Director of Research at PAISRI
Woo, as the person running Buddhists in EA, really excited to see more groups like this! At this point there's enough of us (3 groups) that maybe it's time to start thinking about an EA Interfaith group. :-)
This is pretty long. Is there something like an abstract or executive summary of the post? Skimming a few of the expected places didn't feel like I was quite getting that without reading the whole thing.
True, but what you can do is have explicit values that you publicize and then ask candidates questions that assess how much they support/embody those values. Then you can reasonably say "rejected candidate because they didn't demonstrate value X" and have notes to back it up, or say "rejected because demonstrated ~X". This is harder feedback for candidates to hear, especially if X is something positive that everyone thinks they are like "hard working", but at the same time it should be made clear this isn't about what's true about the candidate, but what could be determined from their interview performance.
My vague understanding is that there's likely no legal issues with giving feedback as long as it's impartial. It's instead one of those things where lawyers reasonably advise against doing anything not required since literally anything you do exposes you to risk. Of course you could give feedback that would obviously land you in trouble, e.g. "we didn't hire you because you're [ethnicity]/[gender]/[physical attribute]", but I think most people are smart enough to give feedback of the form "we didn't hire you because legible reason X".
And it's quickly becoming legally the case that you can request not just feedback but all notes people took about you during the hiring process! Many companies use digital systems to keep notes on candidates, and the data in those systems is covered by GDPR, so candidates can make requests for data potential employers have about them in those systems (or so is my understanding; see for example this article for corroboration). Doesn't apply in the US, but does in the UK and EU.
For many of the breakdowns it would be helpful to understand the base rate in those countries to understand what the data means. For example, gender is easy enough since the base rate is usually close to 50/50, but for things like race I have no idea how many people identify as white, black, asian, etc. in each region to compare against. I realize not everything has a base rate to compare against, but for those that do having that data would really help contextualize what's going on here.
I guess I don't understand why w > x > y > z implies w - y = x - y iff w - x = y - z. Sorry if this is a standard result I've forgotten, but at first glance it's not totally obvious to me.
I didn't quite follow. What's the reasoning for claiming this?
From the definition of the four variables, the following equivalence can be deduced:w−y=x−z⟺w−x=y−z
From the definition of the four variables, the following equivalence can be deduced:
Well, I'd say we're all pragmatists whether we acknowledge it or not due to the problem of the criterion.
Not exactly based on EA org experience, but I think one of the biggest challenges orgs face is going from small enough that everyone can sit at the same table (people sometimes call these 2 pizza teams, because you can feed everyone with two pizzas; in practice the number is somewhere between 8 and 12) to medium (less than 150 people, aka the point at which you can personally know of everyone) to large.
EA orgs are most likely to face the first transition, small to medium. The big thing to know is that you'll have to find ways to take what happened and worked organically with a small team and transition it into processes since you'll no longer be able to easily achieve org-wide alignment automatically. This typically means the introduction of planning processes that make sure everyone is on the same page about what you're trying to do and making sure everyone's work is connected to the org's mission.
Exactly what will work is context dependent, but perhaps the first thing to get right is that you can't expect things to keep working the way they did when you were smaller. There is a tipping point where adding one more person breaks your ability to function the way you used to, and at that point you need to be ready to introduce more formal processes (these don't necessarily have to be extremely formal, just more formal than everyone sitting around the table talking and figuring things out without any process).
Dislike the idea. Feels like this will change the character of the site in a way that's negative. It's a bit hard to say way, but part of the vibe of this place is that it's about ideas not about people, and this will take it away from that direction, and I think have more an idea vibe than a personal brand vibe is good for what this forum is for. There's plenty of other places people can have more highly personally identifiable or warmer experience of connecting with others.
If we did this I feel like it would be trying to optimize for something that's not, in my view, the primary purpose of the forum, and thus would make this site worse at being the EA Forum than without this feature.