The BIX results are calculated as follows:
All existing plausible data is used in the first step. This includes data for ca. 2.000 public and ca. 240 academic libraries. Data is available for 65 to 98% of those libraries. BIX for academic libraries uses four indicators that can not be computed from DBS data. The maximal peer group for these indicators are BIX-participants.
Libraries are separated into their peer groups. All further steps refer to the peer groups.
The percentile is calculated for each indicator of each library.
The percentile describes the position of one library in relation to all libraries in its peer group. It is a number between 0 and 1. The library with the lowest value has the percentile "0". The library with the highest value has the percentile "1". All other libraries are in between these two. The library with a percentile of 0,5 is exactly in the middle of its peer group: 50% of all libraries have a lower value, 50% have a higher value. The 0,85 percentile means: 85% percent of all peer libraries have a lower value, 15% have a higher value.
The BIX-profile shows the percentiles of all indicators of the library as a diagramm.
As a rule, BIX rates indicator values as "the higher the better". There are some exceptions, however, in the target dimension efficiency (e.g. costs per visit). These indicators are rated in reverse order (lowest value is best value).
All percentile values of the indicators in one dimension are summed up and divided by the number of indicators in this dimension. The result is again divided into percentiles. They determine the final rating result of the library:
Due to this method, a single top- or bottom value cannot have an undue influence on the overall results of the library.
The rating group of the four dimensions determines the number of BIX-Tops ("stars") that the library gets.