Help Trying to find Averages With Percentage Ranges (work experience)

ckdowling

New member
Joined
Nov 30, 2017
Messages
1
Help! I need the average years of experience for individuals who answered a survey with separate ranges.

If survey respondents answered that...

24.7% have 1-5 years experience
10.9% have 6-8 years experience
18% have 9-12 years experience
27.6% have 13-18 years experience
16.7% have 19-25 years experience
2.1% have 25 years experience

How do I calculate the total average experience of the group??

Any help is greatly appreciated!
 

tkhunny

Moderator
Staff member
Joined
Apr 12, 2005
Messages
9,824
You must pick a representative age in each classification. The whole range will not do, as you have noticed.

If you have more data, you can use a Mean, a Median, or a Mode - whatever you think is most representative. Lacking additional information, a midpoint may have to do. The only real problem with this methodology is the last grouping. Are your '25's REALLY all exactly 25? If the group really means "25 or more", then it may be more difficult to find a representative age for that classification.

Let's see where this leads you. :)
 

stapel

Super Moderator
Staff member
Joined
Feb 4, 2004
Messages
15,945
If survey respondents answered that...

24.7% have 1-5 years experience
10.9% have 6-8 years experience
18% have 9-12 years experience
27.6% have 13-18 years experience
16.7% have 19-25 years experience
2.1% have 25 years experience

How do I calculate the total average experience of the group??
Assume you have some nice number of people; because of the one decimal place for the percentages, let's go with 1,000 (rather than just 100, which would give us fractional people). Then how many people are in each of the classifications?

When working with ranges like this, often a midpoint is picked. However, since the last classification isn't a range of values, you may want to specify that you're using the lower endpoint of each classification, unless the directions for this exercise specified something else (and you should have been given specific instructions).

So the 247 people in the 1-5 classification have a total of 247*1 = 247 years (or, if you use the midpoint, 247*3 = 741 years) of experience. And so forth. How many years do you end up with? Dividing this by the number of people, what then is the average number of years of experience? ;)
 

BeeCuz

New member
Joined
Jun 6, 2016
Messages
17
Here is a graph of the stated percentages
Capture.PNG
Notice how it's not an even randomly distributed arrangement? It's not like the height of people in a theater, or the flips of a coin. There's a pattern that looks a bit like a bell curve, but there's also an exception to the pattern. Average would hide the fact that there's a lot of newbies and a small amount of longest termers.
 

JeffM

Elite Member
Joined
Sep 14, 2012
Messages
3,258
Building from post 4, an average of any type reduces numeric data to a single number and therefore LOSES information. The advantage of an average is that it is easier for the human mind to comprehend one number than multiple numbers. The purpose of an average is to indicate what is the typical value.

In your particular case, you have already summarized the data down to six ranges. The human mind can easily grasp six numbers so the advantage of further reduction is small. Furthermore, in this particular case, what you lose by the reduction is the important fact that you have a bi-modal population, a big group with almost no experience and an even largeer group with substantial experience. If your boss insists on an average, then you can use the advice in posts 2 and 3 to calculate one though I would advise against it.

If using six ranges is considered more information than the likely users can easily comprehend, I suggest either using quartiles or saying " 25% have fewer than 6 years of experience, 45% have more than 12 years of experience, and 30% are in between." I somewhat prefer reducing to three ranges because it shows clearly that those with experience greatly outnumber those with little experience, and that those with considerable experience represent almost half the entire group. A single number will lose both those pieces of information.

In general, there are mathematical techniques to summarize data. That is math. How to present data is about psychology.
 
Top