The undeniable fact that the valedictorian of the thinking assessment (TSA) examination section 2 in 2026 of Hanoi University of Science and Technology achieved 98.98/100 points regardless of being flawed in 9 questions is attracting consideration from candidates and oldsters.
Before these considerations, Assoc. Prof. Dr. Vu Duy Hai – Head of Admissions – Career Guidance Department, Hanoi University of Science and Technology – shared with the press to make clear the scoring mechanism in line with the trendy training measurement mannequin, as an alternative of the conventional calculation technique.
Total scores usually are not solely primarily based on the complete variety of appropriate solutions
According to Assoc. Prof. Dr. Vu Duy Hai, for a few years, exams in Vietnam have used a tough scoring technique, that means every appropriate query is calculated with the similar rating and added to get the closing outcome.
This strategy is straightforward however reveals limitations when it doesn’t distinguish the degree of competence between candidates with the similar variety of appropriate questions, particularly in circumstances the place the questions are completely different in problem.
In truth, two candidates who achieved the similar rating could have appropriately answered completely different teams of questions, with equal problem ranges. At that point, scores don’t precisely replicate competence, and at the similar time trigger difficulties in evaluating outcomes between examination questions or examination phases with completely different ranges.
To overcome this, Hanoi University of Science and Technology applies the IRT (Item Response Theory) mannequin, a contemporary academic measurement technique used in many worldwide exams.
The TSA examination makes use of a 2-parameter IRT mannequin, together with the problem and differentiation of the questions. Accordingly, every query in the examination is assigned its personal parameters, reflecting the problem degree and talent to categorise candidates. At the similar time, every candidate can be estimated a sure degree of competence via the total test.
Therefore, candidates who do many tough questions appropriately could have a bonus in points. While candidates who solely do simple questions appropriately could have a decrease rating.
4-step IRT rating course of with 2 parameters in TSA
The scoring course of in line with the 2-parameter IRT mannequin is carried out in 4 steps.
Step 1: Based on the candidate’s test knowledge, the grading software program will estimate the problem and differentiation of every query in the present examination.
Step 2: The grading software program estimates the thinking capacity rating of every candidate primarily based on the test outcomes and the scale of the present examination.
Step 3: Convert the scale of the present examination to a standard reference scale with earlier exams to convey the candidate’s thinking capacity rating to the similar scale.
Step 4: Convert the thinking capacity rating of the referenced candidate to the TSA scale.
Assoc. Prof. Dr. Vu Duy Hai mentioned that grading in line with IRT helps to evaluate extra precisely and pretty, when the rating is just not solely primarily based on the variety of appropriate questions but in addition the results of the technique of assessing competence primarily based on the total work.
At the similar time, the scale of the examination questions will probably be transformed to a standard reference scale between examination durations to make sure equivalence and equity.
The software of this measurement mannequin is taken into account a step ahead in enrollment innovation, approaching worldwide exams, however candidates want to know clearly to keep away from confusion when evaluating scores.