Wittenberg

Members of the Ziu:

I have been working on some background changes to make our data more useful and accessible, and there's a few interesting points to note.  I thought I would report them to our legislature, and this data might be the basis for new policy changes or simply informative for the future.

There are a couple of trends to notice. 

Population and Activity Trends

First of all, our total population over the last twenty years.

(https://imx.talossa.com/i/c58ab32b-435e-4121-9ee8-63a31b0707ab.png)

As we can see, this was once much higher... I think our high-water mark was 258 citizens.  Currently, we have a population of 139 or so.  And I think it's interesting to notice that this is because of two trends:

• Three-Strikes removal of inactive citizens
• Generally decreased interest in Talossa

These are separate trends, and I think we can see that in the numbers elsewhere on Infotecă (https://talossa.com/infoteca/).

The first trend is the striking out of citizens who neither vote nor respond to a census over the course of two years.  From 2015-2016, the Three Strikes came into real effect and we saw purges of inactive citizens.  These were people who hadn't voted in years, and very few of them have ever shown any interest or returned in the years since.  I don't know if our civic life lost much when they were purged, and the current system might be a better reflection of who's really a Talossan.  So the total number of citizens shrank dramatically.

At the same time, we also can see that people were just becoming less interested and active in Talossa, generally.  In 2021, we got into a serious existential crisis... there were months when there were 10 or 20 total topics posted on all of Wittenberg, with less than 200 total posts. 

This problem is also reflected in the voting numbers.  The Jun 2021 election had only 73 voters, with 50% turnout -- and this is after nearly a hundred non-voting citizens had been removed from the rolls!  The total voters dropped precipitously over the course of a decade, landing down at double digits in 2019 for the first time in a long time.

If we look at naturalization records, we can see there are long spans with zero naturalizations from that time.  At this point, we weren't just shedding citizens who'd been inactive for a long time -- the country was shrinking in a very real way.

Those of us who were around at the time can remember this time, and how eerily quiet and dead everything was.  Very little was happening, and there was very little interest from anyone.

I think it's safe to say that we were in serious danger at this time, and we might be lucky we were even able to pull out of that nosedive.  It's easy to imagine the country losing the ability to appeal to others: a handful of hardcore citizens exchanging offices and talking into the void.  How long could that be sustained before people just gave up?  Or before someone seized effective control and just remade the place in their own image?

Fortunately, we seem to have mostly plateaued, and now our task is going to be to build our way up into the future.

The Virtuous Cycle

So then, when we look at the data, I think it's reasonable to say that strong Witt activity and naturalization numbers are a decent window into national health.  They're not perfect and not complete, but they're real.¹  What do we see when we take a look?

• Interestingly, posts and threads are almost perfectly correlated -- we might think that there could be months with a lot of discussion on just a few things, but the data shows that's not true.
• But as we might expect, Wittenberg activity correlates decently with naturalization and retention.  Terms with higher activity levels in general tend to have fewer people depart for inactivity at the following election.
• That effect doesn't reverse: there's no correlation between the number of people who get purged for inactivity and the activity of the next term.

Overall, there's a virtuous cycle visible in the data: Wittenberg activity becomes applications (not perfectly, r is like 0.4 - 0.5 cumulatively²), and then applications become naturalizations (again, not perfectly, r is about 0.4), and then naturalizations weakly feed back into Wittenberg activity (naturalizations affect future Witt topics at r = 0.31 – 0.37).

The caveat here is that these correlations are all real but not super-strong.  This is what we'd expect, because the whole thing is a little noisy.  There's no variable that tracks "hey we went crazy viral for two days but in a group with very low English proficiency," but that happened and it led to a month with sky-high applications but with a surprisingly low conversion rate.  That doesn't have much to do with anyone's choices, and the effects can be pretty big since our community is pretty small.

Conclusions

Overall, I think this tells us a few things.  First of all, this stuff isn't all random.  When we're active, more people tend to immigrate -- and when more people immigrate, it tends to make us more active.

Second of all, that means that interventions that boost any of these things will have positive consequences for the rest of them.

And because of those consequences, it also seems clear that policy decisions and governance can have a real impact on these trends.  That trend isn't immediate or perfect, and luck plays a big role.  But it's real.  When you vote, your choice matters for the future of Talossa.

I know we have some other folks who like numbers around, so I hope some of them will chime in.

Notes
1. These numbers could be "gamed," and that's important.  If we were to take them as gospel, that'd make them too easy a target to mess with (cf Goodhart's Law (https://en.wikipedia.org/wiki/Goodhart%27s_law%5B/url)).
2. "r" (formally called a correlation coefficient) measures how closely two things move together, on a scale from -1.0 to +1.0.  So if two things have an r of 1, then they happen in lockstep: if A, then B.