CHE University Ranking
If the student survey has a low number of cases, even stricter standards are set for the acceptance in the Ranking and assignment to the middle group.
In order to continue to prevent the small departments from being systematically excluded from the survey, we decided, after intensive discussion, to stand by the minimum number of 15 opinions. However, we did introduce a number of methodological changes so that we can take borderline cases with low response rates out of the ranking: For departments with participation rates of between 15 and 30 students, an additional criterion was introduced for the assessment which specifies that the response rate must be no lower than 10%.
Even stricter standards have now been set for assigning the values in the middle group. As before, the individual indicators are also taken into consideration for assignment to the Top, Middle or Last group, thereby taking account of the deviation from the nationwide mean and the confidence interval. A new development is that two additional borders have been drawn around the nationwide mean. If the confidence interval is so broad that it extends beyond these borders, these values will not be included in the ranking, because it cannot be clearly assigned to a specific group. This applies particularly to departments with response rates of between 15 and 30.
(a) Small departments still have a chance to reach the ranking;
(b) The validity of the group's assignment is still assured, even in cases with low numbers of student opinions, borderline cases are removed from the ranking by the relevant indicator.
Criteria with thoroughly positive or hardly differentiated student opinions are not ranked.
The CHE University Ranking stands out in that as many valid indicators as possible can be drawn on for the ranking. Over the years, numerous departments have improved conditions for their students, also as a response to the assessments in the ranking. In some respects, it can be seen that Germany performs well throughout, and in some cases the students hardly give any poor assessments at all. For the CHE University Ranking, this means that no artificial differences are shown where non exist. This is why indicators that are assessed as good or better throughout are no longer included in the ranking. To ensure that the information on this continues to be maintained, the values without group assignment continue to be recorded in the web version of the ranking. Similarly, no ranking group categories are made any more when the difference between the best and the worst value is less than 1.
Advantage: Ranking groups are only formed when clear differences exist; this makes it possible to adequately depict the differences without over-exaggerating them.