<aside> šļø Also see this article: How much does performance differ between people? See some interesting comments at the bottom of the post. Although this article is done by non-specialists (by which I mean they are not IO Psychologists or some similar type of scientist) and it is specific for the effective altruism community, I think it is still worth reading and considering.
</aside>
<aside> šļø I (Joseph) put this page together, but the text here is almost exactly copy-and-pasted from a textbook on Performance Management: Performance Management (4th edition), by Herman Aguinis. Iāve made a few small changes to the wording, and Iāve tweaked the formatting a bit, but in general you should NOT view this as my own work. You should view this as a lengthy quotation/excerpt of someone elseās work.
Any mis-spellings or type-os should be considered mine, rather than Herman Aguinisās. If you do spot any errors, please do leave a comment to point it out to me. I would appreciate your input!
I canāt do proper footnotes in Notion, so Iāll use red text to indicate the numbers of footnotes, and then list the sources at the bottom of this page.
This is taken from Chapter 5, under the heading 5-2-1-1 The Nature of the Performance Distribution, which starts on page 136.
</aside>
Using a forced distribution system implies that performance scores are forced to fit under a particular distribution shape. But in reality the performance distribution is usually not normal (i.e. , bell-shaped). This means that there is a minority of employees who perform at substantially higher levels than others. These star performers are critical for an organization's success.
If the distribution of performance is truly bell-shaped, then the majority of employees are grouped toward the center, with a small minority of employees who are very poor and very good performers. With this assumption, we ārationā the number of top performers. Thus in an organization of 100 employees, 15-16 of them would be rated as strong performers (more than one standard deviation above median), regardless of how many people are actually strong performers.
If we use a five-point scale, many companies might instruct managers that "no more than 10% of the direct reports gets the highest rating of 5." But what if this is an abnormally outstanding department that has 12% or 15% of the employees performing at a high-level (perhaps due to an excellent applicant pool, selective hiring, and offering training and development opportunities). Or conversely imagine that this department is abnormally weak, yet 10% of the employees must be ranked as 5 out of 5? Why should we force the highest ratings to the top 10% if our team is composed of 40% star performers?
Research has challenged the claim that performance is distributed along a bell curve. In fact, several studies based on more than 600,000 workers, including the number of publications authored by more than 25,000 researchers across more than 50 scientific fields, as well as productivity metrics collected from movie directors, writers, musicians, athletes, bank tellers, call center employees, grocery checkers, electrical fixture assemblers, and wirers have revealed that performance is distributed following a heavy-tail. 29
Figure 5-3 shows a critical difference between these two types of distributions. Under a heavy-tailed distribution, we expect to see many "star performers" (i.e., those far to the right of the mean). But under a normal distribution, the presence of such extreme scores is anomalous. Also, Figure 5-3 shows that performance differences between the top and average performers are much larger under a heavy-tailed compared to normal distribution.
The existence of performance distributions that are roughly normal is expected in certain situations, however. For example, workers in a manufacturing plant cannot work faster than the speed of the assembly line, so matter how good their hand-eye coordination, nor how fast their reflexes. So, situational constraints impede the emergence of star performers. But in the twenty-first-century economy dominated by the services industry, ceilings on performance are less common than they used to be in decades past. 30 Constraints such as geographic distances, lack of good communications, inability to access information and knowledge, and slow technological dispersion are lessened, given the spread of the internet and the increased flow of information. Consequently, normal distributions are the exception, rather than the rule.
An important consequence of "debunking the myth of the normal curve" is that we are now more aware of the existence of star performers (individuals whose contributions are much larger than the rest). 31 Star performers not only do well in terms of their individual performance, they also have a large positive influence on numerous key outcomes (survival of the organization, retention of clients, new product development, and many other indicators of organizational performance). 32