what is the cut off value of coeffcient of variation ( %) , above which the variation is considred to be significant ?????or At what coefficient of variation the data is considered to be homogenous???????

The 'correlation coefficient' is designated by 'r' where r is:

r = (1/n)(SUM[(x-xbar)x (y-ybar)]) divided by (sigma subx)(sigma suby)

where n is the number of pairs and x & y are the values of the two variables. To test for whether the apparent correlation is significant use a 't' test where:

t (no. of sigma) = r multiplied by the square root of n

If t is greater than 3 consider there is a significant correlation

Rick,

I was just scanning the forums and noticed your reply to this query. She is asking about the significance level for a coefficient of variation or CV! It is defined as STDEV/MEAN *100.

It is hard to quantify exactly where it becomes a significant value since it depends upon how much variation in whatever is being measured your business can tolerate. I would use it as a relative measure when comparing groups of data. Similar to comparing STDEV between populations. Hope this helps.

I've really never seen much use for the coefficient of variation.

Does anyone out three use it regularly? Any tips?

Marc is correct.

The coefficient of variation is defined as the standard deviation divided by the mean as a percent [(s/Xbar)*100]. It is used primarily when comparing two distributions that are either:


The data are different units, or
The data are in the same units, but the means are far apart.


The coefficient of variation has no units. It measures relative variability, that is, variability relative to the magnitude of the data.

<font color=#c001c0><blockquote>…what is the cut off value of coefficient of variation (%) , above which the variation is considered to be significant ?????</blockquote></font>

Since CV is used based upon the above uses, CV in and of itself is not significant. Significance applies when comparing two or more variances. Typically, when determining significance for two variances of different units, some method would need to be applied to convert both variances to the same unit.

<font color=#c001c0><blockquote>At what coefficient of variation the data is considered to be homogenous???????</blockquote></font>

Coefficient of variation does not determine homogeneity.

Ken K.,

Yea, I use it some, but rarely.

Regards,

Don

------------------
I was better but I got over it.