You suspect that the pattern relating sales to levels of advertising is not linear (perhaps a log transformation is needed to represent diminishing marginal returns11). Explain how you can still use...

You suspect that the pattern relating sales

to levels of advertising

is not linear (perhaps a log transformation is needed to represent diminishing marginal returns¹¹). Explain how you can still use the SRM as a framework for modeling the variation in the data.

Suppose that large diamonds (more than 1.5 carats) sold at retail tend to be of very mixed quality, whereas small diamonds have consistent quality (more uniform color, clarity, and attractive cut). If this is the case, can we use the SRM to describe the relationship between price and weight among diamonds of widely varying size?