In Andy Field's Discovering Statistics Using SPSS he states that all variables have to be transformed.

However in the publication: "Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation" they specifically state that only the non-normal variables were transformed.

Is this analysis specific? For instance, in a comparison of means, comparing logs to raw data would obviously yield a significant difference, whereas when using something like regression to investigate the relationship between variables it becomes less important.

In Andy Field's Discovering Statistics Using SPSS he states that all variables have to be transformed.

However in the publication: "Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation" they specifically state that only the non-normal variables were transformed.

Is this analysis specific? For instance, in a comparison of means, comparing logs to raw data would obviously yield a significant difference, whereas when using something like regression to investigate the relationship between variables it becomes less important.

Edit: Here is the full text page in the "Data Transformation" section:

And here is the link to the paper: http://www.sciencedirect.com/science/article/pii/S0048969708009121

In Andy Field's Discovering Statistics Using SPSS he states that all variables have to be transformed.

However in the publication: "Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation" they specifically state that only the non-normal variables were transformed.

Is this analysis specific? For instance, in a comparison of means, comparing logs to raw data would obviously yield a sig difference, whereas when using something like regression to investigate the relationship between variables it becomes less important.

In Andy Field's Discovering Statistics Using SPSS he states that all variables have to be transformed.

However in the publication: "Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation" they specifically state that only the non-normal variables were transformed.

Is this analysis specific? For instance, in a comparison of means, comparing logs to raw data would obviously yield a significant difference, whereas when using something like regression to investigate the relationship between variables it becomes less important.