How to calculate test reliability with different exam versions

Hello All,

I'm new to this forum and this is my first questions.

We have 1500 students and each student will receive a test that has 30 question.

We have a test bank that has 763 questions. The test for each student will be different (i.e. different version with different 30 questions). The 30 questions are randomly pulled from this test bank.

How can we calculate the reliability of the exam when each student is practically solving a different version of the exam?

I am referring to this guide which explains how to compute reliability:

The data of the exam I'm talking about can be found here:

Your help is much appreciated.