Preventing Unintentional Plagiarism: A Method for Strengthening Paraphrasing Skills
Walker, Angela L., Journal of Instructional Psychology
Plagiarism may result from faulty cognitive processing and thereby be unintentional (Marsh, Landau, & Hicks, 1997). The current study tested the effectiveness of paraphrasing training designed to prevent unintentional plagiarism. Thirty-six students enrolled in research methods participated, one group received paraphrasing training; a control group did not. Both groups paraphrased a moderate and a difficult one-paragraph passage midway through the semester and a lengthier passage at the end. Results indicated no difference for the moderate paragraph; however, the training group performed better than the control on the difficult and lengthier passages. Interestingly, both groups reported similar levels of confidence in paraphrasing skills. Findings demonstrate the potential for paraphrasing training and provide direction for further development.
In an early lecture on plagiarism, students appeared puzzled when reviewing examples of paraphrased and plagiarized work. Afterwards, one student stood and said, "Hi, my Name is 'X' and bravely admitted, "I didn't know it, but I've been plagiarizing my entire academic life"--inciting an animated conversation wherein students discussed their genuine confusion. Students indicated that identifying plagiarism was simple when a writer failed to provide citation; left the original unchanged, or modified the work by one or two words. When a writer made several minor changes, however, students believed the paraphrasing was accurate suggesting that when writing students may unintentionally plagiarize.
Given the growth of writing services and highly publicized cases of plagiarism, conceiving of students as plagiarizing unintentionally may border on the preposterous; however, research suggests otherwise (Roig, 1997, 1999, 2001). Students, for example, evaluated rewritten versions of an original paragraph and indicated whether the versions created by the researchers represented accurate paraphrasing or plagiarism (Roig, 1997). The plagiarized versions contained moderate and superficial alterations--substitutions of synonyms for original words, additions, and/ or deletions of one to four words, and reversal s of the sentence structure--close inspection of the modified versions revealed that the originals remained primarily intact with a few "patches" which Howard (1995) aptly defined as patchwriting. In the study, most participants correctly labeled the paraphrased items, but nearly half judged plagiarized versions as accurately paraphrased showing that students are unaware of the extent that they must change the material further indicating that students can plagiarize unintentionally (Roig).
In a separate study, Roig (1999) explored the possibility that unintentional plagiarism was related to the misunderstanding the definition of plagiarism as well as to readability. To test this possibility, students paraphrased both a simple and a complex paragraph. Results showed that students plagiarized more when tackling a complex paragraph than when facing a simple paragraph; providing data that students do in fact possess skills necessary for paraphrasing but students may be impeded from applying those skills when dealing with rigorous text (Roig).
Surprisingly, students are not the only writers guilty of committing unintentional plagiarism. Roig (2001) proposed that students struggle with paraphrasing because of modeling their professors' practices. Given the exact tasks as students in the Roig (1997) study, 44% of professors judged a plagiarized item as correct and one third lifted five-to-nine word strings. Roig argued that the errors result from a lack of agreement about the definition of plagiarism within the field.
Using the dual model of persuasion (see Chaiken, Wood, & Eagly, 1989), Marsh, Landau, and Hicks (1997) suggested that inadvertent plagiarism results from faulty cognitive processing. First, simply thinking about paraphrasing requires considerable cognitive energy and once the physical process of writing begins, people have little resources left to automatically engage in thoughtful, systematic processing to determine if they paraphrased sufficiently; as such people experience cryptoamnesia or are unaware of plagiarizing (Marsh, Landau, & Hicks). Furthermore, as paraphrasing is cognitively demanding, students are likely to engage in less effortful heuristic processing increasing the likelihood of using short cuts--patchwriting--particularly as the material becomes more complex (Marsh, Landau, & Hicks; Roig 1999). These cognitive processes, coupled with the fuzziness of the definition of paraphrasing--within and across disciplines-set the stage for unintentional plagiarism (Roig, 2001).
Familiar with the complicated nature of plagiarism, researchers designed techniques to reduce plagiarism--unintentional and deliberate--in student writing (Landau, Druen, & Arcuri, 2002; Barry, 2006). Providing students with concrete examples of plagiarized work along with detailed explanations, for instance, increased student knowledge of plagiarism (Landau et al.) Additionally, students in introductory psychology courses who practiced paraphrasing one paragraph segments over a course of six weeks developed more sophisticated and complete definitions of plagiarism--failing to provide citation and misrepresenting another's idea as one's own--than students in a control group (Barry, 2006).
Following in the tradition of researchers, I designed and tested the effectiveness of paraphrasing training by providing students enrolled in Research Methods the opportunity to discuss examples and to practice paraphrasing on multiple occasions (Landau, et al. 2002; Barry, 2006). I further compared students who received training to a control at two times. Based on research about complexity (Roig, 1999), I predicted that students who received training would plagiarize less than a control given a difficult paragraph, but not a moderate paragraph during assessment one. I also compared the groups using a lengthier passage in assessment two, as research regarding reading difficulty has been limited to the evaluation of a single paragraph (Roig). In fact, Roig suggested that plagiarism might decrease if students received additional information. In contrast, I proposed that length operates similarly to complexity; both may lead to heuristic processing resulting in an increase in instances unintentional plagiarism--unless the students are encouraged to avoid plagiarizing (Marsh, Landau, & Hicks, 1997). Therefore, I predicted that training would provide the encouragement needed to paraphrase the lengthier passage more accurately--or commit fewer errors--than the control. Lastly, given that people exhibit an overconfidence bias particularly when viewing themselves as accurate (Klayman, Soil, Gonzalez-Vallejo, 1999), I predicted similar confidence rating for both groups perhaps providing evidence that students genuinely misunderstand the definitions of plagiarism and paraphrasing and may commit plagiarism unintentionally.
Thirty-six students--33 women, 3 men--enrolled in two sections of Methods in Psychology I taught by two instructors at a New England university participated. Nineteen students were in the training group; 17 students who did not receive training, served as the control. All students in the training condition participated in both assessments. In the control, 16 participated in the first assessment and 17 in the second, reflecting class attendance rates.
Materials and Procedure Paraphrasing Training
Paraphrasing training began during the second week of the term during one class session with a brief review of the University's Academic Integrity Policy and a discussion of the rules for proper citation. Next, I introduced of definitions based on the work of Roig (1997 & 1999) and Howard (1995) of plagiarism or patchwriting, which included: l) word strings, lifting exact phrases from the original consisting of five-to-nine words or more 2) substitutions, replacing original words with synonyms 3) additions, introducing one to four words to the original 4) deletions, eliminating one to four words from the original and 5) reversals, reversing the sentence structure and/or word structure. I further defined accurate paraphrasing as expressing the original idea in one's own words and producing a summary of the original devoid of word strings, substitutions, additions, deletions, and reversals. During the same session, I demystified the research process by describing the time psychologists devote to reading published research and writing research papers as well as explained the steps involved in peer-review. I encouraged students to develop and appreciate their own written voices. Following, the class reviewed samples of original excerpts and corresponding examples of plagiarized and accurately paraphrased work.
Next, the students prepared paraphrasing notes formatted into two columns. In the left column(s), students wrote the APA style reference on page one, followed by direct quotes of the critical information about each section of an empirical article. In the fight column(s), students wrote paraphrased summaries, allowing for a direct comparison of their work and the original, a technique recommended in methods textbooks (e.g. Harris, 2001). To facilitate focused note taking, I created guidelines using the work of Jordan and Zanna (1999). As suggested by Roig (2001), I emphasized summarizing the material and explained that should readers desire more detail, readers must return to the original--hence the necessity for the reference list. After the lecture, students independently constructed paraphrasing notes for the same article and critiqued one another's work in class.
Throughout the semester, students compiled paraphrasing notes to accompany their research papers. For the first report, students submitted notes for five articles and summarized three in their papers. In the second (research proposal), students submitted notes for an additional five articles (10 total) and summarized eight. For the full report, students submitted paraphrasing notes for an additional five articles (15 total) and summarized 12. Although the paraphrasing notes were not graded, students could not submit papers without notes. I also informed students that if I found gross plagiarism, they would receive a failing grade for the papers.
As is standard with the Methods course offered at the university, the instructor for the control group presented a brief lecture on plagiarism and paraphrasing at the beginning of the semester. Similar to the training group, students in the control completed original research projects and wrote APA style empirical papers with similar requirements for the number of articles included in the literature review. Students in the control, however, neither participated in an interactive session on plagiarism nor generated paraphrasing notes.
I developed the measures used for plagiarism in the current study based on previous research (Roig, 1997 & 1999) consisting of the following averaged measures: 1) word strings, that is lifting exact phrases consisting of five-to-nine words from the original 2) substitutions, modifying the original text by using one to two synonyms; 3) additions, including one-to- two new words to the original 4) deletions, eliminating one-to-four words from the original; and, 5) reversals, rearranging sentence order or interchanging phrases. I created a plagiarism score using by adding all five items (see Appendix for example).
A brief questionnaire regarding experiences with paraphrasing included items rated on a 7-point Liken-type scale ranging from 1 (not at all) to 7(very much). For both groups, two items assessed confidence: confidence in paraphrasing ability and confidence in the ability to assist others in avoiding plagiarism. For the training group, the questionnaire included additional items measuring the "helpfulness" and "usefulness" of the paraphrasing training.
The first assessment occurred during the middle of the semester. Similar to previous research (Roig, 1997 & 1999) students were a moderate and a difficult paragraph--each consisting of four sentences. I selected the moderate paragraph from an article about procrastination (Chu & Choi, 2005) and the difficult paragraph--containing more psychological terms--from an article regarding actor-observer difference (Moore, 2005). Independent raters agreed upon the readability of both paragraphs. The students then paraphrased the paragraphs, taking as much time as needed to complete the task. The second assessment occurred at the close of the semester wherein students paraphrased a lengthier excerpt--a three-paragraph segment--concerning research on gay and lesbian couples (Kurdek, 2005). Independent raters judged the lengthier passage as equivalent in readability to the moderate paragraph yet not as complex as the difficult paragraph. As in the first assessment, students took as much time as needed to complete the task. Afterwards, students completed the brief questionnaire about paraphrasing and the instructors thanked the students for participating.
Two independent raters coded the passages and established the interrater reliability for the moderate passage as .90, the difficult passage .89 and the lengthier passage .93. I conducted an independent t-test to determine differences on the dependent measures for the training group and the control. As predicted, there were no significant differences on the measures for the moderate passage, ts < 1. For the difficult passage, in contrast, the training group produced significantly less word strings, substitutions, additions, deletions, and reversals than the control (see Table 1). The overall plagiarism score was reliable ([alpha] = .70) as shown in Table 1; the training group scored significantly lower than the control.
Similarly, for the lengthier passage, the training group produced significantly fewer substitutions, additions, deletions, and reversals than the control (see Table 2). Results regarding word strings did not differ (t < 1). For the overall plagiarism score, which was reliable ([alpha] = .75), the training group had a significantly lower score than the control as displayed in Table 2.
During content analysis, I discovered an unexpected relationship between the amount of writing and plagiarism for the lengthier passage that did not hold for the shorter passages, p > .05. More specifically, the training group used significantly less words than the control (see Table 2) and analysis revealed a significant positive correlation between the number of words and the overall plagiarism score, r = .57, p = .01.
As predicted, confidence ratings between the groups did not differ significantly ts < 1. Lastly, students who received training reported that the training was useful (M = 5.30, SD = 1.26) and helpful (M = 5.35, SD = 1.30).
The present study evaluated paraphrasing training designed to decrease unintentional plagiarism and results partially provide evidence of its effectiveness. Consistent with previous research (Roig, 1999), students did not differ on the moderate paragraph indicating that generally, methods students can paraphrase brief passages that contain little psychological terminology. Particularly for more demanding passages, results suggest that the paraphrasing training facilitated students' performance--as demonstrated by differential performance on the difficult passage at mid-semester and lengthier passage at semester's close. Students who received training, furthermore, performed better than the control suggesting that length and difficulty function in similar ways cognitively (Chaiken, Wood, & Eagly, 1989; Marsh, Landau, & Hicks, 1997). In other words, when presented with difficult or lengthy passages, students typically engaged in heuristic processing--unless trained to avoid doing so. Therefore, paraphrasing training, which encourages using one's own written voice, summarizing ,and continual comparison, may increase the use of thoughtful systematic processing thereby resulting in fewer instances of plagiarism.
There were some unexpected findings. First, the low frequency of word strings is inconsistent with previous research (Roig, 1997, 1999) indicating that students are indeed modifying the original--albeit patchwriting--perhaps explaining the similarity of confidence scores. The confidence ratings also indicate that the definition of paraphrasing is unclear to students enrolled in methods courses thereby providing evidence that plagiarism can be unintentional further demonstrating the need for more extensive paraphrasing training than typically occurs in upper level courses (Landau, et al. 2002; Barry, 2006). Second, the unpredicted positive correlation between the number of words and the amount of plagiarism for the lengthier passage suggests that students who paraphrase accurately focus on a global approach of summarizing rather than attempting to capture every detail. This result supports previous research suggesting that when teaching about writing, instructors should emphasize condensing and summarizing (Roig, 2001).
Despite these interesting findings, there were limitations. First, the study was posttest only and did not include pre-training data--failing to meet the requirements of the classic non-equivalent groups design (Campbell & Cooke, 1979). However, the consistent differences found at both assessments indicate that including a pre-test would lead to similar findings. Second, students did not paraphrase the same passages at both assessments, in order to prevent testing effects. Given the length of time between assessments, however, this may have improved the study. Third, I wanted to provide students with as much time as they needed to complete the paraphrasing so that a time limit would not distract from their efforts, however, I did not measure the time spent paraphrasing--possibly an important variable in determining the effectiveness of the training and the accuracy of paraphrasing.
Although somewhat effective, the paraphrasing training could benefit from modification given that both groups plagiarized to some extent. One revision may include further heightening students' concerns about the accuracy of their paraphrasing skills--easily accomplished by reviewing the present study. This may seem counterintuitive; however, researchers found that students who felt like imposters--or those who are continually concerned about the legitimacy as students--reported cheating and plagiarizing less often in their academic lives than non-imposters (Ferrari, 2005). Therefore, encouraging vigilance by building in activities that require students to review and assess formally their work may further contribute to the effectiveness of the paraphrasing training.
Anecdotally, trained students were better equipped to discuss research than previous classes perhaps showing that the training may not only serve to help students paraphrase accurately but mat also function to increase their engagement in the research process. Future research could empirically examine this possibility. In conclusion, the current study demonstrates the necessity of integrating paraphrasing training in systematic ways into research methods and writing intensive psychology courses and provides a possible vehicle for preventing unintentional plagiarism.
"When (3) explaining (4) the behavior of others, people tend to discount the influence of the social situation 1 more than they should (2) (Jones & Nisbett, 1972). For instance (5), Nisbett, Caputo, Legant, and Maracek (1973) found (6) that (7) participants in their study explained their own selection of a (8) college major and a (9) romantic partner in terms of (10) the features of those majors and partners (11). However (12), when explaining their (13) best friends' choices (14), (15) is people instead (16) relied on (17) the features of their friends' personality traits (18). This (19) actor-observer difference describes the (20) tendency for (21) people to disregard the power of 22 situational influences on others' (23) behavior (Moore, 2005)."
Participant data with several instances of plagiarism from the control:
Social situations are discounted (1) more frequently (2) when individuals are asked (3) to explain (4) the behavior of others. Research done by (5) Nisbett, Caputo, Legant, and Maracek (1973) has indicated (6) that both the (7) selection of their (8) college major and the (9) romantic partner was based on (10) the major's and partner's features (11). This was different (12) when (13) the choice of a best friend (14) was involved (15). Instead, people (16) explained their (17) best friend's personality traits and not features (18) when asked to describe their behavior. The 19 actor-observer (20) tendency causes (21) people to disregard (22) situational influence on other people's (23) behavior.
(1.) Reversal of "discount influence of social situation" to the beginning of the sentence; deletion of "people tend to"; reversal of "discount the influence of social situation" to "social situations are discounted"; deletion of "the influence of"
(2.) Substitution of "more frequently" for "more than they should"
(3.) Addition of "individuals are asked"
(4.) Substitution of "to explain" for "explaining"
(5.) Deletion of "for instance", addition of "research done by"
(6.) Substitution of "has indicated" for "found"
(7.) Addition of "both the",
(8.) Substitution of "their" for "a"
(9.) Substitution of "the" for "a"
(10.) Substitution of "was based on" for "in terms of"
(11.) Reversal of "features of those majors and partners" to "the major's and partner's"
(12.) Substitution of "this was different" for "however"
(13.) Deletion of "when explaining their", addition of "the"
(14.) Reversal of "best friends' choices" to "choice of best friend"
(15.) Addition of "was involved"
(16.) Reversal of "people instead" to "Instead, people"
(17.) Substitution of "relied on" for "explained their"
(18.) Reversal of "personality traits" and "features" phrases, addition of "best"
(19.) Substitution of "the" for "this"
(20.) Deletion of "difference describes the"
(21.) Substitution of "causes" for "for"
(22.) Deletion of "the power of"
(23.) Substitution of "other people's" for "others"
"When explaining the behavior of others, people tend to discount the influence of social situation more than they should (Jones & Nisbett, 1972). For instance, Nisbett, Caputo, Legant, and Maracek (1973) found that participants in their study explained their own selection of a college major and a romantic partner in terms of the features of those majors and partners. However, when explaining their best friends' choices, people instead relied on the features of their friends' personality traits. This actor-observer difference describes the tendency for people to disregard the power of (1) situational (2,3) influences (4) on others' behavior" (Moore, 2005).
Participant data with few instances of plagiarism from training group:
When asked why did you choose him as your boyfriend many will answer with responses associated with individual qualities of "him." When asked why did you choose the major you are in, many will answer with responses associated with individual qualities about their major. Research continually shows that people often associate their own responses to individual qualities and disregard how their environment (1) and situations (2) they are in greatly (3) influence (4) the decisions they make.
(1.) Deletion of "the power of", addition of "how their environment"
(2.) Substitution of "situations" for "situational"
(3.) Addition of "they are in greatly"
(4.) Substitution of "influences" for "influence"
Barry, E. S. (2006). Can paraphrasing practice help students define plagiarism? College Student Journal, 20, 377-384.
Chaiken, S., Wood, W., & Eagly, A.H. (1989). Principles of persuasion. In E.T.
Higgins and A. Kruglanski (Eds.), Social psychology: Handbook of basic mechanisms and processes. New York: Guilford Press
Chu, A. H. C., & Choi, J. N. (2005). Rethinking procrastination: Positive effects of 'active' procrastination behavior on attitudes and performance. Journal of Social Psychology, 145, 245-264.
Campbell, T.D.,& Cook, D.T. (1979). Quasi-Experimentation: Design and Analysis for Field Settings. Chicago: Rand, McNally.
Ferrari, J. R. (2005). Imposter tendencies and academic dishonesty: Do they cheat their way to success? Social Behavior and Personality, 33, 11-18.
Harris, R. A. (2001). The plagiarism handbook. Los Angeles: Pyrczak Publishing.
Howard, R. M. (1995). Plagiarism, authorships, and the academic death penalty. College English, 57, 788-806.
Jordan, C. H., & Zanna, M. P. (1999). How to read a journal article in social psychology. In R. F. Baumeister (Ed.) The Self in Social Psychology (pp. 461-470). Philadelphia: Psychology Press.
Klayman, J., Soll, J. B., Gonzalez-Vallejo, C. (1999). Overconfidence: It depends on how, what, and whom you ask. Organizational Behavior and Human Decision Processes, 79, 216-247.
Kurdek, L.A. (2005). What do we know about gay and lesbian couples? Current Directions in Psychological Science, 14, 251-254.
Landau, J. D., Druen, P. B., & Arcuri, J. A. (2002). Methods for helping students avoid plagiarism. Teaching of Psychology, 29, 112-115.
Marsh, R. L., Landau, J. D., & Hicks, J. L. (1997). Contributions of inadequate source monitoring to unconscious plagiarism during idea generation. Journal of Experimental Psychology, 23, 889-897.
Moore, D. A. (2005). Myopic biases in strategic social prediction: Why deadlines put everyone under more pressure than everyone else. Personality and Social Psychology Bulletin, 31,668-679.
Roig, M. (1997). Can undergraduate students determine whether the text has been plagiarized? The Psychological Record, 47, 113-123.
Roig, M. (1999). When college students' attempts at paraphrasing become instances of potential plagiarism. Psychological Reports, 84, 973-982.
Roig, M. (2001). Plagiarism and paraphrasing criteria of college and university professors. Ethics and Behavior, 11, 307-323.
Angela L. Walker, Department of Psychology, Quinnipiac University.
Correspondence concerning this article should be addressed to Angela L. Walker at angela. firstname.lastname@example.org.
(1.) An earlier version of this article was presented at the 20th Annual Conference on Undergraduate Teaching of Psychology, Monticello, NY, March 2006.
(2.) I especially thank Joshua Taylor for his significant contribution to the content analysis.
Table 1 Mean Scores for Plagiarism Measures by Group for the First Exercise for the Difficult Passage in Assessment One Measure Training Group Control M SD M SD Word Strings .01 .22 .69 * 1.13 Substitutions 2.26 1.88 6.00 * 3.42 Deletions .31 .58 1.37 ** 1.08 * p-values < . 05. ** p-values < .01. Table 2 Mean Scores for Plagiarism Measures by Group for the Second Paraphrasing Assessment Measure Training Group Control M SD M SD Word Strings .21 .91 0.80 1.14 Substitutions 4.78 4.26 9.73 * 3.95 Deletions .84 1.60 3.26 ** 1.43 Additions .89 1.24 3.46 ** 2.26 Reversals .47 .90 1.80 * 1.26 Average Plagiarism Score 1.75 7.49 4.56 ** 6.52 Number of Words 91.00 21.22 120.80 * 17.62 * p-values < .05. ** p-values < .01.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Preventing Unintentional Plagiarism: A Method for Strengthening Paraphrasing Skills. Contributors: Walker, Angela L. - Author. Journal title: Journal of Instructional Psychology. Volume: 35. Issue: 4 Publication date: December 2008. Page number: 387+. © 2009 George Uhlig Publisher. COPYRIGHT 2008 Gale Group.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.