Professional Documents
Culture Documents
Overview
Denison Organizational Culture Survey results are presented in the form of percentiles. The purpose of
this document is to describe what percentiles are, why they are used, and how they are used by Denison
Consulting to help make culture survey results more understandable.
What are percentiles?
Percentiles show how an organizations culture ratings compare to ratings of other organizations. They
indicate the percent of organizations that scored below a certain value. For instance, if your
organizations Core Values index average is at the 57th percentile, it means that 57% of organizations
in the benchmark database had lower index averages than your company.
Why use percentiles?
The most difficult parts of conducting any type of evaluation are determining how well someone or
something actually performed and then communicating that information in a meaningful way. For
instance, suppose that Heather, a fourth-grader, knows 18 out of 30 words on a vocabulary test, or that
Steve, a high school senior, runs a mile in 7 minutes and 40 seconds. Most people would not find these
numbers to be very meaningful. Is knowing 18 out of 30 words an indication that Heather is falling
behind in vocabulary, or is she right where a fourth-grader should be? Is 7 minutes and 40 seconds a
poor time for a senior running a mile or should Steve consider running track in college? A logical
question to ask about the performance of Heather or Steve is Is that good or bad?
Both of these examples use what is known as a raw score. A raw score is the unadjusted value that has
been measured, whether it is a fourth-graders score on a vocabulary test, a high school seniors time in
the mile run, an infants weight, a retail outlets sales revenue, or an IT consulting firms average survey
score on a five-point scale on the Core Values index. By themselves, raw scores may not be very
useful or informative for anyone other than a knowledgeable specialist. In short, raw scores often
cannot answer the question, Is that good or bad?
Percentiles to the rescue!
One way of addressing the Is that score good or bad? problem is to compare the score for one
organization with the scores attained by other organizations. This is known as norm-referencing, and
it is best known for its use in educational testing. Most parents with school-age children have probably
seen achievement testing results reported in percentile terms, particularly if their children are old enough
to have taken the PSAT, SAT, ACT, or any of a number of different state-administered achievement tests
given in the U.S. For instance, if Angie scores at the 82nd percentile on the Math Concepts section of
the Iowa Tests of Basic Skills this means that she scored higher than 82% of same-grade students in a
normative sample that took the same Math test. For Angies parents, this is a useful thing to know.
Relative to other fourth-graders, Angies progress in mathematics is extremely good. Unless they were
highly knowledgeable about the Math Concepts scale of the Iowa Tests, however, being told that
Angie had a raw score of 19 would not tell them very much about her progress.
80th
Percentile
Percentile
30%
5th
30%
95th
Percentile
Percentile
15%
2.62
15%
2.84
3.06
3.31
Frequency of Means
Percentile
3.56
After the normalization procedure was performed, each raw score mean was associated with a particular
percentile. Using this information, a new raw mean, such as an organizations average on the Vision
index, can be converted into a percentile. At this point, the normative database was ready for use with
new survey projects.
Company A
Company B
Percentile
3.96
3.83
3.75
3.70
3.64
3.60
3.56
3.51
3.48
3.44
3.41
3.38
3.35
3.32
3.27
3.20
3.13
3.07
2.98
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
Note that the actual norms have percentiles ranging from 1 to 99.
To save space, this example table only shows every 5th percentile.
These values are meant to serve as an example only and do not actually
represent normative data for the Vision index.
As the table shows, Company As raw mean of 3.60 is at the 70th percentile. This indicates that it is
higher than 70% of raw means obtained by the other organizations in the normative database. Company
A seems to be doing a better job than 70% of the companies in the benchmark group when it comes to
creating a vision for the organization.
On the other hand, Company Bs raw mean of 3.27 is only at the 25th percentile. Company B is in the
bottom quartile when it comes to establishing a clear, shared, long-term vision for the companyits raw
mean is higher than just 25% of the raw means in the normative database.
Company A and Company B both receive survey reports with their percentiles on the Vision index.
Their survey responses started off as raw scores, were averaged into raw mean scores, and now appear
as percentiles. The point of this whole process was to make the culture survey results easier to interpret
to give people a better understanding of their organizations relative strengths and weaknesses.