Military Review English Edition July-August 2015 | Page 66

from command billets. With the Army’s job rotation
system of development, using multirater feedback in
annual evaluations would require significantly more
frequent feedback sessions, overloading our already
busy superiors, peers, and subordinates with another
survey. Additionally, a single multirater feedback
session does not enable behavior improvement in
and of itself, but it requires multiple sessions over
time to demonstrate real change, a process that eventually uncovers a lack of development or toxicity.40
Now that we understand multirater feedback’s
strengths and weaknesses, we can look at how to
integrate it with performance evaluations and
CSL boards, as noted by Gen. Dempsey and Gen.
Odierno.41 Odierno and other senior leaders acknowledge the weaknesses in the Army’s MultiSource Assessment Feedback (MSAF) system, which
includes coaches and which requires an assessment
to be started for completion of an OER.42 The current MSAF system is for developmental purposes
only and offers optional coaching from a third party
outside the chain of command—all supported by
research as appropriate use and likely to enhance
validity and reliability. However, the focus of this paper is not to evaluate the current MSAF system, but
to illuminate issues and conditions that the Army
should take into account if implementing multirater
feedback into evaluations, promotion boards, CSL
boards, or other performance-based assessments.
Where these instruments may work in the civilian
business world, the Army is different in both its binding regulations and culture. Commissioned officers,
warrant officers, noncommissioned officers, and other
enlisted soldiers are all subject to the Uniform Code of
Military Justice, which states that soldiers must follow
lawful orders. “You do not have to like it—just do it,” is
an oft-quoted phrase. The Uniform Code of Military
Justice regulates the military profession to ensure good
order and discipline; this is an artifact of coercion and
authority within the strong culture of the military.43
This hierarchical and authoritative structure drives a
strong culture of following orders, even if the subordinate does not agree with or like the order.
Also within the Army culture, as espoused at
Army leadership schools, is the idea of “owning”
orders that come down the chain of command,
which means to avoid shifting blame to superiors
64

for orders subordinates may not like. This results
in subordinates not always fully understanding the
background of an unpopular order, which can skew
how subordinates view a leader. If a leader follows
this principle, then subordinates may never know
how much the leader corrects, fights, or accepts
unpopular taskings or orders. This knowledge could
change how subordinates view their supervisor and
affect subordinate inputs into multirater feedback.44
Furthermore, even though the Army attempts
to prepare leaders for future jobs with increased
responsibility, most do not comprehend the jobs of
their supervisors fully until they become supervisors. Bailey and Fletcher found that subordinates
and peers may have insufficient experience to rate
effectively, thereby reducing the accuracy and reliability of any performance rating or developmental
feedback.45 Multirater feedback has more consistency, according to Theron and Roodt, in routine
and well-defined jobs.46 However, the Army expects
innovative and adaptive leaders to perform many
complex tasks. Army leaders must solve complex
problems rapidly and perform tasks that are not
routine or well defined—such as individual counseling, unit evaluations, family support, media engagement, combat functions, and many more. Rating
these tasks is unlikely to lead to reliable and valid
data. Using these data in a performance evaluation
or board selection would not only fail to solve the
toxic leader problem but also would damage the
reports so that the board would be promoting or
selecting the wrong leaders based on faulty data.
The most glaring issue with using any multirater
feedback data for a performance evaluation or selection board is that the multirater feedback instrument
cannot predict performance, and it cannot guarantee
that an increase in se