Thursday, November 24, 2011

Simple fix for EPR inflation

I am not sure it is as complicated as the dialog makes it seem.
You only need two things:

- Keep track of the "scores given" by each rater and associate an "average score" with that rater.
- Use a simple formula that amplifies the "score given" to the ratee by the "average score" of the rater.

Done!

The key is to incentivize raters to give more accurate scores - and to make it self-regulating. 
The incentive: the more frank and honest a rater's record is in giving feedback, the more ability the rater has to highlight top performers.   Essentially, "amplify" the rating received by the "average score" given by the rater. 
~ A side benefit would be to highlight raters who did not have the ability to make the tough calls (those with inflated averages).

Here is a chart of how the new numbering system would work  (capped at "0" and "100" for simplicity)



Examples:
EPR #1:
- Rater's "average score" on previous EPRs = 4.7
- Rater gives a "5" (like they apparently do to the majority of their troops)
- End result: ratee gets an overall score of "56" = slightly above average

EPR #2
- Rater's "average score" on previous EPRs = 3.2
- Rater gives a "4" (higher than what they give to the majority of their troops)
- End result: ratee gets an overall score of "75" = above average

EPR #3
- Rater's "average score" on previous EPRs = 3.2
- Rater gives a "5" (way higher than what they give to the majority of their troops)
- End result: ratee gets an overall score of "100" = WAY above average - top of the pack!

Once you create an equation you can add all the variables you like to address other issues - but the key here is it would be self regulating and WAY more effective than the current system.
It would forever STOP EPR inflation.

Here are some examples of other variables that could be added in:
- add a modifier for "top units" (plus 10 points)
- add a modifier for "key positions that only hire ‘elite’" (plus 10 points to 5% of your command's billets)
- add a variable for units who received outstanding unit award (plus 10 points)
- use the "average score" minus the "current year" (minimize a "timing game")
- record the “raters score” against their "average" after subtracting .2 from the score (thus allowing a 10% "top performer push” with no impact on their average for the year)

Perhaps I am missing something - something that could not be "quantified" - I can accept that.  Let us list those items, and start a discussion on the best way to optimally capture all variables.  Meanwhile, we can implement a system now that quantifies and rewards the raters who show ability to make the tough calls.  This would in turn help the AF promote our brightest in a time where we need the greatest fidelity when making tough calls in regards to our manning for the future.

P.S. 
- I hope to post an image of where these "new fields" would be on an EPR in the near future