i am given a sequence of 20 random numbers.
.05,.15,.23,.18,.06,.27,.31,.36,.32,.45,.55,.56,.51,.67,.62,.83,.96,.90,.70,.77.
All in the interval [0,1) And now i am supposed to check, if those numbers satisfy the chi-square goodness of fit test for uniformly distributed random numbers. The significance level alpha = 0.05.
If i just use the formula, do not seperate those numbers into intervals, i get a result of 3,2892 which is way smaller, than the 16,92 i get out of the table.
But ... am i not supposed to check intervals? like on a sample size of 100, or even 500, with expected numbers and so on? Is this task just a cheap trick and can not be done?
If it can be done, how? I am quite lost on that one.