I have confirmed that even in cases like this, the difficulty(b) and discrimination(a) can be obtained.
But how is it possible?
Actually, I am not fully understanding the logic of Bayesian IRT.
Therefore, if the history of solving a specific question does not exist sufficiently (for example, only one)
Using Bayesian IRT repeatedly, I thought the estimated difficulty and discrimination values would change significantly.
However, after repeating, it was confirmed that the values of difficulty and discrimination did not change significantly.
How is this possible?
I have confirmed that even in cases like this, the difficulty(b) and discrimination(a) can be obtained. But how is it possible? Actually, I am not fully understanding the logic of Bayesian IRT. Therefore, if the history of solving a specific question does not exist sufficiently (for example, only one) Using Bayesian IRT repeatedly, I thought the estimated difficulty and discrimination values would change significantly.
However, after repeating, it was confirmed that the values of difficulty and discrimination did not change significantly. How is this possible?