I'm trying to figure out exactly why we do CV in relation to our sample results.
Since CV of less than 10% considered most ideal - but how do we apply that to our results? Do we need to use this percentage and re-calculate our results?
What does '10%' means?
Hello Azalea,
According to https://medicine.yale.edu/labmed/Images/Ed-tox%20assay%20validation%20procedure_tcm45-9304.pdf,
"Accuracy is assessed by determining recovery; tat is , the grand mean of the measured values divided by the target value. The precision is best estimated as the overall standard deviation of the mean and is typically expressed as the coefficient of variation (cv = standard deviation/mean). For most analytes, a coefficient of variation less than 5% represents acceptable performance. Larger coefficients of variati on may be acceptable at the discretion of the Laboratory Director. Coefficients of variations exceeding 10% are rarely acceptable, except at very low concentrations.
The difference between the percent recovery and 100% is bias. The bias should not exceed the coefficient of variation"
Check this out for why we do CV in relation to our sample.
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CCcQFjAB&url=http%3A%2F%2Fwww.researchgate.net%2Fpublictopics.PublicPostFileLoader.html%3Fid%3D54bc01fed039b1017c8b4589%26key%3D03de2533-e9e4-4ece-8521-b298ac2feb12&ei=TTbqVIqwObiSsQTukIGgCg&usg=AFQjCNHw4F0LB1KHDewBJkN4aYJ1BJ2xow
I hope this helps
thanks Isaac Akogwu - I understand why we do inter-assay and intra-assay CV to check for precision and performance of the assay. But how do we apply that to our own sample results?
If let's say the CV is around 11%, so do we need to calculate our data - multiply by 0.11 (11%) and that means our results is truly equivalent to +/- 11% from the actual value? It may be irrelevant or wrong what I'm saying here, but I just hope someone could give some clarity on what does CV means to the assay value of our samples.
In publications, they only report the CV values, but I would like to understand how will that affect our results that we're reporting them. Because I have issues with CV calculation, that some were more than 15%.
The Cv can only be apllied to compare the dispersion of distributions that have no outliers. it is a relative measure of dispersion. If there are outliers you should apliied MAD instead of CV.
Dear Othman,
The coefficient of variation represents the ratio of the standard deviation to the mean, and it is a useful statistic for comparing the degree of variation from one data series to another, even if the means are drastically different from each other.
Inthe other hand, coefficient of variations gives us the degree of variability of obtained data of an experiment. This indicator depends on the variability of measured experimental units. If this indicator is high it means that it is difficult to detect significant differences among the means of treatments considered. However, if this statistician is low, it is easy to detect significant differences among treatment means. Therefore 10%' indicates that the variability degree of data group is 10%.
I hope this solves your request.
Sincerely yours,
Angel
Hello Othman,
C V is ratio relating the measure of dispersion(Standard deviation) of your data points to the central value(mean) i.e. CV= standard deviation/mean. It is a good measure for comparing spread of the distribution datasets/ data groups that are not necessarily in the same unit, since it is measured in percentages. The lower the CV, the better the distribution i.e.if data set A has a CV of 10% and B has a CV of 15%, then A is said to have a better distribution than B .To understand this I want you to look at CV(Coefficient of Variation) as the constant of variation when standard deviation(dispersion) varies directly as the mean(central value) or better still, as the percentage deviation from the mean i.e. a CV of 10% will imply 10% of the data points in the data set deviated from the mean while CV of 15% will imply 15% of data points in the data set deviated from the mean(the center), it then follows that less data points in data set A than in data set B deviated from the representative value (i.e mean) of their respective data set. Hence data set A is more evenly distributed than B.
LIMITATIONS:
1). CV may not be a good measure where are many / significant outliers(extraneous data points).
2) .CV becomes undefined and fails to be useful as the mean tends to zero.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics