There are, in most radios built, one of two types of signal strength indication -
A) Analog ‘Signal Strength’.
B) RSSI digital indicators.
Whilst early non calibrated types all fall into the a) type, regardless of their actual form but only act as a relatively crude indicator which really only aid tuning in at best.
S-meter type analogue meters, which are indicated calibrated and are calibrated against a reference, indicate a relative signal strength and irrespective of how they actually display, are analog relative indicators that you can derive some information from. Notably, a radio reception report would normally include a reference to received relative strength indication and an estimated grading of how stable it is.
Digital receivers use the b) type, RSSI, which in numeric display mode show a dBM/ASU pairing or one of the two. dBM indication roughly equates to an equivalence to an S-meter relative indication but are much more consistent in how accurate the reading is and like readings on a digital multimeter, fluctuate a lot which in itself is actually useful. In bargraph mode, RSSI meters actual indicate a derived relative value which indicates SQ or signal quality.
What you can see in SQ mode, isn’t any equivalence at all to an S-meter, in fact in classic radio reporting terms, SQ would be what you’d use today to determine the S in RST values of a reception report, where RST was never a set of absolute values, just a graded table you estimate quality of Readability (clarity and a rating of how disrupted the resolved and demodulated content is), S for how good the reception is in terms of ‘above the noise’, so S9+30 of a received signal vs an ambient noise SIG level would be an easy 9 for S in RST - that kind of distinctive gap would also be a pretty solid 9 in R terms if the content was clear, less if whilst the SIG was high, clarity wasn’t so.
So you could say, on one level, that RS in RST is effectively represented by SQ in pure digital RSSI indication as usually (unless there’s a case of garbage in garbage out of content) a high SQ usually had a high quality of readability in the transmission.
Another RSSI indication of SQ will be a dual numeric indication of SQ and number of errors (where errors here means bad data packets/frames). So if you could switch to numeric RSSI, you’ll see up to four number sets on some equipment, the SQ reading and Error rate but also the dBM and ASU readings.
If you’ve got all four RSSI elements accessible, you have a regular swiss army knife of a tool you can really use to relatively grade performance of a setup receive/link wise, and by developing a good feel for how they translate, makes refining alignment and optimal positioning very easy indeed.
So that’s why you shouldn’t see a low SQ reading as bad, if it’s steady low and whatever is sent is making the trip in a resolved clear fashion, it’s with threshold and usable. Not that cell phone users would appreciate the significance of it, but when you are trying to align a long distance link setup, you are looking at getting a stable SQ even if low - further refinement will result in improved readings as you fine adjust. If you hit a point where you get a huge SQ but its fluctuating between two wide extremes, that’s usually indicating over signal level received and you would want to somehow lower the relatively received SIG level to minimise that. Also, under that condition you’ll see in the numeric view that you see error rates climbing despite having a high SQ peak.
Equally, a solid SQ level with huge error rates is equally a sign over overly high SIG level received. In both extremes, it’s a sign of the receiver getting overloaded, the SNR inside the receiver has gone to ■■■■ rapidly and a lot of those errors are actually due to noise and resdidual artifacts in the receiver caused by the overload.
As I said in other replies and articles, and no more definitely demonstrated in digital transmission modes, there are absolute thresholds where above a certain point, no amount of extra gain on the receiver will benefit matters and actual be the start of a slippery steep downhill slope to deterioration again as much as below lower thresholds of relative SQ and received strength has a rapidly decreasing SQ and usability as you move towards the noise.
The exact way this affects the content depends highly on the content and how it’s transmitted, but ultimately when you’re listening/resolving deteriorated signal content, the end result is degrees of awful through useless - especially for voice mode and telegraphic modes where high data rates and large groups of data is packet transmitted (video being a good example).
A reading is just that, a reading - but until you can resolve what it’s really telling you, it’s just ambient data of very little conclusive truth without having the context to use it properly.