Reaction time is but a measure of a broader concept - mental speed or processing speed - that many investigators are using for a so-called scientific measure of intelligence. It is widely believed by many that IQ tests are intrinsically arbitrary - insofar as IQ tests reflect any real differences between people at all, these differences endure because they are created and perpetuated by lasting social and educational injustices that are believed to be prevelant in the Western world. In other words, intelligence, as it is currently measured or defined, relys on criteria which are subjective and social rather than objective and scientific.

The thrust among these critics has been to search for a true single information-processing function underlying intelligence. Many of the early attempts at finding this function were attempts to measure speed of "output" or hypothetical central processes. Recently many researchers have been focussing on the brain's immediate reaction to sensory input (i.e., reaction, response, or inspection time). Psychologists have a more general term for these - average evoked potential (AEP).

Getting to your concern of some people with wandering attention or the tendency to block out secondary stimuli. For a study like I outlined, with an individual reading a passage and responding to a secondary stimuli, this may be a concern. However, it is really a secondary "task" that the person is to perform. The tone in this case isn't background noise but something the subject is required to respond to. Tho this doesn't entirely address your concern (which is very valid) it does offer some mitigation.

Moreover, in most research aimed primarily at relating processing speed with IQ, the reaction time is a primary, not a secondary response. And as for the wandering attender, most such tests require a base response of some sort. Someone not attending would not receive a score. Empirically, it's not a problem, although it remains a problem when it comes to assigning an IQ score to that individual.