The actual implementation that I'm interested in is a set of five scores from 1-5 (whole numbers) that are averaged together to compute an overall score (decimal values). However, I prefer to look at the simpler case of a single score from 1.0 to 5.0. The complete evaluation criteria is known to everyone and well defined.
Let's say that I'm asking a group of people to rate themselves (and, perhaps in the future, others) on a scale from 1-5, with 1 being the low end of the scale. Over time, I want people to change their behavior to some minimum standard, which is represented by some point on this scale (improve their score). I envision a "minimum standard" score that everyone is expected to achieve (or show measurable progress toward achieving) in the near future.
I see two options in formulating the range of values:
- The high end (a score of 4-5) is the desired state. A score of around 3 or 3.5 would probably be considered to be acceptable. Anything lower than a 2 may be considered unacceptable.
- The desired goal is in the middle (about 3). The low end of the scale (1) is unacceptable, the high end (5) of the scale is extremely difficult to obtain.
I'm leaning toward the first implementation, with the thought that people inherently want to make the numbers reach the high end of the scale. If people look at the high end and see that it's incredibly difficult to achieve, they will lose interest in measuring themselves and tracking progress. Therefore, a 5 should be hard, but not overly difficult, to achieve, with most people eventually working from a present state of 2-3 to a final state of 4-4.25.
I believe that this might be related in some way to gamification, but there's no reward system. It's simply expected behavior that one would rate yourself and embedded in the culture that you strive to demonstrate improvement in some way (in this particular case, by rating yourself and others on a well-defined scale).
Is there any research that can support or disprove my initial thoughts?