The witness, Dr. D. Laikin: You can't look, you can't relate to statistical differences if you don't speak or don't take such tests.
Adv. Dr. Tal Rotman: So it is not possible to draw a statistical inference here from Table No. 2.
The witness, Dr. D. Laikin: Yes, but we are talking about descriptive data, in numbers they are different, there is a very, very large sample here, but from the data that is presented it is not possible to relate..
Adv. Dr. Tal Rotman: It is true that descriptive data in general is not possible.
The witness, Dr. D. Laikin: I pointed out, it's impossible. But you can get a pattern.
(pp. 1314-1315).
And later on -
Adv. Dr. Tal Rotman: Is the answer, for your information, or Do you have any information, is the answer to any such question, its internal distribution was a normal?
The witness, Dr. D. Laikin: Oh, of each of the statements?
Adv. Dr. Tal Rotman: Yes.
The witness, Dr. D. Laikin: I don't know. I didn't check.
Adv. Dr. Tal Rotman: You didn't check.
The witness, Dr. D. Laikin: Nope.
Adv. Dr. Tal Rotman: Are you aware that in the test of Manova A condition for running the test is that the distribution in each of the sections of the answers will be approximately normal?
The witness, Dr. D. Laikin: I know, I also know that many researchers use parametric tests for pragmatic reasons, even when the data do not always meet the assumption of the normal distribution.
Adv. Dr. Tal Rotman: I suggest that a distribution, where there are only 4 possibilities, can never be a normal distribution, right??
The witness, Dr. D. Laikin: Because of the... In the end you have a result, you end up weighing it, making it almost always a score.
Adv. Dr. Tal Rotman: No, it's not a distribution, the normal distribution is how many people are in a histogram of not at all, how many are in the histogram of occasionally, how often and how much all the time, and you expect that in a normal distribution there will be a Gaussian bell, right?