Hello Ahasan,

I suggest that you start simple - no transformation yet. Using residuals of fish lengths and weights, check if assumptions (eg normality, etc) of your analysis are satisfied. I suggest that you do data transformations only if you can extract additional information that is not available from the simple analysis.

You mentioned reason of performing log-transformation: to remove the size-dependent variation. Again, ensure that you can communicate what resulting log transformed coefficients mean.

Using residuals of fish lengths and weights, check if assumptions (eg normality, etc) of your analysis are satisfied, to remove the size-dependent variation. Again, ensure that you can communicate what resulting log transformed coefficients mean.

You can do it without log transformation. You can do it very simply in Excel.

thanks all.

Hello Ahasan, in my opinion, the most used expression for this relationship corresponds to the allometric equation where weight is expressed in terms of length, and its parameters are estimated by linear regression of log-transformed data. Since the variability in weight usually increases with length, this transformation has the advantage that tends to stabilize the variance of weight (Hayes et al., 1995).

If you take log transformed data, you can take a linear relationship. This may useful to determine the growth (isometric or allometric), and also environmental healthiness, using condition factor

It would be better if  you check the data normality and transform it with log, since lenght and growth is the function of logaritmic. in addition, this transformation will help you minimize the error or variation and your data will have better performance if you'd like to make it in to regression.