I don't think it is quite so black and white as that.
It might not be nowadays, but it certainly started out that way. In fact, the two ranges only appeared when Openreach brought out the option for self-install. For me, at least, the old single range (when only engineer installations were available) mapped directly into the clean range.
An engineer install should be deemed not impacted as checks should be made for poor line quality and local bridge taps but I very much doubt that all engineers do these tests as they are usually under time pressure or may not even know they should be doing them.
The installation guys I have had (all Openreach) told me that they have to run the tests. The installation cannot be closed off without test results.
In addition, IIRC, line tests had already been performed before they turned up.
On the other hand my self install is very unlikely to be impacted as my drop wire goes directly into the master socket and there are no extensions.
The determination of clean or impacted has nothing to do with whether the line is actually clean or actually impacted. Modem statistics are, unfortunately, irrelevant.
It is about whether Openreach can
trust that it can be classified one way or the other. It is all about the
guarantee that comes after the installation.
While the engineer
ought to ensure a properly-performant line, Openreach are willing to
guarantee that performance by setting a higher range (and higher fault threshold) by virtue of having an engineer install. They embody the guarantee by agreeing to send out an engineer in future at higher thresholds.
The obvious scenario that Openreach wanted to avoid when self-install was introduced: They didn't want to make it cheaper for punters (by allowing them to not pay for an engineer), to only then have to offer an engineer callout free-of-charge in order to fix the problem that only exists because the punter didn't want to pay for an engineer. Less income, higher costs ... not a good business move.
What was needed, alongside the option for self-install, was a lesser guarantee that would avoid any extra demand on Openreach-funded engineer callouts (or "truck rolls" as the Americans call them).
The lower range is the method that Openreach have chosen to apply a lesser guarantee. I'm not so sure it was the best way - it leads to inconsistencies if you get an engineer install on day 1, but migrate to a self-install supplier; why should a lesser guarantee apply in these circumstances?
For myself, I will always choose an engineer installation, with an ISP that quotes their minimum speed using the clean range. That way, if faults happen in the future, I'll be able to get an engineer involved at the higher speed rather than the lower one.