This depends on the nature of the data you have collected. Logratio works pretty well

Dr. Agre,

If you go through the weed science paper most of the time they use square root transformation, rather than log transformation. Can you please explain the nature of data collected.

First check that there is no outliers

Next, using the mini tab software, you can check if the data is normal

If your data was not normal, you can normalize your data with methods

Radical,

Exponentiation

And percentage

With respect

There are many techniques for transforming data and each has pros and cons.

However, after transformation, the interpretation should be carefully considered. See the link

https://www.researchgate.net/post/How_can_we_trust_in_the_back-transformed_confidence_interval_and_statistical_analysis

There will be higher treatment variance, block variance as well as replication variance as we are dealing with 2 control treatments (weedy check and Weed free) plus other treatments...

Weedy check treatments always corresponds to higher value of Weed data(weed dry matter and Weed density) while zero value with weed free treatment...if we analyze treatment difference , we will get higher variance..there is assumption of analysis of variance (ANOVA) to have equal variance( others are independence with normality)..square root transformation is used to transform the data resulting in approximately equal variance with variable critical difference..and I think at all the stages if crop growth, one should used same transformation..log transformation is also employed in Weed researches...

Mr Himanshu,

As for analysis in ANOVA is concerned, we expect variation between the treatment, whether it is weed-free or weed check or any other treatment. The very purpose of the transformation is to reduce the within treatment variation, rather between the treatment.