Question

the equation for q test, outlier is q (adjacent number - q-suspect-outlier)/(q-with-most-distance - q-suspect-outlier). Use this equation we will get a number, and we need look at q test table with particular confidence, and to see is the number bigger than the number in table as accept or smaller as reject. By the way, how to get the number in the q table? There is only "confidence" and "N", how people get number there. Like 95%, 10 group population, The possibility is 0.466, how they get this number without know any standard deviation, mean and suspect outlier?

Answer 1

The Q-test is also known as Dixon's Q test. In the table of Dixon' Q test, columns are corresponding to level of significance (confidence e.g. 0.05, 0.01, 0.1 etc) and rows are corresponding to sample/population size N.

Now to lookup for the critical value for a given confidence and particular value of N, one has to look for the column of that 1-confidence value and row of that N value and where these column and row cross, that cell value will be the number at which one decide whether to reject or may not reject the hypothesis.

For example, if N=10 and confidence is 95%=0.95, then lookup for column of 0.05 and row of N=10, then that value will be 0.466.