What is a percentile value?
A percentile ranking compares one value against a distribution of values, by breaking the distribution out into 100 buckets. A percentile value of 60% implies that the score is better than that of 60/100 respondents.
How does the 360 assessment report compute percentiles?
The percentiles shown in the 360 report are calculated as compared to the full set of people who have completed the 360 assessment. These scores do not consider self-assessments.
Why are percentiles useful in this case?
A primary case for using percentiles is when the raw value is not particularly useful. This is true in this case, for a couple of reasons, at least:
First, average values and variance make raw scores hard to interpret. E.g. responses often average around "agree", and answers of "strongly disagree" are rare.
Second, these values vary significantly between questions and sections. The same value for different questions might imply different things, since the average response and variance are different.
By using a percentile, one can adjust for these factors and give a more relevant, useful measure.
How should one interpret and apply Percentile scores?
The primary intention is that the score from a 360 helps one learn more about themselves, and better enables setting specific and valuable goals. The relative percentile measure allows one to better gauge what scores are strong overall, and how scores on one section relate to scores in another.
In other words, use these values to help set a direction to work on and to understand overall areas of strength and weakness.
How should one NOT interpret and apply Percentile scores?
We discourage using the 360 scores in order to precisely evaluate one person versus another.
First, this is contrary to the spirit of the Torch service, which is focused on generating development, not performance evaluation. Too much emphasis on evaluation (by any role involved) can often interfere with the quality of growth that results.
Second, the assessment scores are not designed to support such usage. Applying them in this way is likely to result in inaccuracy, due to a host of factors, including variance among organizations and feedback providers.