I would argue that judges, especially those who began during the 6.0 era, which honestly is probably still most of them, are in the habit of thinking about overall score.
Except the 6.0 scores were arrived at in a holistic way, not by building up points for elements and five different components.
Judges may indeed be thinking "A was best at jumps, B was best at presentation, I think that overall the best performance was from A (or B, or C who was not best at anything but had no weaknesses), which is how they would have thought under 6.0, with more specifics but no official base values etc. But that's completely different from trying to keep to keep track of the IJS scores for elements within a point or two much less tenths and hundredths of points.
It may not be in their job description, but I am quite willing to bet they still are able to do it and that many of them do. Especially when they get curious about why certain scores seem not to match their impressions- that would make me pay more attention!
They won't hear or see the scores that don't match their impressions until after they have finished scoring. They'll hear/see the total scores when they're announced right after the skate, and they'll be able to check the detailed protocols after they get off the judges' stand. But if they have specific questions/curiosity about the event in progress, they won't get any answers at a time that will affect their scores for anyone in that event.
As for the argument that they have too much to think about and not enough time, I say rubbish. They can mark a skill quickly, checking off GOE points as they see them, and then pay attention to other qualities, especially since they can, as you clarify, take notes! Still enough time to be aware of levels, even, though they don't control how those are called. I imagine a program full of high level skills will give an impression that an athlete has better skating skills, so being aware of what was completed might be something they'd want to notice.
The spin levels don't really relate to the skating skills. Maybe certain kinds of difficult entries or exits. Maybe the change of edge feature.
Seeing that a skater is attempting difficult features in a spin and executing them well (or poorly) would likely affect the GOE the judge chooses to award for that spin, as well as the Performance/Execution score.
When judging a spin, all judges need to think about all the positive GOE bullet points and any necessary GOE reductions if applicable and make their decision about what GOE to award.
For a combination spin with flying entry and change of foot, for example, all that shows on the screen is FCCoSp. The base values for that general code range from 1.1 (two positions, no features, "V" designation for inadequate fly and/or failure to reach the landing position within 2 revolutions and hold it for 2 revolutions) to 3.5 (three positions, four features, full credit).
Will even the judges who also have tech panel appointments remember the base values for each possible permutation of level, number of positions, and inadequate fly/landing, and also the values of the positive and negative GOEs (0.5 and 0.3 each, respectively)?
Or even remember the base value differences between that and the FCSp (flying camel), FCCSp (flying camel with change of foot), and FCoSp (combo spin with flying entry and no change of foot), depending what different skaters choose to do as their freeskate flying spin?
For step sequences, the features are more directly related to skating skills, and the judges probably will be thinking about how the skills demonstrated in the sequence will affect the SS component as well as the GOE for the sequence. They will notice if the skater is turning in both directions and doing a variety of difficult turns and steps and using their upper body and whether they do several difficult one-foot turns on the same foot in succession. But will they be counting exactly how many of each kind of turn and step they do and how much of the time they turn in each direction? And will they remember that the base values for level B through level 4 range from 1.5 to 3.9, and the positive GOEs are the same for levels B-3 but higher for 4, and the negative GOEs are lower for levels B and 1, same as the positive ones for level 2, and same as the level 4 positive for both levels 3 and 4?
A skater missing credit for one or two turns not on clear edges could make the difference between level 2 and level 4 without necessarily affecting the judge's GOE. So even a judge who had memorized the features and the scale of values but didn't see those specific edges as too flat wouldn't know if their +2 would put the score for that sequence at 3.6 or 5.3. They're more likely thinking "Good musical phrasing and nuances, choreographed to the theme of the program, good energy, good flow, mostly clear edges" and not "Three counters, three rockers but all the same direction, two loops, was that a bracket or a counter?, does that count as twizzles or traveling threes?" Counting the turns is the tech panel's job, not the judges'.
You asked about whether or not gymnastics has any sort of PCS. The short answer is "not really." The longer answer is that there are a few potential tenths worth of deductions available each for artistry (on floor and beam), dynamics, and amplitude. If you thought a gymnast presented a very poor quality routine overall with no real attempt at artistry and little evidence of physical conditioning relative to the skills thrown you would be able to take about a point off their score between these three deductions, but I have never, ever seen that done. More likely you'll see a tenth off here and a tenth off there, totaling maybe .2 to .5. Not generally enough to make a very big difference except, maybe, if you use them to differentiate between event finalists, who probably have less to take in these areas anyways.
Good to know, thanks.