Inter-Ranking Consistency Index (IRCI)

The problem IRCI solves

Two institutions can hold the same average standing for very different reasons. One may be placed in a similar percentile band by all three systems. The other may be placed very highly by one system and much lower by the others, producing the same average from far weaker agreement.

Averages alone cannot distinguish these cases, yet the distinction matters: convergent placement by independent methodologies is much stronger evidence than a single outlying result offset by weaker ones.

How consistency is measured

After each published result is normalised into a percentile, IRCI is derived from the mean absolute deviation between those percentiles. Small deviation means the systems agree and consistency is high; large deviation means they disagree and consistency is low.

Because it operates on normalised percentiles, IRCI is unaffected by differences in table size or banding conventions between publishers.

Why it is deliberately small

IRCI carries three percentage points of the composite — enough to distinguish well-evidenced scores from thinly evidenced ones, but not enough to let a reliability adjustment outweigh measured performance.

This reflects a general principle in the GURI framework: reliability indicators should qualify a result, not replace it. Ranking Achievement remains the dominant component at 72%.

Reading low consistency correctly

Low consistency is not a defect. It usually describes a genuine profile: an institution with outstanding research output but a modest international profile will be placed very differently by bibliometric and reputation-weighted systems, and low IRCI is the correct signal that its strengths are concentrated rather than broad.

GURIndex therefore reports IRCI on institution profiles and in ranking tables rather than folding it silently into a single number, so the disagreement between sources stays visible.

Frequently asked questions