Jump to content
Beckman Coulter User Forums

Kaluza 1.3 Geometric Mean and HPCV calulations


Recommended Posts

Please see information requested from a K 1.3 user below which asks for the differences between V1.2 and V1.3 regarding the calculation of the Geometric mean and the HPCV statistics.

 

"I’ve compared the two help files side by side, but cannot work out what differences there are in the actual calculation of geometric mean and HPCV statistics. There is a change in the actual formula syntax required to display statistics in an information table, but this is not restricted to the two mentioned above, nor will it directly affect their calculation. Is it possible to get more information on what the specific calculation updates were for the Geo mean and HPCV?"

 

With thanks

 

Maree

Link to comment
Share on other sites

Maree,

 

The changes made to the geometric mean and HPCV calculations were relatively minor and in many cases there will be no visible difference. Because Kaluza's testing requires exact statistics matches between versions, it was important to publish that changes had been made which cause minor differences.

 

For HPCV the change was related to the number of bins assigned to the half peak. This requires some explanation: HPCV is calculated by creating the frequency distribution histogram for the designed population, locating the highest point (bin with the largest value), dividing that size by two in order to determine the half-peak, and then seeking in both directions until the highest and lowest bins on either of the peak which represent the half peak levels are found. In version 1.2 the way the resulting number of bins were calculated sometimes resulted in excluding a bin that should not have been excluded. Version 1.3 corrects this problem. For many populations the difference between the approaches may be zero or too small to see at the default statistics precision.

 

The geometric mean change was broader in scope but the change in results is usually small. In version 1.2 geometric mean is calculated using the frequency distribution histogram. In version 1.3 it is calculated using individual values. In practice means that values will generally be similar to what they were in v1.2, but you'll find that for populations closer to zero they deviate to a greater degree. This change results in geometric means that track more closely to the results produced by reference sources.

 

Hope that helps,

 

Ernie

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...