Amendments tomorrow: QX (quorum based on past attendance)

28 views
Skip to first unread message

Alex Dehnert

unread,
Sep 27, 2026, 9:45:06 PMSep 27
to dsnc-di...@googlegroups.com

It looks like we'll be taking up amendment QX again tomorrow night. It's been a while since we last talked about it, so I figured I'd refresh people's memories before the meeting happens and maybe kick off a little advanced discussion and share my views.

The voter guide, with the amendment text and the arguments in favor and against are at https://docs.google.com/document/d/1gkD1kiPpmWh0PzL0k0HUCb_qMl8HnVd793f2lDrL9t4/view

A short summary is that amendments QX would make (general&special) meeting quorum based on attendance at recent meetings, rather than any fixed number (starting last meeting, it's always 20).

It's a complicated amendment with three subquestions making it up:
A: Use membership attendance or total attendance?
B: Use median or mean?
C: What fraction of recent attendance?

I expect subquestions A&B to be resolved by plurality vote, giving us specific, fully filled in amendments (one per percentage) to potentially adopt by a 2/3 vote.


Getting more into my opinions (I've included the short version of my thoughts on each subquestion in brackets):

General [no]

In general, I have some concern about these amendments.

There's a lot to be said for a quorum requirement that grows and shrinks with attendance, but it also has practical disadvantages: it means that we really need to have an accurate count of attendees at each general meeting. Even if there's nothing on the agenda that will involve a vote besides approving the previous minutes and adjourning, we need accurate attendance numbers, because attendance will impact the next meeting. For most quorum requirements, once the quorum requirement is met, it's okay to stop counting -- with a requirement like this, a careful count is always required.

Another question is what "attendance of a meeting" actually means, and how to count it. I unfortunately haven't been carefully following the Zoom side of meetings, but my impression from comments in our August meeting was that it's fairly common for people to drop in and out. Does attendance mean the number of people present at some particular time (say, the start of the meeting, or ten minutes in)? Max concurrent attendees? Anybody who was present for any part of the meeting? It's easy enough for our current flat quorum -- at the time of any given vote, you need at least 20, and if there's any doubt you can pause then to count (or just make sure there's at least 20 yeses+nos) -- but with this kind of fractional quorum we need to compute a number to feed into calculating the next meeting's quorum, not just "yup, it's at least 20".

I worry that if we adopt this, we'll be wasting significant time trying to carefully count attendance at every meeting -- and/or dealing with concerns about whether are counts were sufficiently careful.

That brings my mind back to the question "why do we have a quorum requirement?". To my mind, quorum is a tool to make sure some tiny portion of the membership doesn't sneakily hold a meeting to get something through. I think notice requirements serve a similar role (making sure that people know about a vote to show up). The DSNC has very robust notice requirements (some mandated by city ordinances 7-221a5&6): general meetings are scheduled for the last Monday of the month in the bylaws and can only be moved by the members (bylaws section 6.1), all meetings require 2 week notice (section 6.5), and anything being voted on requires 1 week notice (section 6.5). Any sneaky faction doing something high-stakes and horribly unpopular[1] would need to provide their opponents lots of notice. To my mind, this is a sign a high quorum requirement isn't necessary for the DSNC, and our flat 20 person threshold is probably good enough without introducing the complexity of something like QX.

[1] The May special meeting about the Elmway keeps coming up in these discussions. A special meeting with 11 members does strike me as iffy, which is a big part of why we've already increased the quorum threshold to 20, from 8 at the time. I also don't think the Elmway proposal was the sort of high-stakes and horribly unpopular thing I worry about passing -- there were several meetings at which it was discussed, dozens of people signed in support[2], the meeting was announced to the membership multiple times[3], and there was only one email reply to the proposal, which didn't appear to object to sending the letter. As far as I can tell, it's not that people thought it was high-stakes and hated it -- most of the membership just didn't care enough or figured more-engaged people were dealing with it. It strikes me as very weak evidence that our existing notice requirements wouldn't keep something high-stakes and unpopular getting through.
[2] Per the letter sent to the city
[3] https://groups.google.com/g/davissquarenc/c/asm2-UdKQNw/m/MO6uhMITAQAJ


A: Membership or total attendance? [total]

If we're going to pass some QX variant, there's a question of whether to count exclusively members or count everyone in the room. Membership is the more natural thing to do, but for logistical reasons, I'd lean towards total attendance. Counting the number of members present requires identifying whether each person is a member or not. As previously mentioned, we'd need to do this for every meeting, not just the ones with significant votes, because the count factors into subsequent meetings' quorum requirements. Recent meeting notices ask people to show up 30 minutes before the meeting starts for check-in so we can accurately figure out who's a member and give people voting cards -- do we want to lock ourselves into doing something similar even for a meeting that's ~just some panel discussions? We'll presumably figure out ways to get more efficient, but still, it's administrative complexity.

Total members are much easier to count -- somebody can look around the room, count the bodies, look at the Zoom attendee count, and subtract anybody who joined both. It seems like it will be a much easier number to count. It does include non-members, which is a little odd, but I suspect that we typically have a fairly small number of non-members, so I wouldn't normally expect this to be much of a distortion. If some tiny cabal is holding a secret meeting, it does allow them to recruit non-members to fill in the numbers -- but really, is it so much easier to recruit enough non-members to show up at a meeting to be worth it? Especially since the non-members still wouldn't be allowed to vote, so depending on non-members makes it easier for any members who do show up to vote down whatever sketchiness is proposed. The theoretical risk here doesn't seem big enough to justify the complexity of identifying and counting just members.

(The total attendance proposal also doesn't exclusively consider total attendance: it still requires at least 8 DSNC members, in addition to the main fraction of recent meetings threshold. The intent there is that it ensures some basic level of DSNC member participation, while being low enough that just looking at the first couple of rows and saying "yup, I recognize those eight people as members" works -- and since it's a fixed number, it doesn't require carefully counting the total number to feed into the next meeting's quorum calculation.)

B: Median or Mean? [median]

I'd prefer to do median -- median is generally less impacted by outliers than mean. Under mean, if we have a particularly small meeting (e.g., snowstorm), the next three meetings will have a particularly small quorum, and if we have a particularly large meeting (e.g., controversial developer gives an update), the next three meetings will have a particularly large quorum. Under median, it's the attendance of the middle meeting, so we'd need two big or two small meetings to make the quorum jump around. Also, median isn't vulnerable to counting errors at particularly small or large meetings (if your median meeting has 75 people, it doesn't matter whether the weird snowstorm meeting has 10 or 50 or your giant town hall has 100 or 200), which seems like a potential improvement over mean. (Of course, attendance changes at the median meeting do shift quorum one-for-one, whereas mean is three-to-one, but my gut is this is less impactful.)

C: Fractions [reject all or 50%]

In general, I think a "normal" meeting should comfortably hit quorum. If it's hard to tell at a glance whether quorum has been reached, that's stressful for the people running the meeting -- will we be able to vote on whatever important business was scheduled? What happens if things need to be postponed? Is it worth trying to hunt down absent attendees? It's fine if the rare e.g. general meeting during a snowstorm misses quorum, and for a special meeting "can we reach quorum?" can be a legitimate gauge of interest in the subject matter, but if organizers are routinely stressed about hitting quorum at a normal-looking meeting that's IMO a sign the quorum requirement is too stringent. We have a spreadsheet of past attendance numbers -- the red cells indicate when quorum wouldn't have been met, and the yellow cells represent this "organizers would be stressed" idea by highlighting when 75% of the attendance wouldn't have been enough to hit quorum (meaning glancing out at the room it'd be hard to be confident without counting that we have quorum, especially a few minutes before the start, when people are still drifting in).

Using 100% of the past three meetings mathematically guarantees that either we continually increase attendance (seems very hard! people go on vacation and also eventually we run out of space and new people) or we will routinely fail to hit quorum. Looking at our history, we're early enough to be seeing general attendance growth, but we would still fail to hit 100% quorum six out of the past fifteen meetings, with another six close calls. (Honestly, this seems like it might introduce weird follow-on behaviors, like some people leaving meetings when quorum seems set to make it easier to meet quorum in the future, or pre-planning some meetings to have light attendance, miss quorum, and reset the bar lower.)

The argument in the voter guide compares our meeting attendance to our membership numbers. I think it's fine for meeting attendance to be much lower than membership (currently around 10%), because giving up two hours once a month is a pretty high bar, whereas we strive to make becoming a member easy. It's fair to note that this means meetings might not be representative of the organization (or, IMO more importantly, the neighborhood), but making quorum nearly impossible to meet doesn't seem like a good way to fix that, and 50, 100, and 200 meeting attendees are all not very representative of a 700 member (and growing) organization (or tens of thousands of neighbors) -- if we want more representative decision-making, we should probably be doing it with out-of-meeting votes (like board, CBA, and hopefully soon, bylaws) or shifting power towards representative bodies (like the board or city council). 

75% seems much more reasonable, but quorum would still be a source of stress. Looking at the spreadsheet, we would fail to meet quorum once and be close (and accordingly stressed) six times, which seems like a lot of stressed meetings to me.

50% quorum would have been consistently met. It would have been close (June 2026: quorum of 40, attendance of 45) one time -- that meeting really does look small, but still bigger than almost all 2025 meetings, so "stressed but met quorum" seems pretty reasonable.

My general inclination is, again, that the QX amendments are too complicated to implement for the (non-)problem they solve, so I'd prefer not to pass any of them. If we are passing any, though, I think 50% is the only threshold that doesn't seem like it'll introduce significant stress and unpredictability for the organizers. 75% seems stressful but sorta functional. 100% seems like it'll make missing quorum a routine thing.

Anyway, those are my thoughts -- anybody else want to share?

~~Alex

Colin McMillen

unread,
Sep 27, 2026, 11:01:25 PMSep 27
to Alex Dehnert, dsnc-di...@googlegroups.com
I am also against any proportional amendment, especially now that the baseline has been raised to 20 members. As you mention, the existing notice requirements already make it pretty hard for a small group of people to ram through an unpopular proposal without the active consent of the voting membership.

However, re the first choice -- total attendance or only the subset who are members -- I think "membership only" is the only feasible option. There's too many reasons why attendance of non-members might spike for a given monthly meeting, and it feels wrong that this could change the threshold required for quorum months down the line. Just thinking of a few possible innocuous situations that could plausibly occur:

1) A Somerville High School class decides to require (or give extra credit) for attending a local civic meeting and reporting on what they observed. As a result, 100 non-member students dial in to our Zoom call some random Monday night.

2) Some largeish local org, like the Somerville employees union or Somerville Police's union, wins some labor concession that we argued in support of, and a few dozen members of that union show up at a meeting to give their thanks.

3) A couple dozen candidates choose to run for open at-large city council & mayor seats, and show up to our meeting to see what we're interested in and mingle with us socially afterward.

In all of these cases, I wouldn't want to turn away the non-members (they have legitimate reasons to attend, even if they can't vote), but I also wouldn't want that to count *against* DSNC being able to do business in future meetings.

There's a couple less-benign reasons why "total attendance" is a bad idea:

4) Let's say it's July and we know a contentious vote is likely to be held in September's meeting (maybe a developer plans to give an update on a major proposal in early September.) People who would like the default of "no action" decide to enlist 200 people to show up to the *August* meeting -- whether they are valid voting members or not -- for the explicit reason of making it impossible for DSNC to conduct business for the next 3 months. Essentially, the mere presence of unaffiliated people at a meeting can be used as a "pocket veto" for any future business.

5) The same situation but swapped: a *developer* feels like DSNC is likely to make an official statement *against* a proposed development, or to require significant concessions, so they spike the vote by paying 200 people to show up the month before so that we can't vote *against* their proposal either.

I agree that "voting attendance" is harder to track than total attendance, but it's not *that* much harder: just count the number of "+" roles handed out on Zoom plus the number of voting cards handed out to in-person members.

(The logistical overhead of "do we count peak attendance? Attendance when the meeting starts?" etc are problems for both approaches, and honestly for *either* approach the overhead involved is not worth the cost, in my opinion. But if we're going to have some overhead anyways, we might as well pick the approach that's robust to random fluctuations in attendance by non-members and also prevents non-members from being able to intentionally interfere with DSNC business.)
--
You received this message because you are subscribed to the Google Groups "DSNC Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsnc-discussi...@googlegroups.com.

Colin McMillen

unread,
Sep 27, 2026, 11:06:21 PMSep 27
to Alex Dehnert, dsnc-di...@googlegroups.com
Perhaps implicit in my argument, but I might as well make it explicit too:

If we are going to do this, I agree with Alex that "median" would be a better option than "mean", for being robust to one-off outliers, and a threshold of "50%" is the only one that wouldn't put undue stress of needing to stress about quorum for every meeting (and possibly, as Alex mentions, the potential need to re-assess quorum constantly throughout a meeting if people wander away from Zoom, step out of the physical room, etc.)

Alex Dehnert

unread,
Sep 27, 2026, 11:42:58 PMSep 27
to dsnc-di...@googlegroups.com

For 1-3, a weirdly popular one-off meeting wouldn't have any real impact if we're using median (we'd use the higher of the two normal meetings). (If we do mean, this matters more. If 100 bonus non-members show up, under mean, we're looking at a 33 person higher quorum.) Administratively, member attendance also means that we'd need to figure out which of those people are members and which aren't -- and for e.g. students dialing in, my guess is a bunch of them are going to be at least eligible...

For 4-5, it's true there's a potential attack here. Recruiting large numbers of people to show up seems hard/expensive, and I'm not really sure it's that much easier to recruit a bunch of randoms than a bunch of people who live in the region. Performing this attack would also look really bad -- turning the news story from "DSNC votes for to condemn project" into "developer pays hundreds of people to immobilize the DSNC" doesn't seem like a good luck.

Dunno, I do get the concerns and membership does feel more right, I just think it's not a big enough practical difference on "when do we reach quorum" to be worth the administrative hassle.

I agree that "voting attendance" is harder to track than total attendance, but it's not *that* much harder: just count the number of "+" roles handed out on Zoom plus the number of voting cards handed out to in-person members.

Until ~July, we didn't do "+" roles or voting cards -- my guess is that if we don't have much to vote on and don't need to track it for quorum, we'd go back to that. Being able to do that is the real win for total membership, I think -- it's not "no need to count the voting cards", it's "we don't even need to hand those out, we can just wave at people as they go by". I think there's a big admin hassle gap between doing nothing for checking people in (or counting people in the physical&Zoom room) and being rigorous enough about membership checking to know how many members attended.

~~Alex

Penelope Doherty

unread,
Sep 28, 2026, 7:22:15 AMSep 28
to Alex Dehnert, dsnc-di...@googlegroups.com
Hi Alex- I think my remaining question is: if checking verified voters isn't worth the administrative hassle, why do we go through the hassle of verifying voters to begin with? 

Not to open a hornet's nest, but in most models a board is elected for a reason - to vote on issues for the membership, not in addition to. Then elections are held for the few special items that need a full membership vote.  

Is there a reason we're not considering that democratic model? Seems like it would be a solution to all these calculations. A thought. 

P


From: dsnc-di...@googlegroups.com <dsnc-di...@googlegroups.com> on behalf of Alex Dehnert <al...@dehnerts.com>
Sent: Sunday, 27 September 2026 23:42:52
To: dsnc-di...@googlegroups.com <dsnc-di...@googlegroups.com>
Subject: Re: [DSNC-Discussion] Amendments tomorrow: QX (quorum based on past attendance)
 

PJ Santos

unread,
Sep 28, 2026, 8:56:42 AMSep 28
to Penelope Doherty, Alex Dehnert, DSNC Discussion
Hi Penelope,
A big reason why we need to do all this now is for bylaw amendments. The three major things we require member votes on are electing the board, amending the bylaws, and passing CBAs. 
Historically, we've asked membership to vote on smaller things (sending various letters), though I'm coming around to the view of that process adds a lot of chaos and is less representative than having the board vote on stuff.
Best, 
PJ 

Reply all
Reply to author
Forward
0 new messages