How to Recognize Plagiarism
Changes in the Tutorial and Certification Tests
The tutorial and test on How to Recognize Plagiarism was originally developed for use by students in the Instructional Systems Technology (IST) department at Indiana University, starting in September, 2002. As others have since discovered these online resources, usage has been increasing dramatically each year.
In 2008, web logs indicated that about 181,000 Certificates were issued, while in 2012 over 661,000 Certificates were awarded, with about 97% of the traffic to the site coming from outside Indiana University. In 2013, there were over 6.9 million accesses to the online tutorial and test. A Google search for "plagiarism test" typically yields a link to this site at the top of the first page of the search results, which means that many other sites (also with high Google page ranks) now provide links to our web site.
We have learned through correspondence initiated by instructors that many now require their students to take this test. We have no control over who uses our tests and for what purposes. Our goal has always been to help people understand what plagiarism is, so that they do not commit plagiarism in their writing and presentations. Anyone is welcome to use our learning resources and tests on this site for non-profit educational purposes. Test item pools remain proprietary and are not available for duplication elsewhere. This service is provided free of charge. Tests and content are maintained by a few volunteer graduate students and an emeritus faculty member in IST at Indiana University who have donated their time and effort.
Based on feedback we have received from college and high school instructors whose students use our online resources, we made the following changes in 2013-14:
- August, 2013: Questions on the original 10-item test were re-orderd, due to evidence of wide-spread cheating. E-mail from several instructors had identified a YouTube video with an aswer key to the plagiarism test. Each time that we re-mixed the items, new answer keys were soon posted in comments to that YouTube video. The number of views of that YouTube video nearly doubled by mid-August. Consequently, we started planning for ways to improve test validity and to make it harder to cheat on the test.
- Sept. 7, 2013: The number of questions was substantially increased to create a very large inventory. Many new questions included more subtle forms of plagiarism.
- Each 10-item test is a random sample from the large inventory of items.
- Since there were more difficult items, a minimum of 9 out of 10 correct answers was set as the criterion for passing a test.
- There are literally billions of different 10-item tests possible with the new item pool.
- Distinctions between IU and non-IU test takers were no longer made. Instead, once a test has been passed, test takers can supply their name and e-mail address to receive a Certificate with this and other unique information embedded in each Certificate.
- Certificates were e-mailed in addition to displaying them in the user's web browser.
These changes were made in order to make it harder to cheat via use of a distributed answer key. We had asked Google several times to remove the YouTube video. Google complied in September, 2013, due to IU copyright violation by the person who had originally posted the video in December, 2012.
- Sept. 13, 2013: Specific numbers of questions answered correctly and incorrectly on the 10-question tests were no longer provided in test results. This helped prevent obtaining a Certificate by simply guessing answers, using the 'Back' button, making a change, and then getting a new evaluation of the same test to see if the score improved. There was considerable evidence that this strategy was used by many individuals until they passed the test, without actually understanding what plagiarism is.
- Sept. 21, 2013: Only the first attempt for each randomized test provided results and the opportunity to earn a Certificate on a specific computer. This policy of restricting each test to a single attempt was based on evidence of wide-spread cheating via the "back button strategy."
All 10 questions must be answered on each test within 40 minutes. Since these 10-item tests are taken anonymously, the IP number of the computer used for each test indicates when that test is attempted and failed at least once. A new test must be taken for another evaluation of answers.
The IP number was reported on the Certificate, along with the date and time the test was passed, how long it took, and the name supplied by the test taker.
After someone has failed a test, an error message of "too many failed attempts" occurs when the same test is attempted again at the same computer at that IP address. This can be easily remedied by clicking a button to take a new 10-item test. It also prevents multiple attempts of answering the same 10 questions until that test is passed by one or more individuals at the same computer. This is by design. It is not a technical error with the testing system.
In summary, the same computer at the same IP address can be used repeatedly if a new 10-question test is taken each time.
- Sept. 23, 2013: The practice test was enhanced by providing new examples of subtle forms of plagiarism. Specific feedback on each question was given in the practice test, so that students have opportunities to see their mistakes and correct them. Of course, the full tutorial continued to be available for student learning.
- August 13, 2014: A choice of Certification tests was made available for persons who are:
- Undergraduate students and advanced high school students
- less difficult test, large inventory of questions
- 10 randomly selected items for each test administration
- must correctly answer at least 9 out of 10 questions to pass
- primarily for college undergraduates and high school students
- Graduate students
- more difficult test, very large inventory of questions
- randomly selected items administered one at a time
- number of items vary for each test administration (using a computerized classification test method; test lengths typically vary from 8 to 20 questions)
- test ends as soon as pass/fail decision is reached at the 95 percent confidence level
- primarily for college students at the master's and doctoral level.
- Undergraduate students and advanced high school students
- August 23, 2014: Test takers and their instructors can check the validity of a Certificate, if the test was passed after 12 noon (U.S. Eastern Time), August, 13, 2014. The unique Test ID, combined with one other specific fact reported on the Certificate, is required to validate a Certificate. A URL for test validation is displayed on the Certificate under the student signature. If a Certificate is questionable, patterns that could indicate attempts to "game the testing system" are reported, so that instructors can judge for themselves whether further action is warranted.
- August 25, 2014: If a Certificate is determined to be valid (as described above), it can be viewed again in a web browser. This is intended to address the problem of "lost Certificates" (e.g., "My dog ate my homework."). We do not send new Certificates if they have been lost. However, if test takers retain their unique Test IDs, they can retrieve their Certificates themselves.
Certificates earned before August 13, 2014, are not eligible due to changes in the tests. There will be a time limit for how long after a test is passed that this can be done. The very large amount of test records of Certificates earned will be a limiting factor, but we hope to keep records for several months in each semester (fall, spring, and summer) before archival.
- August 30, 2014: Computers sharing the same IP number could now take tests at the same time without causing the "too many failed attempts" issue. This had been reported as a problem in Local Area Networks (LANs) where computers share a router and its IP address. Individual IP numbers within the LAN are hidden behind the router's firewall for security--a good practice (e.g.,in a computer lab or home network). Uniqueness of a test is now determined by a combination of IP number and Test ID.
- September 3, 2014: New examples of 17 patterns of plagiarism and non-plagiarism were created. This was done in response to requests by users for examples in addition to those previously available (5 each on word-for-word and paraphrasing plagiarism). Now there are a total of 25 examples with explanations of how a writing sample can be changed to avoid plagiarism (with side-by-side comparisons) and 2 further examples of appropriate quotation and paraphrasing to avoid plagiarism.
- September 4, 2014: Feedback on types of mistakes made is provided when a Certification Test is not passed. While feedback on specific answers to test questiions is not provided in order to help maintain test item security, an analysis of patterns of errors made on a test is now conducted on each test taken. On the test results page, specific links are dynamically provided based on analysis of types of errors made by each test taker (e.g., to Clueless Quote, Pointless Paraphrase, etc.). This was done in response to requests by test takers for better feedback to help rectify their mistakes in identifying word-for-word plagiarism, paraphrasing plagiarism, and non-plagiarism.
- September 5, 2014. Test items were modified to make it easier to spot similarities in the original source material and student versions. Particularly relevant parts of the source were bolded for each item, so that test takers can better focus on whether the student version is plagiarized, and if so, what kind according to the criteria we use.
Second, the language on the test buttons and Certificates was changed to clarify the main target audiences (undergraduate and advanced high school students).
- September 7-11, 2014. Examples and practice sections of the tutorial were improved. Over the past 12 years, inadvertent inconsistencies in language used to describe plagiarism had been introduced. Some students reported getting confused by these inconsistencies in language. As part of the content revision for the Examples and Practice, inconsistent wording and layout was fixed, and new color cues were introduced to highlight and distinguish important terms.
Immediately following the changes introduced September 3-7, 2014, the rate of passing a Certification test improved by 141 percent. Over 34,000 new Certificates issued Sept. 5 - Oct. 31.
- September 9, 2014. Feedback in the practice test was enhanced. In addition to the previous hints on questions answered incorrectly, specific links to further examples are now provided with the hints for the kind of error made.
- September 11, 2014. The method of generating unique Test IDs was changed. The test ID is now a long string of numbers, and no longer includes a decimal point, as of 8 a.m. on Sept. 11. The validation service will accept both the previous and current forms of test IDs.
See AECT presentation on history of plagiarism test, 2002-2014 (PDF, 5.7mb)