Employee Ranking Systems LO17250

Richard Goodale (fc45@dial.pipex.com)
Mon, 02 Mar 98 11:00:19 GMT

Replying to LO17182 (and in preparation for a reply to "Hosts Note"
to LO 17219)

Hi Doug

Sorry for the tardiness in this reply, but I've been in e-mail/LO limbo
for the past week or so due to petty skirmishes with the two most powerful
people on the Web, Generalissimo Gates and Major Domo. But that's another
story, another thread... However, because of those snafus, most of my
e-mail history is buried under impenetrable layers of MS Code (are these
guys really 100 times better than their peers?).. So, please my apologies
for any errors of memory or false attribution.

The quote you attribued to me in the above referenced post was actually
made by Eric Bohlman, in a reply to one of my posts on the relationship
between Ranking systems and Dentistry (for those of you who missed this
fascinating sub-thread, you had to be there...).

My apologies to Eric for being inadvertently tarred with my brush.

[Host's Note: My apologies as well... Mis-attributions happen easily
enough, I try to watch for them. ...Rick]

I read your story (about how the US Air Force in the 70's tried to
"force rank" officers by sub-organisation, regardless of the overall
"quality" of people assigned to each sub-organisation) with interest.
As I've mentioned in previous posts, I was involved in US Army's Officer
Personnel Directorate in the very early 70's, and know well of the system
of which you speak.

I very much agree with you that forcing some sort of bell curve over small
groups of people is idiotic, if that ranking is then transferred to some
larger pool, consisting of groups with varying missions and varying
capabilities. In my Army experience, this was not done. All ranking was
done over the larger pool, against established "pool-wide" standards. So,
someone who scored 238 in an OER in an elite unit was (at least
theoretcally) equivalent in ranking to one who scored 238 in a unit of

Let me add a few other points from my memory of this experience, because I
think it may be useful (at the very least to me) in thinking through the
broader issues of "Therefore, LOrger's, what?" as posed by our host in

The Army in those days did keep track of individual's career average
OER's, to the first decimal point (i.e. a scale of 0.0 to 240.0--although,
effectively, users "translated" it into a 5 point scale--scores below 235
were pretty rare). However, it used the forced ranking of these data ONLY
for purposes of "triage" (similarly to what Rol Fessenden has eloquently
described in many of his posts). In general terms: those below 230 got
booted out of the Army; those who conistently scored at or near 240 were
marked for early promotion and higher callings; those between 230-235 were
assessed to see if they could make some specialised contribution--if not
they were booted out; and those between 235 and 239.5 had a "normal"
career porgression (in fact, those at the higher end of that band tended
to get the better and more challenging assingments).

These "general" terms, however, oversimplify the situation. All key
personnel decisions incorporated a number of other factors than the forced
ranking of the OER's. Even in the most hopeless of cases, terminations
required a complete file review by senoir officers. The ranking for the
important career decisions (e.g. batallion commands, War College
selection, promotion to LTC and Colonel) involved a system in which OER
averages were only ~2/3 of the total "score." This "score" include other
factors such as education, medals won, and quality of assignments.
Finally, even after a preliminary "ranking" was made of the elegible pool
for any major action, using the above criteria, ALL candidate files were
reviewed in depth by a panel of senior officers. I know all this because,
for two years, for one branch of the Army, I was in charge of making up
these preliminary rankings.

What I learned from this experience is that "forced ranking," when made
against established and organisation-sensitive criteria, and supported and
tempered by human judgement, can be a very useful and humanitarian way of
helping some of us imperfect humans accomplish the very difficult task of
making decisions which will affect the future of some of our fellow

If we believe that one of the most important raison d'etres for having
leaders in the first place is that hard decisions need to be made, ALWAYS
with imperfect information, we should support (and always strive to
"perfect") any tool, like "ranking systems" that can help such decisions
to be made with intelligence and compassion. It IS "lonely at the top,"
but that's why real leaders get paid the big bucks.


Richard Goodale


Richard Goodale <fc45@dial.pipex.com>

Learning-org -- Hosted by Rick Karash <rkarash@karash.com> Public Dialog on Learning Organizations -- <http://www.learning-org.com>