Proof for Mean-Mode=3(Mean-Median)

  • #1
1,007
76
Can anybody give proof of the above relationship algebraically?I have not seen the derivation of it.
 

Answers and Replies

  • #2
DrDu
Science Advisor
6,107
806
I don't think that this is a generally valid relationship. I would guess that it holds for a PDF expressed in terms of a lowest order Edgeworth series.
 
  • #3
1,007
76
I don't think that this is a generally valid relationship. I would guess that it holds for a PDF expressed in terms of a lowest order Edgeworth series.
I was in a hurry and when one is not familiar with the derivation ,often one messes the formula.
Yeah the correct relationship is Mode=3Median-2Mean.What is the derivation for it?
 
  • #4
DrDu
Science Advisor
6,107
806
That's the same formula you wrote before, only solved for Mode. It is not a general valid equation. For example, there are distributions which don't even have a mode, but a median and a mean.
I suppose you can get it using an Edgeworth expansion including the skewness and the curtosis:
http://en.wikipedia.org/wiki/Edgeworth_series
 
  • #8
DrDu
Science Advisor
6,107
806
I hardly know more about this relation than you. In the link I found there are all the references to original articles you need.
 
  • #9
1,007
76
I hardly know more about this relation than you. In the link I found there are all the references to original articles you need.
I have seen that link in Mathematics stack exchange before and I did't get it that's why I have posted it here.When you first posted that it is not a valid relationship I thought that modification in statistic may have came for this formula.
 
  • #10
1,007
76
No problem if you don't know but if you can help for some initial steps It would help me.
 
  • #11
statdad
Homework Helper
1,495
35
Karl Pearson (I believe) developed that guideline (not a rule) from observations of many slightly to moderately skewed data sets and distributions. The equality you've written really should be taken as "approximately equal to", since the intent of the relationship was to have a quick way to approximate values. I don't know whether he published a derivation or simply mentioned it in an aside or lecture.
 

Related Threads on Proof for Mean-Mode=3(Mean-Median)

Replies
6
Views
862
  • Last Post
Replies
9
Views
4K
Replies
7
Views
3K
Replies
1
Views
1K
Replies
7
Views
7K
  • Last Post
Replies
1
Views
1K
  • Last Post
Replies
1
Views
14K
Replies
2
Views
1K
Top