Entering edit mode
10.0 years ago
stat.1405
▴
30
There are two questions:
First: I have a column its dim is 1*10000, each row of it is the total differences between 20 indiviuals at a position, so the values in the column are 0,1,2,3,....20 I need to test the randomaizaition between this column and the positions. what type of test will works,
Second: I have also missing data ./. in my data which looks like this 0/0, 1/0 , ... 2/2, 3/0, .. It is a variants, Is the missing values randomly distrbuted or not?
Is this the follow up to your previous question? Please consider editing your previous question instead of creating a new one.
I deleted the previous one
Also, there is a misconception in your question, you cannot apply a test to establish significance for that the data is "random", because that is the null hypothesis, it is the outcome with the highest probability any way (imagine a coin-flip, you can establish significant deviation from uniform distribution by repeated experiments, but you cannot give a significance for the coin being fair). You can only test for significant deviation from the null-hypotheses. What "random" means depends of your model of the random process. Even the outcome of a manipulated coin is "random", just that the probabilities for each side are not equal.
therefore, unless you model your random process by some concrete random distribution (binomial, poisson, gaussian, ...) your question for "randomness" is void or always true: "yes, your data is random", you do not need a test to establish this, but this doesn't mean much.
Possibly you meant "independence" not "randomization"?
ok if i need to test is ind. what the test