I want to calculate pooled genetic gain across 5 environments but trials across 5 environments are in different designs and different number of replications. So practically speaking I cannot do combined analysis. Therefore I have done individual analysis. But I would still like to see genetic gain across all five environments. Can I use mean values of each of these environments to get a pooled mean across 5 environments and use pooled mean values to estimate genetic gain across 5 environments? If not want is the best way to do pooled genetic gain?
Bharath
One option for pooling the means to get pooled estimates is to use the concept derived for meta-analysis. Here, the means are weighted by their sample sizes. So you'll need the means, sd, and N for each group. The measures you can choose from are: Cohen's d, Hedges's g, Glass's Delta, (among others).
Any meta-analysis package can calculate these for you, or you can do it manually if necessary.
I hope this helps
Ariel
I did not get full insight into your experiment if you could use a mixed model for combined analysis. But if you decide to make weighted mean, it is done like this:
When we compute a simple arithmetic mean of a set of data, we assume that all the observed values are of equal importance and we give them equal weight in our calculation. In situations where the numbers are not equally important we can assign to each a weight which is proportional to its relative importance and calculate the weighted mean (see attachment for an illustration).
@ Gemechu Keneni.... Weighed mean makes sense when population size is not big and there is one type of experimental design. But in my case my population size is 100 entries and I have augmented design in 1st year and alpha-lattice in second year. As per my knowledge there is no statistical method to combine these two experimental designs to do combined analysis. If there is any please let me know
Yes you can use mean of means in your case. When you have a large sets of populations and non-equal number of observations, the means are taken on the basis of number of observations in each case individually and then these means can be computed for overall mean and to estimate various genetic components. Please see my publication in this kind of situation. Heritability estimates for winter hardiness in lentil under natural and controlled conditions.
A. Ali, D. L. Johnson
Plant Breeding 06/2008; 119(3):283 - 285. DOI:10.1046/j.1439-0523.2000.00491.x · 1.34 Impact Factor
Yes, You can do. Then, use each environment as a block. Then, gain is there. But comment with caution. NZ
Thank you Dr. Nusret Zencirci,
Is there a way to combine augmented design study with alpha-lattice design for a population size of 100 entries? Some people consider each environment as a block and analyze it as an RCBD for breeding purpose but I think block size is too big to address field variability. Is there a better way to do combined analysis under this circumstances
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics