sentence in language usage, citation
original passage aIld includes structures or expressions, though with some and
disp lays writer's fairly irrelevant which have not been used in minor graI11Jnatical reference
78 Choi, Yeon Hee

good understanding of specifics, original passage, though and lexical errors. is

content of original which paraphrase may contain some partially
passage, though some demands similar phrase-level accurate.
points may not be more expressions from original
accurately addressed. conciseness. passage.

3 Paraphrase includes half 3 Paraphrase 6 Paraphrase displays partial use 3 Occasional noticeable
the key points of original includes of writer's own sentence grammatical and
passage and displays the ilTelevant stmctures or expressions, lexical errors and
writer's partial specifics at wh ich have not been used in limited range of
understanding of content times that original passage, though vocabulary usage.
of original passage. lack paraphrase may occasionally
conciseness. contain similar phrase-leve l
expressions from origina l
passage.

2 Paraphrase includes less 2 Paraphrase 4 Paraphrase displays lack of 2 Accumu lation of

than half the key points includes a writer's own sentence structures grammatical and
of original passage and number of or expressions that have not lex ical errors with
displays the writer's in'elevant been used in original passage, negative effect on
limited understanding of specifics and paraphrase contains a comprehension of
content of original with serious noticeable number of similar content.
passage. prob lems in phrase-level expressions from
conciseness. original passage.

I Paraphrase does not I Paraphrase 2 Paraphrase replicates almost all I Serious and frequent
include any key points of includes sentences fi'om original grammatical and
original passage and substantial passage. lexical errors that
displays no sign of the number of interfere with
writer's understanding of irrelevant comprehension of
content of original specifics content.
passage. with no sign
of
conciseness.

o Paraphrase merely copies o Paraphrase oParaphrase merely copies o Paraphrase merely o/No
words from original merely words fi'om original passage. copies words ii-om citations
passage. copies words original passage. and no
from original references.
passage.

Applicable levels: tertiary

Key words: paraphrase/paraphrasing, academ ic writing, L2 writing, citation, plagiarism, teacher-led, Web-
based, Internet, using source texts
 Pa raphrase Practices for Using Sources in L2 Acade mi c Writing 79

Yeon Hee Choi

Dept. of English Education
Ewha Womans Uni versity
52, Ewhayeodae-gil , Seodaemun-gu
Seoul , 120-750, Korea
Tel: (02) 3277-265 5
Email : [email protected]. kr

Received in March, 201 2

Rev iewed in April, 201 2
Rev ised version received in May, 201 2