Confidence in the certificates of the MEP or “Vasya's dad is strong in mathematics”
The popularity of Coursera, Udacity and other sites for distance learning is gaining momentum. Everyone is familiar with the common structure of the MEP-courses: video lectures, when viewed which provides mini-tests for learning, homework with a deadline of implementation, minimal opportunities for joint discussion of the course by students. The student’s identity confirmation is minimal: just an email is enough to start the course, and in the end to receive for some courses - a paper certificate signed by a large university.
But is such a structure enough to trust the MEP-certificates of universities? We tried to isolate the shortcomings of the MEP and supplement the existing system with methods to control and protect against cheating, and to test these hypotheses, we implemented our prototype of a web service for distance learning.
Consider the now generally accepted structure of online courses in terms of its security and protection from cheating and abuse.
Take an example from a familiar IT area. Suppose you are looking for a Python-developer (or Java, C ++) in a team, with knowledge of machine learning, AI algorithms, and also consider sensible juniors without experience. There is no one candidate Vasya - there are several Coursera paper certificates, Udacity for specialized disciplines, supposedly “his” project with source codes on github, while there is no real programming experience in the team - but he seems to be answering your technical questions and promises to learn quickly . Maybe this will be the case, and it will make a good developer.
But, what if Coursera certificates and “their” project are ordered “to outsource” from an assistant through a freelance service at a reasonable price?
')
In total, we have - due to the lack of control over the online training services and the opportunity to complete the courses with an assistant, Vasya can greatly overstate his “external” rating and expectations from his work, and it’s not a fact that later he will be able to meet those expectations. Everybody loses from this: those who trust certificates - they will lose their time with Vasya, and Vasya himself - trying to deceive the service, instead of really learning and mastering technologies - he loses his time and yours. Perhaps, thanks to this type of thinking, Vasya will eventually become a good entrepreneur or a hacker (we hope white), but a good developer for whom he will try to pretend to himself will not immediately become. Given the high demand for any programmers - he is likely to be able to go through 1-2 bad HR filters. It is possible not to become attached only to IT, and apply this analogy to the majority of other areas where online courses are read and validated by the MEP certificates are issued that affect the user rating outside the online training service. Therefore, the unprotected rating and lack of control is unprofitable for everyone, including the online services themselves, as this reduces the credibility of the received certificates.
Let us take another example of an easily winding rating - some public ballots (for photos, for selecting music bands for the festival) take into account the simple likes of Facebook or VKontakte as votes. Everyone knows that “likes” can be easily screwed up, and you can google several services that will do it quickly and inexpensively. Thus, as a result of a winding up rating of participants, if there is no additional commission of experts in the voting, or there is no technology for eliminating markups and punishing them - the first prizes will be received by mediocre participants, and interesting newcomers without “like” support will not be able to get to the top. Voting for the quality of the product of the participants turns into a banal "clog up more likes." I think these two examples are enough to understand the value of the fight against cheating in any field. Consider the technology cheat existing online services MEP.
So, Vasya decided to carefully prepare for the interview for your position - google the FAQ for the profile interview, read “Python (java, C ++) in 24 hours” or an analogue, put some money into the Chinese freelance assistant (let's call him Dad), who knows the necessary discipline and with whom he will take courses and receive certificates. I registered on Coursera or analog, found the necessary courses, ticked off "I will not write off", signed up for the necessary courses. Then Vasya passes the access to Coursera to the Pope, and the Pope begins to take all the courses for him. Within the framework of the existing system, it is impossible to take any measures against Vasya, since in no lecture or test his personality is confirmed, and Vasya can easily pass the course to the Pope, as well as homework. Dad is a good specialist, and passed all courses in good faith, deserved the issuance of a confirmed certificate for them. At this moment, Vasya returns, sends his personal data to confirm the identity - and a pack of certificates for the courses that Pope has completed in Vasya's pocket. Vasya received the necessary certificates, Papa - a good 50-100 $ for passing the courses by Chinese standards, both parties are satisfied. The probability to burn Vasya’s scheme with existing technologies of MEP-courses is 0%. To improve the plausibility, Vasya can give the Pope a proxy / VPN on his IP to go to Coursera, so a banal check on the IP - Vasya is also easy to manage. As a result of his actions, Vasya is ready for an interview with the conditional general “external” rating of 50, having only 10 in reality.
It is clear that an experienced recruiter paired with a technical specialist will be able to easily weed out Vasya at an interview or according to the results of a probationary period, but in some companies certificates play a very large role, and there they are ready to take newbies with almost no experience. The purpose of the example is not to blame Vasya, he just found a vulnerability in the MEP, but to show that the mere presence of an easily winding rating and certificates undermines trust in them, and in this area it is important to deal with cheating, as some value of these certificates is declared, in perspective, close to real diploma. Most of the services on online education do not do real anti-cheat protection, or they do it very poorly, for example - the tick “I observe fair play” on Coursera, which does not guarantee anything.
In addition to passing the course with an assistant, other methods of cheating are possible - consider them.
Suppose Vasya still felt remorse and decided to refuse the help of the Pope in principle, to do all the tests and tasks himself. Let us consider what opportunities it has for cheating the test results without the Pope, within the framework of the existing system of MEP courses:
There are quite ample opportunities, and Vasya is happy with this state of affairs, because it is possible, without the expenses of an assistant, to significantly wind up a rating in the course of the course.
After such a debriefing, questions arise:
Question I. If the online education services almost do not interfere with the above methods, then what percentage of their students will be able to wind up the results and get the certificates, repeating after Vasya?
The answer is . According to the LA Times, a survey of the students of the online service of the Online Journal of Distance Learning Administration back in 2010 revealed that 32.7% of students tried cheats at least once in an online course [1] . This is the percentage of only those who admitted this themselves, and in 2010 the MEP was not so common. In today's situation, when the MEP becomes really massive, but so far they do not add serious control - you can assume any percentage of cheaters up to 50% (accurate data on cheat can not be obtained, as long as there is no real control of markups) users who are curious enough to to google the answers, or they can afford a dad's assistant or a bot Petya in barter. There is no real counter-wrap-up method yet.
Question II . What value are Coursera Course Certificates? Can you trust them, since they are signed by Harvard or Stanford?
The answer is . In general, it is doubtful, since under the current MEP system it is very difficult to identify a conscientious apprentice or cheater with assistant Pope and nerd Petya in the certificate holder Vasya.
Question III . What to do?
The answer is . Try to deal with the cheating, in the next section we will try to suggest exactly how.
Consider the possible methods of dealing with the listed cheats.
Having studied and realized the problems listed above, we tried to implement some of the hypotheses proposed above in our Tutortalk online learning service: video control with warnings and a countdown timer on test questions. Creating answer generators for technical tasks, controlling keystrokes and blocking the switching of the / Alt-Tab tab from the lesson application is in the process of implementation.
Imagine conducting a traditional university offline course, consisting of lectures + seminars + exams + control on exams, tests or homework. On such a course, the teacher usually combines the functions of a teacher (actually transferring knowledge, giving lectures, analyzing tasks) and functions of an arbitrator (controlling students from cheating on a test or an exam, controlling the accumulation of coursework and dissertations, other ways of controlling students' knowledge), or transferring functions of an arbitrator assistants (they lead seminars for small groups of students, each are responsible for their own group). Any online course should strive for all the listed parameters to approach the offline course. So, if we supplement the MEP courses with webinars and exams with the ability to control and the functions of the arbitrator, with the right approach, we will get an almost equivalent online equivalent of the usual offline training without the above mentioned problems with cheating. The functionality of the majority of the MEP reflects only video lectures and tasks, while bypassing the lessons of webinars and student control, or vice versa - there are services that implement only webinars. We decided to combine these 2 directions in our project so that the win was for both formats - the students can listen to the usual video lectures at a convenient time for them, take post-lecture tests:
and then work out standard tasks on webinars with a teacher, take exams on the same webinars, but with the function of video control by the arbitrator and issuing fines for violations of fair play.
At the same time, of course, you will have to take into account the teacher’s labor costs on the webinar or the arbitrator who exercises video control, and the cost of their time (this cost will be small for the arbitrator assistant teacher), and also check the arbitrator for impartiality - this remains in the competence of the teacher, as and in offline courses. But the positive result is obvious - all students will be put on equal terms, the substitution of the student’s personality, cheating or quitting will be problematic, the restraint system will act as a warning and minus points from the arbitrator, and malicious violators can be expelled from the course or do not give the right to the certificate. It is also obvious that in this case it is necessary to keep a journal of visits, late pupils and teachers on such critical tests or webinars, because the time of the arbitrator or teacher has value and is scheduled according to the schedule - therefore, students should accept the fact that despite the fact that This is an online course, they will have to pass some critical tests only at certain times - if they want to get a valid certificate that is trustworthy. Of course, the student’s video control in class is unlikely to be adequate if he sits with a laptop or tablet in a subway car or in a noisy room, where he is constantly distracted - there are also some limitations that are familiar to offline courses, and any lessons are not calculated to conduct in such conditions, you can only listen to video lectures.
Thus, it is possible to significantly improve the MEP and credibility of the MEP-certificates of universities, and private teachers and tutors will be able to conduct full-fledged online courses for small groups, combining the functions of a teacher and a referee, as it happens on a regular offline course with exams.
We proceed from the fact that all students should be put in equal conditions, and also from the fact that anonymous online learning (which is currently allowed by MEP-services) with a teacher is ineffective for several reasons:
1) on condition of anonymity, it is impossible to resist the listed cheatings, to organize control, anonymous and at the same time protected from cheating rating is impossible.
2) training with a teacher is based on trust and personal contact, it is almost impossible to anonymously
3) it is assumed that both parties (the teacher and the student) have nothing to hide from what concerns their educational life
4) the majority of freelance services, payment systems and other services where it is important to control users - similarly, they offer an increase in user privileges in exchange for a decrease in his anonymity, this is a common practice
Other settings for the privacy of the service — for example, whether to publish your results of courses completed, or your students list — can be customized by users at their discretion. Whether to pass verification and use control, a system of warnings and fines - the final decision rests with the teacher, depending on his teaching methods.
Knowing about such a control system, the young programmer Vasya, the Pope and the botanist Petya, will think about it ten times before trying to look for ways to circumvent such a system. So, from pluses:
There are also relative disadvantages, which for some people will be pluses:
In the prototype of our Tutortalk online training service, we tried to eliminate most of the listed problems and implement the proposed anti-cheat techniques. In the near future, several new courses will appear on our service, which will be conducted using the student control and anti-wrap capabilities, as well as a redesign with improved UX. Our service has passed a preliminary selection for a grant from Start Fellows, where the main points of our vision for the future project are discussed in the discussion, we plan to further improve and test hypotheses as part of our service, and gradually enter the mass market of the MEP.
The analysis of all the problems listed in the article was carried out without being tied to a specific service; most of the popular MEP platforms (Coursera, Udacity, Edx) have these disadvantages. As a result, we come to the conclusion that in the absence of anti-cheat protection for the majority of the MEP, the percentage of abuse of students can be arbitrarily large, everything depends only on the curiosity and resourcefulness of students, there are no real limitations. Estimates of cheating on online courses, according to various sources, are in the interval from 30 to 40% of students [1] and significantly exceed the percentage written off on offline courses because of the actual lack of resistance to cheating. Certificates on online courses signed by Harvard or Stanford have an obvious value, in the long term close to the value of a regular diploma - therefore such certificates are of interest for many experts, and therefore for cheating or receiving them in dishonest ways. Therefore, for the MEP services, as well as for freelancing services or payment systems, which also need to carefully monitor their users and protect themselves from abuse - protection against cheating is simply necessary. At the same time, developing the MEP control system, as well as for any security system, it is important not to overreact and to maintain a balance between security and acceptance of the system by users, so that a simple online test does not turn into a kind of polygraph test.
Materials and articles on the control of the MEP:
1. Using technology to fight cheating in online education (LA Times)
2. Online Test-Takers (NY Times)
3. MOOC Teaches How To Cheat in Online Courses
UPD1: Appeal to the minus , who at least mastered the article: before minus, decide -
1) Do you generally a) for equal conditions for students of the course, or b) for the possibility of cheating (this is a “military trick”, everything is OK)? More directly - for equal competition of students or for corruption?
2) Do you understand, a) that equal conditions without control are impossible or you think b) that control is not necessary and certificates are not needed by anyone and will not be needed?
3) Are you a) for the evolution of the forms and methods of online education and the gradual replacement of offline education, or b) for keeping the MEP courses forever as vulnerable to cheats and optional as they are now?
The article justifies answers a), if you have 2-3 answers b) - we have nothing to discuss with you, and the article is not for you.
UPD2: If you have something to argue with reasonedly, it’s great, I suggest you write a detailed article-refutation or detailed commentary about the fact that no one needs control, that everyone is happy with everything, and that the MEP is already established and will not move anywhere. Only with weighty arguments, please, and links to sources. So that we believe you and give up to implement “useless” control.
But is such a structure enough to trust the MEP-certificates of universities? We tried to isolate the shortcomings of the MEP and supplement the existing system with methods to control and protect against cheating, and to test these hypotheses, we implemented our prototype of a web service for distance learning.
Consider the now generally accepted structure of online courses in terms of its security and protection from cheating and abuse.
Rating problems that are not protected from cheating
Take an example from a familiar IT area. Suppose you are looking for a Python-developer (or Java, C ++) in a team, with knowledge of machine learning, AI algorithms, and also consider sensible juniors without experience. There is no one candidate Vasya - there are several Coursera paper certificates, Udacity for specialized disciplines, supposedly “his” project with source codes on github, while there is no real programming experience in the team - but he seems to be answering your technical questions and promises to learn quickly . Maybe this will be the case, and it will make a good developer.
But, what if Coursera certificates and “their” project are ordered “to outsource” from an assistant through a freelance service at a reasonable price?
')
In total, we have - due to the lack of control over the online training services and the opportunity to complete the courses with an assistant, Vasya can greatly overstate his “external” rating and expectations from his work, and it’s not a fact that later he will be able to meet those expectations. Everybody loses from this: those who trust certificates - they will lose their time with Vasya, and Vasya himself - trying to deceive the service, instead of really learning and mastering technologies - he loses his time and yours. Perhaps, thanks to this type of thinking, Vasya will eventually become a good entrepreneur or a hacker (we hope white), but a good developer for whom he will try to pretend to himself will not immediately become. Given the high demand for any programmers - he is likely to be able to go through 1-2 bad HR filters. It is possible not to become attached only to IT, and apply this analogy to the majority of other areas where online courses are read and validated by the MEP certificates are issued that affect the user rating outside the online training service. Therefore, the unprotected rating and lack of control is unprofitable for everyone, including the online services themselves, as this reduces the credibility of the received certificates.
Let us take another example of an easily winding rating - some public ballots (for photos, for selecting music bands for the festival) take into account the simple likes of Facebook or VKontakte as votes. Everyone knows that “likes” can be easily screwed up, and you can google several services that will do it quickly and inexpensively. Thus, as a result of a winding up rating of participants, if there is no additional commission of experts in the voting, or there is no technology for eliminating markups and punishing them - the first prizes will be received by mediocre participants, and interesting newcomers without “like” support will not be able to get to the top. Voting for the quality of the product of the participants turns into a banal "clog up more likes." I think these two examples are enough to understand the value of the fight against cheating in any field. Consider the technology cheat existing online services MEP.
Technology cheating online learning service "course with an assistant"
So, Vasya decided to carefully prepare for the interview for your position - google the FAQ for the profile interview, read “Python (java, C ++) in 24 hours” or an analogue, put some money into the Chinese freelance assistant (let's call him Dad), who knows the necessary discipline and with whom he will take courses and receive certificates. I registered on Coursera or analog, found the necessary courses, ticked off "I will not write off", signed up for the necessary courses. Then Vasya passes the access to Coursera to the Pope, and the Pope begins to take all the courses for him. Within the framework of the existing system, it is impossible to take any measures against Vasya, since in no lecture or test his personality is confirmed, and Vasya can easily pass the course to the Pope, as well as homework. Dad is a good specialist, and passed all courses in good faith, deserved the issuance of a confirmed certificate for them. At this moment, Vasya returns, sends his personal data to confirm the identity - and a pack of certificates for the courses that Pope has completed in Vasya's pocket. Vasya received the necessary certificates, Papa - a good 50-100 $ for passing the courses by Chinese standards, both parties are satisfied. The probability to burn Vasya’s scheme with existing technologies of MEP-courses is 0%. To improve the plausibility, Vasya can give the Pope a proxy / VPN on his IP to go to Coursera, so a banal check on the IP - Vasya is also easy to manage. As a result of his actions, Vasya is ready for an interview with the conditional general “external” rating of 50, having only 10 in reality.
It is clear that an experienced recruiter paired with a technical specialist will be able to easily weed out Vasya at an interview or according to the results of a probationary period, but in some companies certificates play a very large role, and there they are ready to take newbies with almost no experience. The purpose of the example is not to blame Vasya, he just found a vulnerability in the MEP, but to show that the mere presence of an easily winding rating and certificates undermines trust in them, and in this area it is important to deal with cheating, as some value of these certificates is declared, in perspective, close to real diploma. Most of the services on online education do not do real anti-cheat protection, or they do it very poorly, for example - the tick “I observe fair play” on Coursera, which does not guarantee anything.
In addition to passing the course with an assistant, other methods of cheating are possible - consider them.
Other technologies of cheating and cheating
Suppose Vasya still felt remorse and decided to refuse the help of the Pope in principle, to do all the tests and tasks himself. Let us consider what opportunities it has for cheating the test results without the Pope, within the framework of the existing system of MEP courses:
- You can find a good botanist Petya fumbling in the course, unite with a small group of “like-minded people” around him, and he will do each “like-minded person” of his task or pass a test, barter or just like;
Estimated probability of detection by the MEP system: ~ 0.1% (if Peter himself accidentally burns himself, or the answers are too identical, and if it will be monitored somehow) - You can google the correct answers to the tasks already laid out by good Petya in public or on a private forum, and just drive them in, quickly getting a 100% result on the course;
Estimated probability of detection by the MEP system: ~ 1% (significant investments are required to track the publication of responses in all open / closed forums, promptly closing branches with published answers, in a nutshell - the task is comparable to the task “remove something from the Internet forever”) - you can find another, physical Daddy, who will also be strong in math / AI / machine learning, sit behind your back and be on your ear or on Skype to advise all the right answers to inexperienced, but daring Vasya;
Estimated probability of detection by the MEP system: ~ 0.1% (with the existing system, the MEP is close to zero, there is almost no monitoring on the tests) - you can google each specific question in the context of each test — in conditions where the response time to the test is unlimited and you can freely switch to another browser tab or application — you can google almost anything;
Estimated probability of detection by the MEP system: ~ 0% (the student’s attention is not being monitored only for the test)
There are quite ample opportunities, and Vasya is happy with this state of affairs, because it is possible, without the expenses of an assistant, to significantly wind up a rating in the course of the course.
After such a debriefing, questions arise:
Question I. If the online education services almost do not interfere with the above methods, then what percentage of their students will be able to wind up the results and get the certificates, repeating after Vasya?
The answer is . According to the LA Times, a survey of the students of the online service of the Online Journal of Distance Learning Administration back in 2010 revealed that 32.7% of students tried cheats at least once in an online course [1] . This is the percentage of only those who admitted this themselves, and in 2010 the MEP was not so common. In today's situation, when the MEP becomes really massive, but so far they do not add serious control - you can assume any percentage of cheaters up to 50% (accurate data on cheat can not be obtained, as long as there is no real control of markups) users who are curious enough to to google the answers, or they can afford a dad's assistant or a bot Petya in barter. There is no real counter-wrap-up method yet.
Question II . What value are Coursera Course Certificates? Can you trust them, since they are signed by Harvard or Stanford?
The answer is . In general, it is doubtful, since under the current MEP system it is very difficult to identify a conscientious apprentice or cheater with assistant Pope and nerd Petya in the certificate holder Vasya.
Question III . What to do?
The answer is . Try to deal with the cheating, in the next section we will try to suggest exactly how.
Methods of dealing with cheating (hypotheses)
Consider the possible methods of dealing with the listed cheats.
- Course with Assistant, Dad from China, transfer account to Dad
To counteract such a method - you need to control the student’s personality on each critical test, one email / pass / IP control is obviously not enough. Our suggestion is that the critical test should include students' video control (via webcam and microphone), and an arbitrator person should be present (for example, the teacher himself or his assistant), who can visually determine by the video that the student passes the test and also ask him oral questions if there are any doubts, or set a fair play violation warning during the test. - Botanist Peter and "like-minded"
By this method - the opposition is similar to the previous solution. - Google ready answers for the whole course
Possible solutions are to do task generators for tests, for example, for a technical task — if we have a problem whose solution is F = ma (determined by the formula), then by varying the parameters m in the test problem, we will get a lot of unique answers (but not according to the formula) variants of questions + answers. Hypothetically, you can vary in the response and parts of the formulas - for those tasks that suit you, or you can vary the testing conditions and input data for problems like "write a program on a task." For text tasks such as “select the year of the Kulikovo battle” or “select the article a, an, the” this is not suitable, but for technical problems with formulas it is quite. In this case, the publication of ready-made answers is complicated. - Physical Dad with tips in ril time
To combat this method, by analogy with 1) - suitable video control with the arbitrator, warnings. - Googling on the context of the question during the test
Video control is suitable, as well as the limitation of the response time for each test question. In this case, the student will have critically little time for googling (he will see the timer for 30 seconds, for example) - and he will have to focus only on the question or skip the question (the pass will be counted as no answer = wrong answer). It also fits the student's keystroke control (if he presses too much, it means that the answer may be googling, because in tests with options, one click of the mouse is enough to answer) and preventing switching to another application (Alt-Tab) or another browser tab from the application with a lesson.
Having studied and realized the problems listed above, we tried to implement some of the hypotheses proposed above in our Tutortalk online learning service: video control with warnings and a countdown timer on test questions. Creating answer generators for technical tasks, controlling keystrokes and blocking the switching of the / Alt-Tab tab from the lesson application is in the process of implementation.
Imagine conducting a traditional university offline course, consisting of lectures + seminars + exams + control on exams, tests or homework. On such a course, the teacher usually combines the functions of a teacher (actually transferring knowledge, giving lectures, analyzing tasks) and functions of an arbitrator (controlling students from cheating on a test or an exam, controlling the accumulation of coursework and dissertations, other ways of controlling students' knowledge), or transferring functions of an arbitrator assistants (they lead seminars for small groups of students, each are responsible for their own group). Any online course should strive for all the listed parameters to approach the offline course. So, if we supplement the MEP courses with webinars and exams with the ability to control and the functions of the arbitrator, with the right approach, we will get an almost equivalent online equivalent of the usual offline training without the above mentioned problems with cheating. The functionality of the majority of the MEP reflects only video lectures and tasks, while bypassing the lessons of webinars and student control, or vice versa - there are services that implement only webinars. We decided to combine these 2 directions in our project so that the win was for both formats - the students can listen to the usual video lectures at a convenient time for them, take post-lecture tests:
Interface player video lectures and tests in lectures
and then work out standard tasks on webinars with a teacher, take exams on the same webinars, but with the function of video control by the arbitrator and issuing fines for violations of fair play.
The interface of the test with the ability to assign penalties for violators
At the same time, of course, you will have to take into account the teacher’s labor costs on the webinar or the arbitrator who exercises video control, and the cost of their time (this cost will be small for the arbitrator assistant teacher), and also check the arbitrator for impartiality - this remains in the competence of the teacher, as and in offline courses. But the positive result is obvious - all students will be put on equal terms, the substitution of the student’s personality, cheating or quitting will be problematic, the restraint system will act as a warning and minus points from the arbitrator, and malicious violators can be expelled from the course or do not give the right to the certificate. It is also obvious that in this case it is necessary to keep a journal of visits, late pupils and teachers on such critical tests or webinars, because the time of the arbitrator or teacher has value and is scheduled according to the schedule - therefore, students should accept the fact that despite the fact that This is an online course, they will have to pass some critical tests only at certain times - if they want to get a valid certificate that is trustworthy. Of course, the student’s video control in class is unlikely to be adequate if he sits with a laptop or tablet in a subway car or in a noisy room, where he is constantly distracted - there are also some limitations that are familiar to offline courses, and any lessons are not calculated to conduct in such conditions, you can only listen to video lectures.
Thus, it is possible to significantly improve the MEP and credibility of the MEP-certificates of universities, and private teachers and tutors will be able to conduct full-fledged online courses for small groups, combining the functions of a teacher and a referee, as it happens on a regular offline course with exams.
Privacy
We proceed from the fact that all students should be put in equal conditions, and also from the fact that anonymous online learning (which is currently allowed by MEP-services) with a teacher is ineffective for several reasons:
1) on condition of anonymity, it is impossible to resist the listed cheatings, to organize control, anonymous and at the same time protected from cheating rating is impossible.
2) training with a teacher is based on trust and personal contact, it is almost impossible to anonymously
3) it is assumed that both parties (the teacher and the student) have nothing to hide from what concerns their educational life
4) the majority of freelance services, payment systems and other services where it is important to control users - similarly, they offer an increase in user privileges in exchange for a decrease in his anonymity, this is a common practice
Other settings for the privacy of the service — for example, whether to publish your results of courses completed, or your students list — can be customized by users at their discretion. Whether to pass verification and use control, a system of warnings and fines - the final decision rests with the teacher, depending on his teaching methods.
Conclusion
Knowing about such a control system, the young programmer Vasya, the Pope and the botanist Petya, will think about it ten times before trying to look for ways to circumvent such a system. So, from pluses:
- The listed schemes of cheating will become meaningless: Vasya will not spend his time to bypass the system (and will probably accept the new rules of the system’s game, or he will do something else), Dad will not try to “promote” Vasya with his clues, and Peter will probably something more interesting than “pulling out” your colleagues for like
- The recruiter will not waste his time on Vasya, eliminating him immediately on a set of certificates that are not credible
- The teacher will also benefit - he will get more control over the online learning process, and will be able to answer for his online students almost at the same level as for offline students, and if he wants, he can fully switch to online teaching with little or no loss of quality course, where also receive a new market students
- Universities participating in such a system will gain more confidence in their signed MEP certificates.
- A “teacher with a small selected group of students” training model that mutually arranges one another and enjoys mutual respect - it can be much more effective than a tough offline-drill, as well as the existing non-binding and vulnerable format of the MEP
- The overall rating from the service with the anti-wrap system obviously deserves more confidence than from the service without such a system.
- In conjunction with the proposed, you can introduce additional measures to strengthen the control of the personality of students and teachers - verification of educational documents, SMS-confirmation
There are also relative disadvantages, which for some people will be pluses:
- A teacher or an arbitrator will have to spend his time on student video monitoring on a schedule, but most private tutors and university professors are used to combine the functions of a teacher and an arbitrator, or to transfer the functions of an arbitrator to assistants.
- A more careful approach to the compilation of test problems or homework is required so that it is impossible to lay out ready-made solutions for them, but this is only a slight complication of the preparation of tasks, which in any case takes the lion's share of time in preparing an online course
In the prototype of our Tutortalk online training service, we tried to eliminate most of the listed problems and implement the proposed anti-cheat techniques. In the near future, several new courses will appear on our service, which will be conducted using the student control and anti-wrap capabilities, as well as a redesign with improved UX. Our service has passed a preliminary selection for a grant from Start Fellows, where the main points of our vision for the future project are discussed in the discussion, we plan to further improve and test hypotheses as part of our service, and gradually enter the mass market of the MEP.
The analysis of all the problems listed in the article was carried out without being tied to a specific service; most of the popular MEP platforms (Coursera, Udacity, Edx) have these disadvantages. As a result, we come to the conclusion that in the absence of anti-cheat protection for the majority of the MEP, the percentage of abuse of students can be arbitrarily large, everything depends only on the curiosity and resourcefulness of students, there are no real limitations. Estimates of cheating on online courses, according to various sources, are in the interval from 30 to 40% of students [1] and significantly exceed the percentage written off on offline courses because of the actual lack of resistance to cheating. Certificates on online courses signed by Harvard or Stanford have an obvious value, in the long term close to the value of a regular diploma - therefore such certificates are of interest for many experts, and therefore for cheating or receiving them in dishonest ways. Therefore, for the MEP services, as well as for freelancing services or payment systems, which also need to carefully monitor their users and protect themselves from abuse - protection against cheating is simply necessary. At the same time, developing the MEP control system, as well as for any security system, it is important not to overreact and to maintain a balance between security and acceptance of the system by users, so that a simple online test does not turn into a kind of polygraph test.
Materials and articles on the control of the MEP:
1. Using technology to fight cheating in online education (LA Times)
2. Online Test-Takers (NY Times)
3. MOOC Teaches How To Cheat in Online Courses
UPD1: Appeal to the minus , who at least mastered the article: before minus, decide -
1) Do you generally a) for equal conditions for students of the course, or b) for the possibility of cheating (this is a “military trick”, everything is OK)? More directly - for equal competition of students or for corruption?
2) Do you understand, a) that equal conditions without control are impossible or you think b) that control is not necessary and certificates are not needed by anyone and will not be needed?
3) Are you a) for the evolution of the forms and methods of online education and the gradual replacement of offline education, or b) for keeping the MEP courses forever as vulnerable to cheats and optional as they are now?
The article justifies answers a), if you have 2-3 answers b) - we have nothing to discuss with you, and the article is not for you.
UPD2: If you have something to argue with reasonedly, it’s great, I suggest you write a detailed article-refutation or detailed commentary about the fact that no one needs control, that everyone is happy with everything, and that the MEP is already established and will not move anywhere. Only with weighty arguments, please, and links to sources. So that we believe you and give up to implement “useless” control.
Source: https://habr.com/ru/post/187054/
All Articles