Symptom
- Is it possible to enforce distribution guidelines set in Calibration Template?
- How is distribution guideline determined?
Environment
SAP SuccessFactors Calibration
Resolution
It is possible to set the distribution guidelines to be required where the session cannot be finalized until the guidelines are met.
Enforcement of distribution guidelines is supported in Bin View and Matrix View. Once the distribution guideline is set, the Bin View is where the user can see how many subjects he/she should remove or add for each bin. The Matrix view does not have any visual indicator for the guidelines. The Bin View should be the view to use to confirm guidelines have been met.
A warning message is triggered when requirements have not been met and the user is attempting to finalize the session. The warning message will indicate tab names to review.
How to Enable Guidelines Enforcement
Feature is enabled by going to Admin Center > Manage Calibration Templates > Advanced > General > Enable Guidelines Enforcement
The guideline values are entered at Admin Center > Manage Calibration Templates > Data. Next to the calibration element, there is a option to set Distribution Guidelines.
How is guidelines distribution determined and calculated if the total number of subjects cannot be evenly distributed?
We first determine which bin has the largest gap below its ideal distribution. Use the following example; assume the following distribution of 33.3%, 33.4% and 33.3% with 4 total subjects. The ideal target distribution should be (Total subject count * distribution percentage) which yields 4*0.333 = 1.332, 4*0.334 = 1.336, 4*0.333 = 1.332. Next, the system will sort the decimal portion of the ideal distribution
|
Low |
Medium |
High |
Target Distribution % |
33.3% |
33.4% |
33.3% |
Ideal distribution (4 subjects) |
1.332 |
1.336 |
1.332 |
Largest whole number of Ideal distribution |
1 |
1 |
1 |
Decimal portion of the ideal distribution |
0.332 |
0.336 |
0.332 |
Expected Distribution |
1 |
2 |
1 |
The algorithm to determine the expected distribution is sort the decimal portion of the ideal distribution. The system will distribute from the largest to the smallest until all subjects are accounted for. In the situation where the remainder are the same, it will distribution from the left to right (lower bin to higher bin).
The expected distribution for the above example, the system will expect 2 person to be in the Medium bin, while low and high will have 1 person.
Here is another example:
Assume 7 total subjects and distribution are 20%, 20%, 20%, 20% and 20%.
|
lowest |
low |
medium |
high |
highest |
Target Distribution |
20% |
20% |
20% |
20% |
20% |
Ideal Distribution (7 subjects) |
1.4 |
1.4 |
1.4 |
1.4 |
1.4 |
Largest whole number |
1 |
1 |
1 |
1 |
1 |
Decimal portion of ideal distribution |
0.4 |
0.4 |
0.4 |
0.4 |
0.4 |
Expected Number |
2 |
2 |
1 |
1 |
1 |
The decimal remainder portion of ideal distribution are all the same (0.4), the system will distribution from left to right (low to high). Therefore, lowest and low bin would be expected to have 2 person while all other bin will have 1.
See Also
- KBA 3045355 - Enhancement to Distribution Range and Group of Ratings in Guidelines for Performance Calibration Sessions - 1H 2021
- Calibration Implementation Guide - Distribution Guidelines
Keywords
calibration, guidelines, enforce, distribution, range, subjects, session, cal, admin, system , KBA , csg_q , LOD-SF-CAL-SES , Sessions, Creating, Managing etc. , How To