plagiarism

16
The Influence of the Internet on Plagiarism Among Doctoral Dissertations: An Empirical Study David C. Ison 1 Published online: 21 April 2015 # Springer Science+Business Media Dordrecht 2015 Abstract Plagiarism has been a long standing concern within higher education. Yet with the rapid rise in the use and availability of the Internet, both the research literature and media have raised the notion that the online environment is accelerating the decline in academic ethics. The majority of research that has been conducted to investigate such claims have involved self-report data from students. This study sought to collect empirical data to investigate the potential influence the prevalence of the Internet has had on significant higher education artifacts by comparing dissertations written prior to widespread use of the Internet with those written in a period in ubiquitous Internet use. Due to the prestige associated with the doctoral degree and the fact that the majority of the effort necessary to achieve such a degree resides within the dissertation, this study utilized Doctor of Philosophy (PhD) dissertations written in English and published by accredited universities in the U.S. and Canada. A sample of 384 dissertations were analyzed by Turnitin plagiarism detection software. The mean similarity indices for pre- Internet and post-Internet eras were 14.5 and 12.3 %, respectively. A Mann Whitney U test (Mdn=13, U=30,098.5, p <0.001) indicated that the differences between groups was significant, however opposite than has been purported within the exigent literature. When comparing the counts of dissertations for each time era considering those with plagiarism versus those that had little/no evidence thereof, there was no statistically significant difference (χ2 [1, N =368]=2.61, p =0.11). The findings of this study suggest that the Internet may not be significantly impacting the prevalence of plagia- rism in advanced levels of higher education. Keywords Plagiarism . Academic Integrity . Doctoral . Doctorate . Dissertation . Higher education . Turnitin . Similarity index . Empirical . Ethics J Acad Ethics (2015) 13:151166 DOI 10.1007/s10805-015-9233-7 * David C. Ison [email protected] 1 College of Aeronautics, Embry-Riddle Aeronautical University Worldwide, Daytona Beach, FL, USA

Upload: m-faisal-riyad

Post on 09-Dec-2015

3 views

Category:

Documents


1 download

DESCRIPTION

Recent Study on Plagiarism

TRANSCRIPT

The Influence of the Internet on PlagiarismAmong Doctoral Dissertations: An Empirical Study

David C. Ison1

Published online: 21 April 2015# Springer Science+Business Media Dordrecht 2015

Abstract Plagiarism has been a long standing concern within higher education. Yetwith the rapid rise in the use and availability of the Internet, both the research literatureand media have raised the notion that the online environment is accelerating the declinein academic ethics. The majority of research that has been conducted to investigatesuch claims have involved self-report data from students. This study sought to collectempirical data to investigate the potential influence the prevalence of the Internet hashad on significant higher education artifacts by comparing dissertations written prior towidespread use of the Internet with those written in a period in ubiquitous Internet use.Due to the prestige associated with the doctoral degree and the fact that the majority ofthe effort necessary to achieve such a degree resides within the dissertation, this studyutilized Doctor of Philosophy (PhD) dissertations written in English and published byaccredited universities in the U.S. and Canada. A sample of 384 dissertations wereanalyzed by Turnitin plagiarism detection software. The mean similarity indices for pre-Internet and post-Internet eras were 14.5 and 12.3 %, respectively. A Mann Whitney Utest (Mdn=13, U=30,098.5, p<0.001) indicated that the differences between groupswas significant, however opposite than has been purported within the exigent literature.When comparing the counts of dissertations for each time era considering those withplagiarism versus those that had little/no evidence thereof, there was no statisticallysignificant difference (χ2 [1, N=368]=2.61, p=0.11). The findings of this studysuggest that the Internet may not be significantly impacting the prevalence of plagia-rism in advanced levels of higher education.

Keywords Plagiarism . Academic Integrity . Doctoral . Doctorate . Dissertation . Highereducation . Turnitin . Similarity index . Empirical . Ethics

J Acad Ethics (2015) 13:151–166DOI 10.1007/s10805-015-9233-7

* David C. [email protected]

1 College of Aeronautics, Embry-Riddle Aeronautical University – Worldwide, Daytona Beach, FL,USA

Plagiarism has long been identified as an issue of concern in the academic environment as wellas among others utilizing text and other media (Wang 2008). The word Bplagiarism^ made itsway into the English language in the early 17th Century presumably to describe a steadilyincreasing phenomenon (BPlagiarism^ 2014a). The amplified interest in plagiarism was likelylinked to the fact that words were becoming much easier to copy, especially in light of therapid growth of printing presses, improved literacy, and more widespread distribution of text(Febvre and Martin 1976). As a result, the concept of copyright was developed among Britishprinters in the 18th Century in order to prevent theft of works (Howard 2007; MacQueen et al.2007). Even celebrated authors, politicians, and leaders are not immune from the temptationsof plagiarism. According to Moss (2005), Shakespeare and Martin Luther King stole parts oftheir works from others. Author Rand Paul and prominent primate scientist Jane Goodall arealso guilty of textual piracy (Bailey 2014). An intelligence report on Iraq concerning weaponsof mass destruction that was plagiarized from various, including online, materials wassubmitted to Prime Minister Tony Blair (Howard 2007). Another highly publicized case ofplagiarism was that of Jayson Blair of the New York Times in which he was found to haveborrowed text for many of his 600 articles (Roig-Franzia 2014).

Higher education also has experienced a prevalence of problems with plagiarism through-out its history (Postle 2009). Cheating scandals involving plagiarism have become morecommon in the headlines both in the U.S. and abroad. According to Perez-Pena (2013),approximately 70 students were implicated in a plagiarism event involving a final exam atHarvard University. A committee at Ohio University uncovered more than 40 master’s thesesthat had been plagiarized, in some cases rather overtly (Leingang 2006). Doctoral degrees havealso been revoked in cases where ex-students were found guilty of stealing words from otherssuch as cases at Ohio State University, Kent State University, Johns Hopkins University, andthe University of Florida (Connell and Mallette 2005; Pyle 2010). Several prominent Germangovernment officials including the Defense Minister and the Education Minister, were found tohave plagiarized their dissertations all of whom subsequently had their doctorate degrees revoked(Diehl and Trenkamp 2013; Gopalakrishnan 2011). In some locations, plagiarism is reported tobe not only an education issue but a societal one, for example in Nigeria, there are rampantaccounts of copying by students and government officials alike (Daramola 2014). China is also acountry facing challenges in curbing plagiarism, with numerous examples of academic trans-gressions and an apparent cultural disposition for acceptance of such misconduct (Rogers 2014).Plagiarism appears to be a multifaceted issue across the globe (Eret and Ok 2014).

In addition to the attention plagiarism has received in the media, it has also been the focusof numerous studies within the research literature. A common thread within exigent researchnotes that plagiarism is a significant problem at all levels of education including withincolleges and universities. Studies have indicated various levels of purported transgressions,with plagiarism being present anywhere from 10 to 95 % of the student population (Scanlon2003). Research has also claimed that the growing use of the Internet has assisted in theproliferation of plagiarism throughout all levels of higher education (Ison 2012; Eret and Ok2014; Li 2013; Postle 2009; Walker 2010). Howard (2007) summarizes the sentiment withinthe literature stating, Ba sense of impending doom hangs over the academy as the specter of‘Internet plagiarism’ threatens to undo the entire education enterprise. [… there is a] near-universal belief that the Internet is causing an increase in plagiarism^ (pp. 3–4). Wang (2008)also supported this notion describing that Bit is speculated that the Internet has aggravatedstudent plagiarism^ (p. 734). Yet as identified by Batane (2010), Bretag and Mahmud (2009),Ison (2012, 2014), and Walker (2010), the majority of focus of previous research has been on

152 D.C. Ison

self-report data from students and faculty. What current research is lacking is specific,empirical evidence to support suppositions within the research. There is virtually no direct,data-supported research on the alleged growth in plagiarism because of increased Internet use(Batane 2010; Ison 2014; Kutz et al. 2011; Ma et al. 2006; Ma et al. 2008; Walker 2010). Inparticular, a shortage of empirical studies exists in analyzing the potential effects of the Interneton the highest form of higher education, doctoral studies (Ison 2012, 2014). This study soughtto close this gap in the literature.

Review of Literature

Defining Plagiarism

According to the Cambridge Dictionary, the definition of plagiarism is Bto use another person’sidea or a part of that person’s work and pretend it is your own^ (BPlagiarism^ 2014b).Specifically addressing plagiarism in the context of research, Title 45 of the Code of FederalRegulations (CFR) Part 689.1 defines plagiarism as Bthe appropriation of another person’sideas, processes, results or words without giving appropriate credit^ (Research Misconduct –Definitions, 2012, para. 4). The combined elements of these definitions, that words, work, orideas used without proper credit or in a means to indicate that they are one’s own, convey thecommon essence of plagiarism as it is recognized in academe (Ison 2012, 2014; Wang 2008).Gibelman and Gelman (2003) remarked that plagiarism is often associated with some sort ofintentional deception; however, Gilmore et al. (2010) argued that it is common that individualsunintentionally omit citations or are not aware of proper attribution techniques. Therefore,there is the possibility that there are cases of unintentional in addition to intentional plagiarism.Even though there are cases when an individual does not seek to deceive, such occurrences aredifficult to identify or prove. Therefore, as noted by Walker (2010), academia typically viewsthe ethical indiscretion of plagiarism much like as in the legal world in which such transgres-sions are deemed to occur if undocumented textual reuse is identified. Although not typicallyviewed as malicious textual reuse in comparison to other types of plagiarism, the undocu-mented reuse of text by the original author, commonly referred to as self-plagiarism, is stillconsidered a breach of academic integrity (U.S. Office of Health and Human Services, 2013).

The definition of plagiarism becomes more complicated when using software tools toanalyze texts as common phrases in academic writing and other identifying information(e.g., school names and database listings) account for some overlap in most if not all works(Ison 2012, 2014; Martin et al. 2011). Different measures of the percentage of overlap requiredto be considered Bplagiarism^ exist in the literature. Allan et al. (2005) cautioned not to use toolow of a threshold because almost all papers have some sort of overlap just to language used inacademic work and other aforementioned attributes, for example, when the cutoff wasidentified as 1 % over 90 % of papers contained textual overlaps. Even in light of the guidanceof Allan et al. (2005), Batane (2010) considered the 1 % threshold as the dividing line betweenlegitimate and plagiarized research. In a study analyzing 40,000 papers using Turnitin®, a textmatching detection tool that identifies a similarity index, i.e., the percentage of textual overlap,used a cutoff of 3 %. Garden (2009) recommended the use of a 5 % similarity index forplagiarism labeling whilst Bretag and Mahmud (2009) utilized 10 %. Sikes (2009) found thatseveral institutions have adopted the 10 % cutoff as sufficient evidence to consider the workhaving plagiarized elements.

The Influence of the Internet on Plagiarism Among Dissertations 153

Beyond the Byes^ or Bno^ existence of plagiarism, research has classified the level ofplagiarism in different ways. Batane (2010) registered Turnitin® results with 1–34 % as low-scale, 35–69 % as medium-scale, and 70 % or more being high-scale plagiarism. Turnitin®color-codes its findings into measures of plagiarism, grouping one word to 24 % as green, 25–49 % as yellow, 50–74 % as orange, and 75 % or more as red (iParadigms 2012). Ison (2012)used a hybrid classification based upon the recommendations of Sikes (2009), Bretag andMahmud (2009), and iParadigms (2012) adopting 0–10 % as no or little plagiarism, 11–24 %as low level, 25–49 % as medium level, and 50 % or greater as high level plagiarism.Irrespective of definition or severity, the commonality among all of the outlined research isthat some level of suspicion exists when there are large textual overlaps and, in such cases,further action by the reviewer may be required (Ison 2012, 2014; Walker 2010; Sikes 2009).

Plagiarism in the Online Environment

While there is no shortage on research on the topic of academic ethics, integrity, andplagiarism, a significant amount of focus in recent years has centered on the online educationenvironment and the use of the Internet in coursework. Many of the available studies purport alink between the increase in Internet use and a deterioration in academic integrity amongstudents. Studying higher education students and institutions, several researchers reportedincreases in increases in plagiarism in the online environment and/or via the use of Internetmaterials (Eret and Ok 2014; Jones et al. 2013; Sheard and Dick 2011). Recent scandals andresearch specific to the topic indicate similar trends in graduate education Connell and Mallette2005; Diehl and Trenkamp 2013; Gopalakrishnan 2011; Ison 2012, 2014; Leingang 2006;Pyle 2010).

History of Internet Use in Higher Education

The history of the use of the Internet on college campuses and online education can be tracedback to the early 1990’s. The introduction of the Gopher browser in 1991 was one of the firstuniversity sponsored approaches to accessing the WWW. The University of Phoenix identifiedthat new methods of distributed learning would have great potential to reach a large populationand began building their institution in 1989. In 1993, the first Bvirtual^ institution to gainaccreditation, Jones International University, entered the higher education realm (accreditationwas gained in 1999) (Jones International University 2014). Perry et al. (1998) indicated that by1997, 92 % of American college students had access to a computer and the majority ofcolleges and universities had a website or web pages between 1995–1996. Faculty memberswere also beginning to heavily utilize online course supplements subsequent to 1995. By 1998,43.8 % of college students reported using the Internet on a regular basis, although it was notclear how much of this access was academically related (Perry et al. 1998).

What little literature that is available on the adoption of Internet use on higher educationcampuses indicated that such occurred soon after the proliferation of the WWW from 1992–1994 (Elon University n.d.; Maxymuk 1996; Pew Research 2014). Online research databasesbegan to become more prevalent shortly thereafter, with common search systems such asProQuest arriving around 1996 (Borgman 1996; ProQuest n.d.). Research articles begin tomention the Internet as a potential concern for plagiarism with increasing frequency from 1993onwards (Berge and Collins 1994; Scott et al. 1994). A multi-database search with thekeywords Bstudent^ and Bplagiarism^with Bonline^ or BInternet,^ limited to scholarly journals,

154 D.C. Ison

yielded only 30 articles from 1991–1993 and none were studies of occurrence within thetimeframe, focusing instead on the impending maleficence. From 1994–1998, the countincreased to 80, with articles beginning to implicate the use of online sources in the conductof plagiarism starting to increase from 1997 onwards. From 1999 through 2014, this articlecount has grown to 3210 with numerous studies on the use of Internet sources in the conduct ofplagiarism and the influence the Internet may have on these occurrences (EAGLEsearch 2014).

Internet Plagiarism As the use of the Internet has rapidly expanded, so has the use of onlinesourcing for the conduct of academic endeavors from the proliferation of online institutionsand programs to the use of online databases and websites for references in papers, theses, anddissertations (Ison 2014, 2012; Li 2013; Eret and Ok 2014; Postle 2009; Walker 2010). Alongwith this, it has been reported in a range of literature that the Internet is influencing or evencausing an increase in maligned used of online material by students: BAs the Net generationhas increasingly easy access to the Internet in their lives, literature has revealed that, in recentyears, more and more young people at school are practicing cheating, and things are gettingworse each year^ (Ma et al. 2008, p. 198).

Stricherz (2001) found that the Internet is a common source for undocumented studentsourcing with more than half of students (n=4500) reported plagiarized online resources.Moreover, Bemails and chat rooms on the Internet are also frequently used for plagiarism^ (Maet al. 2008, p. 199). In support of the theory that the Internet is negatively influencing studentwork, Ma et al. (2008) found that Bthe Internet can be viewed as a constructivist learningenvironment where students engage in meaningful learning that is relevant to their ownpersonal interests, which can, in turn, form a social community that encourage plagiarism^(p. 199). Moreover, Eret and Ok (2014) illustrated that Bwhile plagiarism has been a growingproblem in higher education for a long time, the use of the Internet has made this increasingproblem more unmanageable^ (p. 1002). Further exacerbating the issue is the availability ofonline paper writing services which allows students to easily buy papers, theses, and disser-tations on any subject of their choosing (Howard 2007; Ma et al. 2008). Wang (2008) claimedthe lure of the use of Internet sources was a matter of convenience with the ability for studentsto essentially find an unlimited resource for data on a wide range of subjects that can be easilytransferred into one’s document. In their study of 451 college students, Wang (2008) found that80 % of students believed the Internet has made plagiarism a more serious problem with 75 %of students reporting that the online environment made is easy to cut and paste material. Thirtypercent of students stated that they had used material without citing it and 62 % said theybelieved others had done the same (Wang 2008).

Whilst a large volume of research has been conducted into student cheating withinsecondary and undergraduate education, there is also evidence of a growing problem at thegraduate level. Using Google, McCullough and Holmberg (2005) found that 27 % of a sampleof 210 master theses had evidence of plagiarism. Other studies reported issues of student abuseof online sources among graduate students. Carroll (2013) stated that the rapid demand for textmatching detection software is a reaction to increased transgressions by advanced students.Gilmore et al. (2010) also noted that graduate student plagiarism was of concern and worthy offurther inquiry. Glassman et al. (2011) maintained that the inappropriate use of online materialamong graduate students was alarming enough to dictate the construction of a systematiceducation process to better educate such neophyte scholars on how to avoid source misuse. Inan empirical study of master’s theses, Eckel (2011) established that 53 % (n=100) of thesescontained multiple strings of uncited text. In an evaluation of graduate education at Ohio

The Influence of the Internet on Plagiarism Among Dissertations 155

University, Roig (2010) found that numerous theses and at least one dissertation hadsignificant textual overlap with source material. Lastly, Powers (2008) discovered B‘rampantand flagrant plagiarism’ by graduate students^ (para. 4).

Perhaps the most troubling cases of plagiarism are those that occur at the doctoral level.Because so much of the doctorate degree revolves around and depends upon the performanceand completion of the dissertation, misrepresentation of material in this critical documentessentially makes the degree itself illegitimate (Connell and Mallette 2005; Daramola 2014;Diehl and Trenkamp 2013; Gopalakrishnan 2011; Ison 2014, 2012; Pyle 2010). The stakes arealso much higher at the doctoral level. The degree is generally considered the highest level ofacademic attainment and noted to be a terminal degree. Individuals who achieve the PhD (orequivalent) have a higher level of potential income, the ability to apply for positions inacademia and in government, have a higher level of prestige among peers, and are assumedto have a high level of research skill (O’Connor 2012; Tzanakou 2014; U.S. News and WorldReport 2014). At the same time, the discovery of plagiarized doctoral work has, in many cases,extreme consequences such as removal from faculty or government positions and in somecases the revocation of the doctorate itself (Connell and Mallette 2005; Daramola 2014; Diehland Trenkamp 2013; Gopalakrishnan 2011; Ison 2014, 2012; Pyle 2010). In fact, a study byHarirforoosh et al. (2011) discovered that more than 10 % of dissertations written by applicantsfor faculty positions were found to contain plagiarized materials.

Although the availability of data on the prevalence of plagiarism at the doctoral level is verylimited, Ison (2012) found that 60 % of dissertations (n=100) had some level uncited material.In a comparison of dissertations written by students at online institutions versus those writtenby traditional institution students, it was found that 57 and 61 % of dissertations (n=368),respectively had plagiarized material (Ison 2014). Jeffery and Fries (2011) also discoveredunauthorized use of work in dissertations, although the sample size was limited (n=6). Otherconcerning cases of plagiarism exist within the medical sciences. Eaton (2004) discovered that25 % of medical students had plagiarized during their studies. Also first, second, and third yearpharmacy students sometimes had problems identifying what was considered plagiarism inwriting assignments (DeGeeter et al. 2014).

Has the Internet Really Made Things Worse?

Even in light of the arguments made within the exigent literature, there is little data basedevidence that indicates that there has been an increase in plagiarism following the widespreadadoption of the online environment. Most of the blame on the Internet for a degradation ofacademic ethics or its subsequent effect of increasing plagiarism is unjustified, relying more onnews headlines, student self-report studies, and conjecture. In fact, Howard (2007) argues thatthe purported rise of plagiarism is a function of improved detection and awareness rather thanan actual increase: BThe biggest threat posed by Internet plagiarism is the widespread hysteriathat it precipitates^ (p. 12). Wang (2008), McCabe (2009), Hart and Morgan (2010), and Ison(2012, 2014) found that it is unclear as to whether plagiarism is increasing due to the lack ofavailable evidence to support such claims. Heckler et al. (2013) specifically mentioned thatempirical evidence concerning a rise in plagiarism is missing from the literature. Hart andMorgan (2010) explicitly called for the study of plagiarism trends in the online environment.Considering that there has been little, if any, data driven studies on the potential change in theprevalence of plagiarism since the beginning of the widespread use of the Internet, this study

156 D.C. Ison

sought to provide a pragmatic inquiry into the potential impact the online environment has hadon plagiarism in higher education. Due to the prominence of doctoral degrees, the high stakesrewards that are associated with the degree, and the fact that the degree relies so heavily on theperformance of the dissertation, this study focused on this the most advanced level of highereducation – the doctorate.

Method

This quantitative, comparative study sought to determine if the level of plagiarism has changedfrom a period before Internet prevalence in higher education to the current period in whichInternet use is ubiquitous. Specifically, this study examined the prevalence of suspectedplagiarism at the highest level of higher education, the doctorate, by examining dissertationswith Turnitin® text matching detection software.

Dissertations

Rationale for Sample Sizes In order to insure adequate sample size, an a priori poweranalysis was conducted using G*Power software (Universitat Dusseldorf 2013). Due to thedepartures from normal distribution in previous research on dissertations utilizing simi-larity indices generated by Turnitin®, it was initially assumed that non-parametric analysiswould be most appropriate, specifically a Mann-Whitney U test. Per the recommendationsby Cohen (1988), a medium effect size was assumed; however, in order to protect frompotential under-sampling, the lowest effect size value classified as medium by Cohen(1988), d=0.3, was used for power calculations. The analysis with G*Power determinedthat a minimum sample size for a two-tailed Mann-Whitney U assuming α=0.05, 1−β=0.80, and a medium effect size (d=0.3) was n=368. G*Power was also used to determinerecommended sample sizes for chi square analysis. Assuming α=0.05, 1−β=0.80, and amedium effect size (w=0.3) was n=108 with df=2 and n=88 for df=1 which clearly wouldbe met with the recommended sample size for the Mann-Whitney U analysis.

ProQuest Dissertation and Theses database was utilized to harvest dissertations for analysiswith 184 dissertations randomly being selected from the period before widespread use of theInternet (prior to 1994) and 184 randomly selected from the recent period in which the Internetuse is pervasive (since 2010). Dissertations that were included in this study had publicationdate ranges of 1991 to 1993 and 2010 to 2014. The selection of dissertations was limited tothose written by Doctor of Philosophy (PhD) candidates, presented in English, and publishedunder the supervision of faculty at accredited institutions within the United States and Canada.No restrictions or quotas were placed on the subject areas of dissertations to insure aspectrum of examples as this study sought only to evaluate the presence of plagiarismat this level of study in general. No other restrictions were imposed on the selectionprocess, per the methods used in studies with similar goals of evaluating the generalincidence of plagiarism within a stratus of higher education (Heckler et al. 2013;Mahmud and Bretag 2014; Scanlon and Neumann 2002, and Stuber-McEwen et al.2009). Because some dissertations from this period were simply scanned into PDFdocument type form, all dissertations were confirmed to have optical character recog-nition (OCR) capabilities prior to submitting to Turnitin® so that the software couldaccurately analyze the text.

The Influence of the Internet on Plagiarism Among Dissertations 157

Procedures

Following the random collection of dissertations and the evaluation of OCR, the files werebulk uploaded to Turnitin® for analysis. Common academic and database language such asdatabase identification, school names, title page information, and copyright data were removedfrom analysis. Additionally, reference lists, quotations, and definitions were also removed fromthe similarity index calculation. Dissertations were reviewed to insure there were no cases ofprospective plagiarism, i.e., it was confirmed that similarity overlaps did not occur as a resultof publications written after the dissertation was issued. Microsoft Excel® was used to trackand organize the similarity indices. The Excel® file was then uploaded to SPSS® to conductthe apposite statistical analysis.

Results

Dissertation similarity indices were first analyzed for normality assumptions. Kolmogorov-Smirnov and Shapiro-Wilk tests (p<0.001) indicated that the data did not meet such assump-tions (Razali and Wah 2011). Q-Q plots and histograms were also used to confirm these resultsand subsequently the findings precluded the use of parametric analysis. See Table 1 for theresults of these tests as well as the histograms of the data noted in Figs. 1 and 2.

For the dissertations written since 2010 (post-Internet era), the mean similarity index was12.32 % (SD=8.50). In contrast, the mean similarity index for the dissertations written prior to1994 (pre-Internet era) was 14.53 % (SD=7.71). A Mann-Whitney U test was utilized tocompare the results from each era of dissertations. There was a statistically significantdifference between the similarity indices of dissertations written in the post-Internet era(Mdn=11) and those of dissertations written during the pre-Internet era (Mdn=13), U=30,098.5, p<0.001, r=0.197. Figure 3 shows the box plots for the dissertation samples andmarks the location and numbers of extreme scores as well as outliers.

Plagiarism categories adopted by Ison (2012) based upon the recommendations of Bretagand Mahmud (2009) and iParadigms (2012) were used to group findings. Among dissertationswritten in the pre-Internet era, 33 % showed little or no evidence of plagiarism (≤10 %similarity index) and 53.8 % had a low-level similarity index range (11–24 %). For themedium-level range (25–49 % of similarity), 13.2 % fell into this category. No dissertationsin the pre-Internet sample showed evidence of high levels of evidence for plagiarism (50–74 %similarity or ≥75 % similarity). Grouping the post-Internet era dissertations, 41.8 % werewithin the little or no evidence range, 49.4 % were in the low-level range, 6.5 % were within

Table 1 Tests for normality

Tests of normality

Era Kolmogorov-Smirnova Shapiro-Wilk

Statistic df Sig. Statistic df Sig.

SimIndex Post .223 184 .000 .766 184 .000

Pre .128 184 .000 .936 184 .000

a Lilliefors Significance Correction

158 D.C. Ison

the medium range, and 2.3 % were within the high range. A comparison of the two group setsis displayed in Table 2.

A chi square test for independence was calculated to compare pre/post-Internet results. Dueto the low numbers within the high incidence groupings, these were combined with mediumgroupings. The results indicated that there was no statistically significant difference insimilarity index groupings between pre- and post-Internet dissertations χ2 (2, N=368)=3.79,p=0.15, Cramer’s V=0.101. Similarly, the results of a chi square test for independence of pre-and post-Internet era dissertation groups in terms of little/no evidence of plagiarism versusevidence (low, medium, or high) indicated no significant difference, χ2 (1, N=368)=2.61, p=0.11, Cramer’s V=0.09 (with Yates’ continuity correction).

Discussion

The results of this study indicate that, contrary to what has been reported in the media and inthe literature, evidence does not support the notion that the Internet has had a significantnegative affect on the incidence of plagiarism. Although the current study provides only asnapshot of what may be occurring within academia, it does provide empirical evidence thatmay contradict previous unsubstantiated or perfunctory reports of the Internet causing asignificant negative influence on student work. The findings of this study seem to contradict

Fig. 1 Similarity index distribution of dissertations from before 1994

The Influence of the Internet on Plagiarism Among Dissertations 159

the argument made by Ma et al. (2008) that Bthings are getting worse each year^ (p.198). Such also appears to challenge the assertion of Eret and Ok (2014) who claimedthat the plagiarism has increased within higher education to the point that it is nowBmore unmanageable^ (p. 1002).

The statistical analysis of the dissertations revealed two different approaches to dispellingthe deleterious effects of the Internet on academic honesty. The inquiry into differences in thesimilarity indices of pre- and post-Internet era dissertations showed that there was a statisticallysignificant difference between the collections though it was opposite that purported bycontemporary authors and researchers with the pre-Internet era dissertations having a highermean and median similarity index. This indicates that the level of severity of plagiarism maynot have increased as anticipated. Although the fact that both samples of dissertations havesome level of evidence of plagiarism above 50 % is troubling, this aligns with the findings ofIson (2012, 2014) as well as Jeffrey and Fries (2011). Moreover, this aligns with previousevidence of similar levels (53 %) of plagiarism incidence within theses documented by Eckel(2011). This supports the notion that plagiarism is a continuing problem albeit one that hasbeen around for quite some time (Febvre and Martin 1976; Howard 2007; MacQueen et al.2007; Postle 2009).

Viewing the data from a different perspective, the chi square results showed that theincidence of occurrence (counts) of suspected plagiarism and little/no evidence thereof wasstable between groups. When extending this analysis to the various levels of plagiarism (little/none, low, medium/high), the results were the same – the frequencies of occurrence were not

Fig. 2 Similarity index distribution of recent year dissertations

160 D.C. Ison

significantly different. These findings also may support the need to reconsider the notion,common within the research literature, that plagiarism had increased as a result of morefrequent use or exposure to the Internet (Li 2013; Eret and Ok 2014; Postle 2009; Walker2010). Certainly more empirical evidence is necessary in order to determine the potentialimpact that the Internet has had on student research of all types.

Examining the distributions of each group of dissertations (Figs. 1 and 2) affords a clearerpicture as to the characteristics of the characteristics of plagiarism potential. Within the pre-Internet group, the variation among similarity indices is smaller than that of post-Internetdissertations. One possible explanation for this contravention is that the older the work that issubmitted to text matching software, the more likely sources used in writing the analyzed texthave been integrated into the database used by the software. It may also be related to theincreased number of individual databases that are collected by the software over time (Butakovand Scherbinin 2009; Garner 2011; Zu Eissen and Stein 2006). In contrast, the post-Internetgroup had more extreme cases as indicated by the frequency of extreme scores and outliers

Fig. 3 Box plots of dissertation samples. Note: Extreme scores are indicated with Bo^ and outliers aremarkedwith B*^

Table 2 Comparison of similarity indices: pre-internet versus post-internet

Study Little/none Low Medium High Total plag

Pre 33 % 53.8 % 13.2 % 0 % 67 %

Post 41.8 % 49.4 % 6.5 % 2.3 % 58.2 %

The Influence of the Internet on Plagiarism Among Dissertations 161

depicted in Fig. 3. With this data it can be argued that while the frequency of cases and meansimilarity indices do not support the notion of plagiarism getting worse with the widespreaduse of the Internet, an escalation in the extremity of plagiarism may be of concern. This maysupport the observation by Powers (2008) that while plagiarism has been and continues to becommon, it has become more egregious. Simultaneously, one must examine the fact thatoverall suspected plagiarism was lower in the post-Internet group, specifically the percentageindicated to have little or no suspicious textual overlap. Perhaps the introduction of textanalysis software such as Turnitin® is actually positively influencing students to be morecareful in avoiding plagiarism as has been noted in studies on incidence of academicmisconduct before and after such systems are brought into use (Batane 2010; Heckler et al.2013). Nevertheless, without more stringent controls on the samples, e.g., assurances of similarlevels of inclusion of subject matter, such conclusions should be viewed with caution.

Although the incidence of self-plagiarism was not specifically analyzed in this study, duringthe review of the dissertations, it was evident that such was fairly common among both groups.Self-plagiarism most commonly came in the form of un-cited use of academic papers, journalarticles, conference proceedings, and similar works that preceded the dissertation publicationdate. This study sought to identify the overall incidence of plagiarism, per the definitionsoutlined previously. In accordance with the U.S. Department of Health and Human Services(2013) policy on self-plagiarism, which identifies unidentified textual reuse regardless ofauthor as academic misconduct, this standard was adopted within the current study. Moreresearch is necessary to quantify the incidence and severity of self-plagiarism amongdissertations.

There are potential limitations and delimitations to this study. Theoretically, older disserta-tions may be more likely to have more references available to the Turnitin® database whilstnew dissertations may have some references that have not yet been identified by the system.This could explain why similarity indices of pre-Internet dissertations were higher. Moreevidence needs to be collected in the form of different types of artifacts such as generalstudent papers and theses which would provide a more comprehensive view of the prevalenceof plagiarism in each era. In addition, a broader sample could provide insight into whether thepervasiveness of extreme or blatant cases of plagiarism have increased over time. Anotherlimitation is that Turnitin® cannot definitively detect all examples of plagiarism. In addition,there is the possibility of false positive results although the researcher reviewed the disserta-tions to mitigate this issue. It is also important to note that the findings of this study provideempirical evidence of plagiarism among dissertations within the confines outlined in this study,caution is advised in extrapolating the results beyond the scope of the study such as implica-tions of plagiarism within scholarly research journals or within higher education as a whole.

Per the methods used by Heckler et al. (2013); Mahmud and Bretag (2014); Scanlon andNeumann (2002); and Stuber-McEwen et al. (2009), this study sought only to identify thelevels of plagiarism irrespective of subject area. Because this research sought to measure thegeneral prevalence of plagiarism within PhD dissertations with interest to compare thosewritten in two different time ranges, no quotas or limits were used to assure equal quantitiesof dissertations within subject areas. While the definition of plagiarism does not rely on subjectarea, the researcher recognizes that there could be variations in the type and level of plagiarismamong different foci of dissertations. Also, the researcher did not distinguish between plagia-rism and self-plagiarism. While the uncited use of copyrighted material is, by definition,plagiarism, however some circles may view this as less severe of an indiscretion. Futureresearch should consider the influences of these aforementioned issues.

162 D.C. Ison

Conclusions

The fact that more than half of all analyzed dissertations contain evidence of plagiarism isproblematic. Further, the common occurrence of self-plagiarism requires the attention ofacademic stakeholders to pursue standards for referencing material authored by dissertationstudents. Even in light of these facts, the findings of this study suggest that the Internet may nothave had a significant negative affect on the conduct of plagiarism. Aside from the evidencethat newer dissertations may have more incidences of extreme scores or outliers, the meansimilarity indices and frequencies of plagiarism occurrence discount the notion that the Internetis the instigator or cause of increased textual misuse. With the mean similarity index of14.53 % for pre-Internet dissertations and 12.32 % for the post-Internet dissertations beingopposite of what is proposed in media and the literature, a revision of assumptions about thenegative consequences of the widespread use of online research resources may be in order.Hopefully the findings of this research will provide faculty, administrators, researchers, and themedia with relevant, empirical evidence to guide discussions and policies about plagiarism inacademia. Additionally, the data within this study should encourage academic stakeholders toimprove the guidance that is provided to all levels of students to better inform them aboutproper citation methods and academic ethics.

Suggestions for Future Research

Based upon the results of this study, the following future research is suggested:

1. Collect larger samples of dissertations from each era to confirm the incidence of extremecases of textual overlap. This study should be expanded to a broader sample of worksfrom countries outside of North America.

2. Conduct a larger study to include theses or other academic artifacts to verify the findingsof this study. This study should be expanded to a broader sample of works from countriesoutside of North America.

3. Repeat the current study with analysis of dissertations utilizing one or more alternativetext matching detection systems to validate the results of this study.

4. Investigate the potential influences of various student factors on the incidence of plagia-rism to include the major or subject area of study, the student gender, the nationality of thestudent, geographic location of the institution, and the type of institution.

5. Inquire into the levels of plagiarism within specific segments of dissertations, e.g.,literature reviews versus results, from a global sample of works.

6. Conduct a study of dissertations and theses to evaluate the source materials commonlyused in plagiarized text to determine if they originate from online or other sources.

References

Allan, G., Callagher, L., Connors, M., Joyce, D., & Rees, M. (2005). Some Australasian perspective on academicintegrity in the Internet age. Paper presented at the EDUCAUSE Conference, Auckland, NZ.

Bailey, J. (2014). Top plagiarism scandals of 2013. Retrieved from http://www.ithenticate.com/plagiarism-detection-blog/top-plagiarism-scandals-of-2013#.VElCnfnF-sl

Batane, T. D2010]. Turning to Turnitin to fight plagiarism among university students. Educational Technology &Society, 13D2], 1–12. Retrieved from http://www.ifets.info/journals/13_2/1.pdf.

The Influence of the Internet on Plagiarism Among Dissertations 163

Berge, Z., & Collins, M. (1994). Publishing and editing an electronic journal on the Internet. Retrieved fromhttp://eric.ed.gov/?id=ED373701

Borgman, C. (1996). Why are online catalogs still hard to use? Journal of the American Society for InformationScience, 47(7), 493–503. doi:10.1002/(SICI)1097-4571(199607)47:7<493::AID-ASI3>3.0.CO;2-P.

Bretag, T., &Mahmud, S. (2009). A model for determining student plagiarism: electronic detection and academicjudgment. Journal of University Teaching & Learning Practice, 6(1), 6.

Butakov, S., & Scherbinin, V. (2009). The toolbox for local and global plagiarism detection. Computers &Education, 52(4), 781–788.

Carroll, N. (2013). E-learning–the McDonaldization of education. European Journal of Higher Education, 3(4),342–356.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Psychology PressTaylor Francis Group.

Connell, M., & Mallette, M. (2005). The right of educational institutions to withhold or revoke academicdegrees. Paper presented at the 26th Annual National Conference on Law and Higher Education, Oxford,Mississippi.

Daramola, D. (2014). Plagiarism: the undiagnosed Ebola in Nigeria. Leadership [online version]. Retrieved fromhttp://leadership.ng/news/386536/plagiarism-undiagnosed-ebola-nigeria

DeGeeter, M., Harris, K., Kehr, H., Ford, C., Lane, D., Nuzum, D. S., Gibson, W., et al. (2014). Pharmacystudents’ ability to identify plagiarism after an educational intervention. American Journal ofPharmaceutical Education, 78(2), 1–33.

Diehl, J., & Trenkamp, O. (2013). Plagiarism accusations: Merkel’s education minister has Ph.D. title revoked.Spiegel Online International. Retrieved from http://www.spiegel.de/international/germany/education-minister-schavan-has-ph-d-revoked-in-plagiarism-scandal-a-881707.html

EAGLEsearch. (2014). EAGLEsearch. Retrieved from http://library.erau.edu/home.htmlEaton, L. (2004). A quarter of UK students are guilty of plagiarism, survey shows. BMJ [British Medical

Journal], 329(7457), 70.Eckel, E. J. (2011). Textual appropriation in engineering master’s theses: a preliminary study. Science and

Engineering Ethics, 17(3), 469–83. doi:10.1007/s11948-010-9214-6.Elon University. (n.d.). Imagining the Internet’s: Quick look at the early history of the Internet. Retrieved from

http://www.elon.edu/e-web/predictions/early90s/internethistory.xhtmlEret, E., & Ok, A. (2014). Internet plagiarism in higher education: tendencies, triggering factors and reasons

among teacher candidates. Assessment & Evaluations in Higher Education, 39(8), 1002–1016. doi:10.1080/02602938.2014.880776.

Febvre, L., & Martin, H. (1976). The coming of the book: The impact of printing 1450-1800. London: New LeftBooks.

Garden, C. (2009). What does my Turnitin report mean? Edinburgh: Edinburgh Napier University.Garner, H. R. D2011]. Combating unethical publications with plagiarism detection services. Urologic Oncology:

Seminars and Original Investigations, 29D1], 95–99. Elsevier.Gibelman, M., & Gelman, S. (2003). Plagiarism in academia: trends and implications. Accountability in

Research, 10(4), 229–252.Gilmore, J., Strickland, D., Timmerman, B., Maher, M., & Feldon, D. (2010). Weeds in the flower garden: an

exploration of plagiarism in graduate students’ research proposals and its connection to enculturation, ESL,and contextual factors. International Journal for Educational Integrity, 6(1), 13–28.

Glassman, N., Sorensen, K., Habousha, R., Minuti, A., & Schwartz, R. (2011). The plagiarism project. MedicalServices Quarterly, 30(4), 337–348.

Gopalakrishnan, J. (2011). Revocation of 3 German politicians’ Ph.Ds for plagiarism, 3 reactions. Retrievedfrom http://www.ithenticate.com/plagiarism-detection-blog/bid/60500/Revocation-of-3-German-Politicians-Ph-Ds-For-Plagiarism-3-Reactions#.VElVrfnF-sk

Harirforoosh, S., Bossaer, J. B., Brown, S. D., Pond, B. B., Ramsauer, V. P., & Roane, D. S. (2011). Plagiarismamong applicants for faculty positions. American Journal of Pharmaceutical Education, 75(10), 1.

Hart, L., & Morgan, L. (2010). Academic integrity in an online registered nurse to baccalaureate in nursingprogram. Journal of Continuing Education in Nursing, 41(11), 498–505.

Heckler, N. C., Rice, M., & Bryan, C. H. (2013). Turnitin systems: a deterrent to plagiarism in collegeclassrooms. Journal of Research on Technology in Education, 45(3), 229–248.

Howard, R. M. (2007). Understanding Binternet plagiarism^. Computers and Composition, 24(1), 3–15.iParadigms. (2012). Instructor user manual. Retrieved from http://Turnitin.com/en_us/training/instructor-

training/instructor-user-manualIson, D. C. (2012). Plagiarism among dissertations: prevalence at online institutions. Journal of Academic Ethics,

10(3), 227–236. doi:10.1007/s10805-012-9165-4.

164 D.C. Ison

Ison, D. C. (2014). Does the online environment promote plagiarism? A comparative study of dissertations frombrick-and-mortar versus online institutions.MERLOT Journal of Online Teaching and Learning, 10(2), 272–282.

Jeffery, D., & Fries, J. (2011). Unauthorized uses of a coauthored work and a doctoral dissertation. Ethics &Behavior, 21(2), 118–126. doi:10.1080/10508422.2011.551469.

Jones International University. (2014). Our history. Retrieved from http://www.jiu.edu/about-jiu/our-historyJones, I. S., Blankenship, D., & Hollier, G. (2013). Am I cheating? An analysis of online student perceptions of

their behaviors and attitudes. Proceedings of the American Society of Business and Behavioral Sciences, 20,21–24.

Kutz, E., Rhodes, W., Sutherland, S., & Zamel, V. (2011). Addressing plagiarism in a digital age. HumanArchitecture: Journal of the Sociology of Self-Knowledge, 9(3), 15–36.

Leingang, M. (2006). Ohio college stung by plagiarism charges. The Washington Post [online version]. Retrievedfrom http://www.washingtonpost.com/wp-dyn/content/article/2006/08/21/AR2006082100068.html

Li, Y. (2013). Text-based plagiarism in scientific publishing: issues, developments, and education. ScienceEngineering Ethics, 19(3), 1241–1254. doi:10.1007/s11948-012-9367-6.

Ma, H., Lu, Y., Turner, S., & Wan, G. (2006). An empirical investigation of digital cheating and plagiarismamong middle school students. American Secondary Education, 35(2), 69–82.

Ma, H., Wang, G., & Lu, E. (2008). Digital cheating and plagiarism in schools. Theory Into Practice, 47(3), 197–203.

MacQueen, H. L., Charlotte, W., & Graeme, L. T. (2007). Contemporary intellectual property law and policy.Oxford: Oxford University Press.

Mahmud, S., & Bretag, T. (2014). Integrity in postgraduate research: the student voice. Science and EngineeringEthics, 2014, 1–16. doi:10.1007/s11948-014-9616-y.

Martin, D., Rao, A., & Sloan, L. (2011). Ethnicity, acculturation, and plagiarism: a criterion study of unethicalacademic conduct. Human Organization, 70(1), 88–96.

Maxymuk, J. (1996). Riding the technology waves in search of electronic access: CD-ROM, gophers, the WWWand beyond. Journal of Government Information, 23(3), 327–334.

McCabe, D. L. (2009). Academic dishonesty in nursing schools: an empirical investigation. Journal of NursingEducation, 48(11), 614–623.

McCullough, M., & Holmberg, M. (2005). Using the Google search engine to detect word-for word plagiarism inmaster’s theses: a preliminary study. College Student Journal, 39(3), 435–441.

Moss, S. (2005). A history of plagiarism (not my own work). Retrieved from http://www.guardian.co.uk/books/2005/nov/23/comment.stephenmoss

O’Connor, C. (2012). Top 5 factors to weigh before getting a PhD. Retrieved from http://www.boston.com/business/blogs/global-business-hub/2012/02/the_top_5_facto.html

Perez-Pena, R. (2013). Students disciplined in Harvard scandal. The New York Times [online version]. Retrievedfrom http://www.nytimes.com/2013/02/02/education/harvard-forced-dozens-to-leave-in-cheating-scandal.html

Perry, T. T., Perry, L. A., & Hosack-Curlin, K. (1998). Internet use by university students: an interdisciplinarystudy on three campuses. Internet Research, 8(2), 136–141.

Pew Research. (2014). Internet project: World Wide Web timeline. Retrieved from http://www.pewinternet.org/2014/03/11/world-wide-web-timeline/

Plagiarism. (2014a). Online etymology dictionary. Retrieved from http://www.etymonline.com/index.php?allowed_in_frame=0&search=plagiarism&searchmode=none

Plagiarism. (2014b). Cambridge dictionaries online. Retrieved from http://dictionary.cambridge.org/us/dictionary/american-english/plagiarize?q=plagiarism

Postle, K. (2009). Detecting and deterring plagiarism in social work students: implications for learning forpractice. Social Work Education, 28(4), 351–362. doi:10.1080/02615470802245926.

Powers, E. (2008). ‘Distinguished’ no longer. Inside Higher Ed, 22(2). Retrieved from http://www.insidehighered.com/news/2008/02/22/ohio

ProQuest. (n.d.). History and milestones. Retrieved from http://www.proquest.com/about/history-milestones/Pyle, E. (2010). Plagiarism a persistent problem on campuses. The Columbus Dispatch [online version].

Retrieved from http://www.dispatch.com/content/stories/local/2010/09/15/plagiarism-a-persistent-problem-on-campuses.html

Razali, N. M., & Wah, Y. B. (2011). Power comparisons of Shapiro-Wilk, Kolmogorov Smirnov, Lilliefors andAnderson-Darling tests. Journal of Statistical Modeling and Analytics, 2(1), 21–33.

Research Misconduct – Definitions, 45 C.F.R. § 689.1 (2012).Rogers, S. (2014). Plagiarism attitudes common across cultures. Yale Daily News [online version]. Retrieved

from http://yaledailynews.com/blog/2014/02/11/plagiarism-attitudes-common-across-cultures/

The Influence of the Internet on Plagiarism Among Dissertations 165

Roig, M. (2010). Plagiarism: An ounce of prevention…. ASQ Higher Education Brief, June. Retrieved fromhttp://rube.asq.org/edu/2010/06/continuous-improvement/plagiarism-an-ounce-of-prevention-.pdf

Roig-Franzia, M. (2014). ‘A fragile trust’ exhibits irresponsibility behind Jason Blair plagiarism scandal. TheWashington Post [online version]. Retrieved from http://www.washingtonpost.com/lifestyle/style/a-fragile-trust-exhibits-irresponsibility-behind-jayson-blair-plagiarism-scandal/2014/05/05/596aca28-d45d-11e3-95d3-3bcd77cd4e11_story.html

Scanlon, P. M. (2003). Student online plagiarism: how do we respond? College Teaching, 51(4), 161–165.Scanlon, P. M., & Neumann, D. R. (2002). Internet plagiarism among college students. Journal of College

Student Development, 43(3), 374–385.Scott, T. J., Kallman, E. A., & Lelewer, D. (1994). Ethical issues involving the Internet. In Proceedings of the

conference on Ethics in the computer age (pp. 31–32). ACM.Sheard, J., & Dick, M. (2011). Computing student practices of cheating and plagiarism: a decade of change. In

Proceedings of the 16th annual joint conference on innovation and technology in computer scienceeducation (pp. 233–237). ACM.

Sikes, P. (2009). Will the real author come forward? Questions of ethics, plagiarism, theft, and collusion inacademic research. International Journal of Research & Method in Education, 32(1), 13–24.

Stricherz, M. (2001). Many teachers ignore cheating, survey finds. Education Week, 20(34), 3.Stuber-McEwen, D., Wiseley, P., & Hoggatt, S. (2009). Point, click, and cheat: Frequency and type of academic

dishonesty in the virtual classroom. Online Journal of Distance Learning Administration, 12(3).Tzanakou, C. (2014). The wider benefits of a PhD. Retrieved from http://www.universityworldnews.com/article.

php?story=20140409095727966U.S. Department of Health and Human Services. (2013). Self-plagiarism. Retrieved from http://ori.hhs.gov/

plagiarism-13U.S. News and World Report. (2014). Explore the pros and cons of earning a PhD or doctoral degree. Retrieved

from http://www.usnewsuniversitydirectory.com/articles/pros-and-cons-of-earning-a-phd-or-doctoral-degree_12910.aspx#.VGOicvnF-sk

Universitat Dusseldorf. (2013). G*power [computer software]. Available from http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3

Walker, J. (2010). Measuring plagiarism: researching what students do, not what they say they do. Studies inHigher Education, 35(1), 41–59.

Wang, Y. (2008). University student online plagiarism. International Journal of E-Learning, 7(4), 743–757.Zu Eissen, S. M., & Stein, B. (2006). Intrinsic plagiarism detection. In Advances in Information Retrieval (pp.

565–569). Springer, Berlin Heidelberg.

166 D.C. Ison