If you want to know more about testing, assessment and psychometrics, then the entire 20 minute video is well worth your time. Either way, I found this quote at 8:30:
We don't buy the causal link when we say "John grew 3 inches because he was in Mrs. Jones's class. Why should we believe that John gained 10 points because he was in Mrs. Jones's class?
To know if Mrs. Jones caused this, we would have to know the counter factual. That is, we would have to know how many points John would have gained in someone else's class. But we don't know that.
How can we, in all good conscience, make these inferences?