We should have a basic understanding of what a frequency response graph is, what it tells us, and what it doesn't tell us.
Today, a few pointers on a variety of considerations related to this topic.
Pay attention to the "smoothing" used in the frequency response graph. This will be expressed as a fraction. The smaller the fraction the more resolving the graph will be. For example, 1/12th octave smoothing may show you deviations in the frequency response that 1/3rd or 1/1 smoothing does not. Most manufacturers will use 1/3rd or 1/6th octave smoothing for their product pages.
If comparing two different frequency response curves check that the X and Y axis both use the same scaling. If they don't keep that in mind during your comparison.
This may seem basic but try to make sure it is an actual measurement of the speaker or subwoofer and not a simulation. Simulations can be accurate but that is hardly guaranteed.
If a manufacturer of a speaker or subwoofer doesn't have basic frequency response measurements of their products available on their website take that as a warning flag. This may indicate they don't even have the ability to measure their own product. If you don't think this ever happens, look around the internet a bit...it won't take you long to find examples.
A "Compression Graph" is simply a frequency response graph repeated at progressively louder output levels. This type of graph is informative because it will show you how the speaker or subwoofer maintains its 'composure" as the output levels increase. Ideally, the response trace won't change much. However the reality is all speakers and subwoofers have their output limits. With subwoofers the compression graph deviations tend to manifest themselves in the deepest bass initially.
Power Sound Audio