Thanks for raising the questions about random sampling, we could have done a better job of explaining the concept and why we think it’s worth experimenting with earlier. I’ll address the more tactical questions here and we have addressed the higher level principles behind random sampling in: Experimenting with Random Sampling in the Collective. Please use this thread to discuss guest voters in Round 5 and the other thread to discuss random sampling 
50 citizens were invited to the recent deliberation, as I understand it. Why did only 25 participate / vote?
We reached out to 50 Citizens under the assumption that we would have ~50% opt-in rate among those randomly selected, considering participants were required to attend 3 live sessions at set times. Opt-in rates were in line with our expectations and the experiment was designed with ~25 participants in mind. Citizens that are randomly sampled to participated in Retro Funding voting, will be required to participate.
There seems to be an assumption that if citizens don’t vote, they are lacking confidence
Agree that lack of confidence is just one of many reasons Citizens might not vote! It’s important to understand the Token House and Citizens’ House ability to understand technical proposals. Asking voters whether they are confident enough to cast a veto vote is one way to start to measure that. It helps us identify voters that have already made it to the voting platform, and have ostensibly read the proposal, but weren’t confident enough in their ability to evaluate a technical proposal (otherwise this would just show up as a missing vote.) That doesn’t mean lack of confidence is the only reason someone who is already on the voting platform might not cast a vote, and it definitely doesn’t capture those that never even made it to the voting platform for other reasons. Our goal is not currently to measure all the reasons people might not casting votes, but rather to measure whether the Citizens that are already casting votes are confident in the votes they are casting.
But where does that assumption come from? Not one citizen expressed lack of confidence in the most recent veto vote. How do you increase a percentage that is already 100?
That’s what adding this measurement helps us determine! We had never measured confidence before this vote; collecting this information helps us validate/invalidate assumptions.
It would make sense to actively support those who are already active and offer them a sense of community and purpose.
100% agree with you; I don’t think this is mutually exclusive with random sampling though and explain more about that here. There are many things that can be done to better promote community and a sense of purpose and to allow those that want to engage more to do so. However, there are many ways to achieve those things (Establishing clear Citizen expectations, developing robust reward mechanisms, special non-voting roles for high context Citizens such as mentorship, better Collective context sharing, etc.) that are not at odds with random sampling.
Why not make participation an “opt-in” process or exclude Citizens who did not participate in previous rounds if the current number of participants is too large to manage effectively?
In short, requiring an “opt-in” process invalidates the properties of random sampling (see here.) That said, of the participants that are randomly sampled, they must still agree to participate. Participation in Retro Funding votes is currently a requirement of continued citizenship. The reasons we believe random sampling is a scaling mechanism worth experimenting are outlined in detail here.
Why aren’t Citizens asked if they want a sampling, opt-in, or full participation?
In order to run effective experiments, the Foundation needs to be able to manage the conditions and variables of those experiments. We believe this experimental approach will allow the Collective to develop an industry-leading approach to metagovernance, which will transition to the community over time (See the Path to Open Metagovernance.) It is also very important to make sure metagovernance does not allow for the current set of Citizens (at any time), or a small but active subset of them, to be able to limit the rights of other Citizens to participate in the system (which is partially why Citizens are not asked to vote on this.)