the answer to this question is: cluster sample
In cluster sample, we divide the test subjects into two groups and impose the new standarized test into each group.
If, the standarized test provide similar results between the two groups, we can conclude that the data from the test is somewhat valid