if you are using and applying a statistical method with your data collections already, the CV value is what it is and part of the study.  If you decide to replicate the experiment on another area, that may add to understanding and perhaps contribute to statistical findings.  But it is unlikely if your experimental design was appropriate the first experiment, the new testing will change CV unless conditions are perhaps different.  If you have not measured the conditions such as rainfall, soil moisture, temperature, etc. in the first experiment, you will have no way of directly knowing if there are changes in conditions unless you can find other data sources to help differentiate this.  

If others have done similar experiments and they had always found exceptionally small CV values, you may be concerned that yours showed something else.  You may not know if it is weather, rainfall or lack of, soils, seed source, how application and mixing was done, how measures were taken, equipment going bad or not well calibrated, etc.

Unless you can identify some reasoning, if you applied research grade method  with applicable statistics (whether simple or complex), I see no reason to repeat, as you will be presenting your work, assumptions, conditions, methods, data, and results.  What better an anyone expect? 

if you are using and applying a statistical method with your data collections already, the CV value is what it is and part of the study.  If you decide to replicate the experiment on another area, that may add to understanding and perhaps contribute to statistical findings.  But it is unlikely if your experimental design was appropriate the first experiment, the new testing will change CV unless conditions are perhaps different.  If you have not measured the conditions such as rainfall, soil moisture, temperature, etc. in the first experiment, you will have no way of directly knowing if there are changes in conditions unless you can find other data sources to help differentiate this.  

If others have done similar experiments and they had always found exceptionally small CV values, you may be concerned that yours showed something else.  You may not know if it is weather, rainfall or lack of, soils, seed source, how application and mixing was done, how measures were taken, equipment going bad or not well calibrated, etc.

Unless you can identify some reasoning, if you applied research grade method  with applicable statistics (whether simple or complex), I see no reason to repeat, as you will be presenting your work, assumptions, conditions, methods, data, and results.  What better an anyone expect? 

you can build confidence intervals. (bootstrap) and use hypothesis testing to accept/reject your result at a chosen level

Dear Vö,

CV% is dependent on some of those factors:

1. the experimental design,

2. the replication / plot size,

3. the experimental material you study.

The literature in your study area gives you a clue about the range of your acceptableCV%.

After checking the literatüre and you still think your CV % is higher than you wish to have, these are the things you can do:

1.Increase the number of replication and / or plot sizes,

2. Use a more effective expermental design, i.e. RCBD instead of RCB or triple lattice instead of RCBD, etc.

3. Use more unifom exprimental material, site etc.

Good luck.

it is ok for field.experient

Võ Tính Thiện

The acceptable CV level is depended on the spirit of the research. For instant, in clinical chemistry research, acceptable CV is different with field experiments in agriculture. But in overall you can check: STATISTICAL PROCEDURES FOR AGRICULTURAL RESEARCH (Second Edition) by KWANCHAI A. GOMEZ or you can have a look at: http://clinchem.aaccjnls.org/content/clinchem/35/4/630.full.pdf

The acceptable CV level is depended on

1. the experimental design

2. the number of replication and its size

3. the experimental material of your study and the parameters that you want to record.

The recent literature in your particular field area will give you a clue regarding the range of your acceptable CV%.

If your CV % is still higher, its means that something is wrong? Check your data again and again.

All answers from researchers here are that it makes sense to me. It depends on characteristics of the crops and varieties. For sweetpotato is quite normal with a high cv, but of course, it shouldn't be too extremely high. Based on our agreement with my former colleagues from CIP in 1990s, we could accept it as high as 38%, for example. Then we should give reasons and recommendation on this particular issue. Why is it high?. There are many reasons, including the stability of performance of varieties due to genotype by environment interaction, it might be causing by biotic and in-biotic incidences. Examples can be seen at page 495 of the publication attached. These are commonly recorded in the sweetpotato trials.

CV not greater than 20% is acceptable.

High CV is always harmful to someone who wants to study the accuracy of some variable. Therefore, the higher the dispersion factor, the worse the result.

For field experiments CV of 20% or below is acceptable and for laboratory experiments 5% is the limit.

Hi

If you are deaf in the lab, it must be less than 25, but it is also acceptable on a farm over 30.

What about the clumped distribution of weeds, for instance, you may nave 500 seedlings/m2 in a grid while the next grid may have only 5 or nothing? Imagine you are working in a large field to collect data where different species are dominating in different areas. This a very common scenario of weed spatial distribution out in the field. In a real-world situation is it possible to say a high CV like 250% is invalid? Please correct me if I am wrong.

Any references to these suggested %CV?