|
360 Degree Feedback Survey Response Scales
When a new 360 degree feedback survey is developed, one of
the tasks is to determine which response scale is most appropriate.
This Insight white paper describes some of the issues
and criteria that should be considered, and provides several
examples of possible scales.
Background and Criteria
Industrial-Organizational psychologists have researched the
pros and cons of alternative response scales for many years
(within and outside of 360 degree feedback applications), but
there is no consensus on the "one best scale." What
works best depends on the purpose of the survey and on the types
of items included. However, there is consensus on the basic
criteria for what constitutes a good response scale:
- The scale is simple and easy to understand.
- To the extent possible, raters have a common understanding
about what the scale means they interpret the scale
in a similar fashion.
- The words used to define the levels are clearly ordered
(e.g., it may not be clear which is higher "Excellent"
or "Outstanding").
- The scale is suitable for the types of items in the survey
(e.g., a scale where one of the levels is "Always"
may or may not be appropriate, depending of the types of
items).
- The scale leads to variance in ratings (competencies and
participants are not all rated the same there is
a good spread in ratings).
- The scale is aligned with the main purpose of feedback
(e.g., different scales would be appropriate when the purpose
is developmental versus evaluating performance against expectations/
requirements).
Three other general conclusions can be drawn
from the research:
- Increasing the number of rating levels within a 360 degree
feedback survey will, up to a point, lead to a greater spread
in ratings (which is desirable). In most situations, having
at least five levels is preferable. However, scales with
more than seven levels can lead to more error being introduced.
- The labels or anchor terms tend to confuse raters when
there are more than six levels, and each level is labeled.
- In many cases, regardless of the themes or number of
levels used, raters tend to inflate their 360 degree feedback
ratings, and there is less variance in ratings than one
would desire.
General Types of Response Scales
Listed below are the general types of response scales, based
on the main theme or aspect on which the ratings are made:
- Simple qualitative how would you rate the person
in this area? (e.g., from "Poor" to "Excellent").
- Extent to what extent does the statement describe
the person? (e.g., from "Very little extent" to
"Very great extent").
- Frequency how often does the person demonstrate
effective use of the behavior/skill? (e.g., from "Almost
never" to "Almost always").
- Developmental (strength/development need) how
would you rate the person's capabilities in this area? (e.g.,
from "Significant development need" to "Exemplary;
a role model").
- Comparison with others compared to others with
whom I have worked, this person is _____ in this area (e.g.,
from "Significantly below average" to "Best
I've ever worked with").
- Performance how would you rate his/her performance
in this area? (e.g., from "Does not meet expectations"
to "Far exceeds expectations"). Note that this
theme would make sense for performance appraisal purposes,
but probably wouldn't for a development-oriented 360 degree
feedback process, where peers and subordinates also provide
ratings.
It is possible to construct a response scale that has more
than one theme. For example, an anchor point of "Excellent;
best I've worked with" combines the qualitative and comparison
themes. However, unless this is carefully done, there is a
danger of the response scale becoming too complex, and thereby
confusing.
As a final comment on general types of response scales, it
is important that each level in a scale relate to the same
theme(s). It is not appropriate, for example, that some levels
of a scale relate to frequency, and other levels relate to
capability (developmental). This is a common failing of many
"home grown" response scales used in 360 degree
feedback initiatives.
Anchor Points
Response scales vary in terms of two main factors the
number of levels (5-point, 6-point, etc.), and the theme or
themes captured by the anchor terms and scale definitions. There
are obviously a very large number of possible scales. Normally,
each level in the scale has a descriptive anchor when there
are four or fewer levels, but all the levels do not necessarily
have to be anchored when there are five or more levels. The
example below shows Censeo's standard 360 degree feedback scale
for established surveys:
5 - Exceptional strength
4 -
3 - Competent
2 -
1 - Weak
The theme for the 5-point scale above is developmental, and
only three of the levels are anchored. It is highly questionable
whether adding anchors for the second and fourth levels, using
intermediate terms, would lead to more accurate and valid
ratings. 360 degree feedback raters implicitly form a mental
image of the scale and its meaning, and are able to easily
understand that a rating of 4 is halfway between "Competent"
and "Exceptional strength."
The response scale below is an example of a 5-point frequency
scale where only three points are anchored:
|
Never
|
|
Sometimes
|
|
Always
|
|
1
|
2
|
3
|
4
|
5
|
Getting Started
While choosing the best response scale (or developing a new
scale) for a 360 degree feedback initiative is an important
task, it is not something over which one should obsess. The
reality is that, regardless of the care that goes into scale
development, a good portion of raters will form their own mental
map, with the lowest value being "bad" and the highest
value being "good." This does not diminish the value
of the 360 degree feedback process, but does mean that the people
being rated should not interpret the scale too literally. Information
on relative scores (which competencies are highest/ lowest),
and normative comparisons (how do I stack up against others?),
should be the focus rather than the absolute values of the ratings.
When a new 360 degree feedback survey is being developed,
it's usually advisable to pilot a draft of the survey in a
focus group with typical raters. In addition to input solicited
on the 360 survey items themselves, key questions regarding
the response scale include:
- Do people interpret the scale the same way?
- Does it "feel right" to raters for the kinds
of items in the survey?
- Does the scale work well for all rater roles that will
be included in the 360 degree feedback process (managers,
peers, etc.)?
- Will the scale provide adequate discrimination versus
everyone being rated the same?
Appendix: Examples of Response Scales
| Qualitative Scales |
| 4 - Excellent |
5 - Outstanding |
6 - Extraordinary |
| 3 - Good |
4 - Very good |
5 - Superior |
| 2 - Fair |
3 - Good |
4 - Very good |
| 1 - Poor |
2 - Fair |
3 - Good |
| |
1 - Poor |
2 - Fair |
| |
|
1 - Poor |
| |
|
|
| Extent Scales |
| 4 - Exactly descriptive |
5 - Very great extent
|
6 - Completely true description |
| 3 - Very descriptive |
4 - Great extent |
5 - Largely true |
| 2 - Somewhat descriptive |
3 - Some extent
|
4 - Somewhat true
|
| 1 - Not descriptive |
2 - Little extent |
3 - Somewhat false |
| |
1 - Very little extent |
2 - Largely false |
| |
|
1 - Completely false description |
| |
|
|
| Frequency Scales |
| 4 - Almost always |
5 - Almost always |
6 - 100% of the time |
| 3 - Usually |
4 - Most of the time |
5 - 90+% of the time |
| 2 - Sometimes |
3 - Often |
4 - 80+% of the time |
| 1 - Seldom |
2 - Sometimes |
3 - 70+% of the time |
| |
1 - Seldom |
2 - 60+% of the time |
| |
|
1 - Less than 60% of the time |
| |
|
|
| Developmental Scales |
| 4 - Towering strength |
5 - Exemplary, best possible |
6 - No room for improvement |
| 3 - Strength |
4 - Real strength |
5 - Significant strength |
| 2 - Competent |
3 - Fully competent |
4 - Strength |
| 1 - Development need |
2 - Development need |
3 - Competent |
| |
1 - Weakness |
2 - Development need |
| |
|
1 - Significant development need |
| |
|
|
| Comparison Scales |
| 4 - One of the very best |
5 - Far above average |
6 - Top 5% |
| 3 - Better than most |
4 - Above average |
5 - 80-95th percentile |
| 2 - Better than some |
3 - Average |
4 - 50-80th percentile |
| 1 - Not as good as most |
2 - Below average |
3 - 20-50th percentile |
| |
1 - Far below average |
2 - 5-20th percentile |
| |
|
1 - Bottom 5% |
| |
|
|
| Performance Scales |
| 4 - Commendable |
5 - Far above requirements |
6 - Exceeds all standards |
| 3 - Exceeds |
4 - Above requirements |
5 - Exceeds most, meets others |
| 2 - Meets |
3 - Meets requirements |
4 - Meets most, exceeds others |
| 1 - Does not meet |
2 - Below requirements |
3 - Meets all |
| |
1 - Far below requirements |
2 - Meets most, below on some |
| |
|
1 - Below on many standards |
© 2007 Censeo Corporation
Return to Downloads page
Learn more about 360 degree feedback
|
Learn more about 360 degree feedbackReturn
to Downloads page
|