Damping and Capping Mechanisms

QS utilize a damping mechanism to spread large inter-annual swings in data. This has the effect both of minimizing the impact on rankings of these (possibly temporary) swings in data, and correcting for errors of interpretation in data submission.

To illustrate how and when the damping mechanism is applied, the following concepts are crucial to understand.

Maximum Accepted Variation: the maximum increase or decrease in an underlying value (score, index or ratio) we will accept without adjustments. It is the underlying scores that are damped in Academic Reputation and Employer Reputation as we use normalized scores to calculate Academic and Employer reputation results in the rankings analysis, whereas for other indicators we use underlying ratios or indices due to the nature of those indicators (for example, Faculty Student Ratio or IRN Index).

Damping Threshold: the damping mechanism is only applied when the underlying values are above this threshold. Any underlying value below will not be affected.

Recovery Increment: the rate QS accepts when the value exceeds the maximum accepted variation.

Damping is applied every time when the current value (submitted by an institution or received from a third party this year) exceeds the previous value (used in the calculations last year, i.e. damped value) by more than the maximum accepted variation.

Reputation Damping

Let's consider example values from the table below. For Academic Reputation, we accept in full a maximum year to year increase of 10 points in the underlying scores. In this case, the recovery increment is 20%, which means we will accept 20% of the excess over the maximum accepted variation. However, if the score has not exceeded the damping threshold of 15, no damping would be applied (even if the year to year change exceeds 10 points). A similar approach is used for Employer Reputation, however, the maximum accepted variation will be 3 points, the damping threshold 4 and the recovery increment 20%.

International Faculty/ International Student Damping

For International Faculty Ratio and International Student Ratios, if the variation of the ratio against last year is over 3% while the ratios of both years exceed 5% then the damping mechanism will be applied. In this case, any variation exceeding 3% will be accepted at a recovery rate of 33%. For example, if the International Faculty Ratio of an institution increased from 5% to 9% this year, the variation would be is 9%-5%=4%, which is higher than 3%. What we use will be 5%+3%+0.33*(4%-3%)) = 8.33%. However, if the International Faculty Ratio is below 5% then we will accept the original value. It should be noted that the International Faculty Ratio is capped at 50%.

Having said that, the maximum ratio increase we could accept will be 50% (Cap) so any values beyond that will be discarded.

Faculty Student Ratio / Citations per Faculty Ratio Damping

Damping of the Faculty Student Ratio and the Citations per Faculty Ratio is similar to the damping of the International Faculty Ratio and International Student Ratio. The main difference is in the way we calculate the variations in ratios. For International Faculty and International Students Ratios, this is calculated by a simple subtraction between the current and previous years’ ratios. The variation of Faculty-Student Ratio and Citations per Faculty Ratio is calculated from the percentage growth/decay from last year.

Taking Faculty Student Ratio as an example, if the increased percentage against the previous year is over 10% while the ratio exceeds 4% then the damping mechanism will be applied. In this case, any variation exceeding 10% will be accepted at a recovery rate of 20%. However, if the Faculty-Student Ratio is below 4% then we will accept the original value. It should be noted that the Faculty-Student Ratio is capped at 30%. Having said that, the maximum ratio increase we could accept will be 30% (Cap) so any values beyond that will be discarded.

In the case of the Citations per Faculty Ratio, we allow a maximum year-to-year change of 20% (plus a reference value depending on the average change level year on year). For any excess, we only consider 33%. For example, let’s assume the reference value equals 0. If a ratio changes from 1.0 to 1.3 in a given year, we would only accept the full change to 1.2 and 33% of the remaining 0.1. So, the hypothetical ratio used in a case like this would be 1.233 (and not the original 1.3).

The following table provides a summary of the damping parameters we used in the QS World University Rankings 2020: