Id say we’re interested in short, medium, and long term wins. What I’m suggesting is one of many things that can be done in the short term to try to increase engagement until some of the longer term things come to fruition. Why wait around doing nothing until we have better and more usable products?
If Kusama is known for Chaos, why not add this sort of feature to nomination workflows and see how the experiment turns out? 
I’d be curious to hear your thoughts on what “quality” means. But one way I see it is (from the perspective of the chain), “quality” means voting with a lot of tokens, as this enables referenda to pass or fail more quickly. The chain itself does not necessarily care how “thoughtful” a vote is. Of course it would be nice to try to cultivate thoughtful discussion and feedback on proposals. But the chain itself does not care about this. It only cares about the vote direction (yes/no/abstain), and amounts.
If you’re curious what this looks like in more active governance ecosystems, in Cosmos validators are very active in governace (and many Cosmos validators are also Polkadot/Kusama validators). And people that back them actually do care about these decisions and their outcomes. See this reddit thread on the massive amount of backlash stakefish got for their last minute vote on prop 89, in which case many people undelegated and unstaked with them.
What I’m suggesting with that is a voluntary decision by a user - it would be unchecked by default and only exist there as a very streamlined way to also delegate if they want. Users currently have no incentive to delegate governance votes to the validators their backing, as it gives them no additional rewards. There’s also no good UI at the moment for someone to even delegate if they want to, so this would be an option for them to be able to also delegate in a streamlined manner until those ui’s exist. When they do it would be better to include links to those in the nomination workflow instead. And per what I mentioned above, also have banners / toasts/ reminders for people if they delegated a while ago to revisit their delegations (so if they would delegate to a validator now, they’re encouraged to change it in the future since it might be old).
This is being worked on in a few ways, but will probably take a while until it actually launches.
More tokens voting will improve the network by allowing things to be able to reach quorum to fail/pass more quickly.
If you look at discussions on Polkassembly / Subsquare, some of the most active contributors are Validators. Just because someone participates in the network as a Validator doesn’t mean they are only out to earn rewards and not meaningfully contribute.
Just to put some data to this, I’m aggregating a bunch of metrics for participating over time. I’m looking to segment the activity by various groups of people (if anyone has suggestions on ways to classify more groups of people, lmk). This also is being indexed over time, so we will be able to see how activity changes. As I mentioned above, at the moment one of the things that matters most is the amount of tokens voting, which is why I try to delineate that as well.
(all of these categories only count for votes themselves, and not the amounts that are delegated to them)
All Votes
Total Votes: 25285
Casting Votes: 18138 (71.73%)
Delegated Votes: 7147 (28.26%)
All Extremely Low Balance Votes (<0.5 KSM): 11574 (45.77%)
All Very Low Balance Votes (<1.5 KSM): 13485 (53.33%)
All Low Balance Votes (<5 KSM): 15742 (62.25%)
All High Balance Votes (>5 KSM): 9318 (37.75%)
As we can see from this, the vast majority of current votes are less than 0.5 ksm, which effectively don’t do anything. This is likely from the Proof of Chaos nfts, or trying to inflate the nubmer of yes / no votes, as we can see that accounts like this automate sending a ton of amounts of 0.015 ksm to accounts to then vote.
Validator Votes
Validator Votes: 14617 (57.80%)
Extremely Low Validator Votes (< 0.5 KSM): 6258 (42.81%)
Very Low Validator Votes (< 1.5 KSM): 6743 (46.13%)
Low Validator Votes (< 5 KSM): 6925 (47.37%)
High Validator Votes (> 5 KSM): 7569 (51.78%)
Validator accounts include any accounts that are also apart of the identity of validators. Per this we can see that validators accounts for more than half of all votes currently (at 57.80% of all votes). The percentages next to the amount segregations (for this and all the following) are out of just that group (instead of the total). The majority of them look to vote with above 5 ksm.
Nominators
Nominator Votes: 4668 (18.46%)
Extremely Low Nominator Votes (< 0.5 KSM): 1287 (27.57%)
Very Low Nominator Votes (< 1.5 KSM): 2070 (44.34%)
Low Nominator Votes (< 5 KSM): 3518 (75.36%)
High Nominator Votes (> 5 KSM): 1108 (23.73%)
Nominators are classified as acconts that are currently nominating (and not also a validator). Most of them look to be voting with between 1.5 ksm and 5 ksm.
Non-Stakers
Non-Staker Votes: 6000 (23.72%)
Extremely Low Non-Staker Votes (< 0.5 KSM): 4029 (67.15%)
Very Low Non-Staker Votes (< 1.5 KSM): 4672 (77.86%)
Low Non-Staker Votes (< 5 KSM): 5299 (88.31%)
High Non-Staker Votes (> 5 KSM): 641 (10.68%)
The majority of them vote with < 0.5 ksm, and are probably bots as I pointed out above.
Identity
Identity Votes: 16202 (64.07%)
Extremely Low Identity Votes (< 0.5 KSM): 6467 (39.91%)
Very Low Identity Votes (< 1.5 KSM): 7122 (43.95%)
Low Identity Votes (< 5 KSM): 7611 (46.97%)
High Identity Votes (> 5 KSM): 8398 (51.83%)
Identity votes are votes that include all the accounts apart of all accounts with an on-chain identity. Since most validators also have an identity, a lot of the people that this group is composed of is also likely them (there’s only 668 on-chain identites on kusama at the moment).
Fellowship
Fellowship Votes: 85 (0.33%)
Extremely Low Fellowship Votes (< 0.5 KSM): 5 (5.88%)
Very Low Fellowship Votes (< 1.5 KSM): 7 (8.23%)
Low Fellowship Votes (< 5 KSM): 18 (21.17%)
High Fellowship Votes (> 5 KSM): 42 (49.41%)
As @bkchr mentioned elsewhere, the fellowship might vote with accounts that are not directly tied to their fellowship account, so that might be why there’s a low rate of participation here. But of the 40 something members, the majority have 0 votes, these 85 votes are from a smaller subset of the total.
I’ll be aggregating and indexing this data over time, so will be curious to see what kind of experiments on kusama have different affects and influence it in different ways. If anyone else has ideas on how on-chain engagement and participation can be analyzed, or what kinds of experiments can be done to increase participation, would love to hear it 