Exam cheating scandal in Son La: Difficult to restore true scores.

Awaiting the police investigation.

Irregularities were discovered in the grading process of the national high school graduation exam in Ha Giang (top row) and Son La (bottom row).

According to Mr. Dao Tuan Dat (lecturer at Hanoi University of Technology, Head of the Executive Board of Anhxtanh High School (Hanoi)), the irregularities in altering exam scores in Ha Giang and Son La are not the same.

Firstly, the multiple-choice test grading process involves four steps: scanning the test papers, transferring the image files to a computer to convert them into text files, and then feeding them into an automated grading machine to generate scores.

In Ha Giang, Mr. Vu Trong Luong made edits in step 2, which is the process of converting images to text. Therefore, when the issue was discovered, it was easy to compare the edited points with the original exam paper.

However, in the case of the irregularities in Son La, the person who altered the scores interfered and erased directly on the original exam paper, making it very difficult to determine which corrections were made by the student and which were made by an outsider.

Grading the 2018 National High School Graduation Exam in Hoa Binh. (Photo: My Ha).

"I think it's possible to spot signs of erasure with the naked eye when looking at the exam paper because with 50 questions, it would take a lot of time for one person to erase everything, so the erasure wouldn't be clean. Meanwhile, if the student erases, they will be more careful."

The difficulty lies in the fact that there is no basis to assert that the erasure was done by an outsider if this fraudulent act was not recorded by cameras.

The only option left is to rely on the investigative and handwriting analysis capabilities of the Ministry of Public Security, and if necessary, an investigation by the grading committee, as I believe they will have a list of candidates whose scores need to be altered and by how much. This is the quickest way, rather than having to re-examine the handwriting on thousands of multiple-choice test papers.

Mr. Vu Khac Ngoc, an education expert in Hanoi, also acknowledged that it is very difficult to determine the true scores from the altered exam papers in Son La because it is hard to identify which corrections were made by the students and which were made by someone else. "If we were to examine the handwriting on thousands of multiple-choice exam papers, I think it would be very difficult. Therefore, based on speculation, I think only some papers with very clear signs of alteration might have their true scores determined, while others will be difficult," Mr. Ngoc stated.

How can we fix the "loopholes" in test scoring?

Mr. Bui Viet Ha, Director of School@net, a school information technology company, stated that after understanding the loopholes leading to exam score fraud, he made several contributions to the process and program - the multiple-choice exam scoring software - of the Ministry of Education and Training.

It is very difficult to correct the altered scores on the original exam papers in Son La to reflect the true scores because it is hard to determine which corrections were made by the student and which were made by someone else. (Illustrative image: My Ha).

According to Mr. Ha, there are two obvious loopholes, the easiest to exploit, and two things that need to be changed.

Firstly, since the exam papers don't have unique identification numbers, the candidate's information (SBD) is always displayed on the screen throughout the scanning, recognition, and editing process. Therefore, this needs to be rectified immediately so that when the recognition phase begins before grading, the candidate's information is obscured from the screen. There are many ways to solve this.

Secondly, regarding the automated grading process: There are two loopholes here: firstly, the input data for the grading program is a text file, making it easily hacked and modified; and secondly, humans, who intentionally want to cheat, will find ways to do so. Allowing the grading committees to directly operate the final grading program is a major flaw.

Based on the above two points, Mr. Bui Viet Ha specifically suggested that the Ministry's grading process and grading methods should be changed.

We need an additional technique that allows for two stages of recognition after scanning the candidate's exam paper (sent to the Ministry of Education): recognizing the general information (the top part) and recognizing the exam paper itself.

When identifying the lower part of the answer sheet, the candidate's information needs to be erased, hidden, or obscured on the screen (and possibly the image as well). This adds an extra layer of security to the process.

Secondly, the final, direct machine-based grading should not be done at the local level or within examination boards. Instead, the entire final grading process should be transferred to the Ministry and processed centrally in the Ministry's computer lab. This way, the examination boards will only handle preliminary processing of the grading information.

According to Mr. Ha, the issue lies in how far the Councils will process the data before forwarding it to the Ministry. This could be done as follows: the Councils process the data until they produce the candidates' answer files. If these files are text files, they need to be encrypted immediately before being sent to the Ministry. This encryption should be automated within the program.

"Of course, nothing is absolute, but if we can do the things mentioned above, I think cheating will be much more difficult," Mr. Ha shared.