[wanderer2323]

Have you noticed how you dodged the request to link the related studies by bringing disproportial attention to the rest of grandparent's post in a condescending manner? I'm not judging you for it, it's normal. Just calling you out.

Also can you link what you consider to be representative studies on the topic?

[richmarr]

You're right, it was a little condescending, I was irritated by the way hueving minimised the evidence without apparently knowing anything about it.

(I'm sorry, hueving, that was immature of me.)

wanderer2323, I used to cite studies but no longer consider it a useful way to discuss topics like this on HN. Too often it descends into methodology theatre and too rarely does it result in useful dialogue.

So, with my apologies, no. I won't link to studies. If you're curious I've shared enough background in other comments for you to get started with your own research.

[sheepmullet]

> Too often it descends into methodology theatre and too rarely does it result in useful dialogue.

That is what happens when you present weak evidence.

I have yet to see any convincing studies despite it being straightforward to test in small repeatable experiments.

It is not hard to get 20 different 5 person teams to perform a complex task for a couple of weeks and see what impacts different types of diversity have.

[richmarr]

[Just to address the other points you raise because I was in a rush last time]

> I have yet to see any convincing studies despite it being straightforward to test in small repeatable experiments.

If these experiments are so "straightforward", and your resistance to these ideas is well-founded, then where are all the experiments demonstrating that there is no such effect?

Alternatively if you believe publishing bias and/or feminist conspiracy are combining to quash all this budding anti-diversity research then why don't you run one yourself and show everyone how it's done?

If you're not up for doing the actual experiment I'm happy to pass your protocol around for some feedback, maybe I could find someone to run the experiment for you. Maybe I can even help with funding.

[sheepmullet]

> Alternatively if you believe publishing bias and/or feminist conspiracy

Why does it have to be a conspiracy?

There are two major groups pushing this research:

1) Companies looking to make a buck by providing "solutions".

2) Social justice types.

Both these groups want there to be serious problems caused by a lack of diversity.

And then you look at the expressed ideologies of the researchers themselves and it is clear they also want the same outcome.

As such what makes the research that comes from these groups any different from research coming out of corporate and political think tanks?

I wouldn't blindly trust the Cato Institute so why would I blindly trust these researchers?

[richmarr]

Sure, publishing bias exists, as does confirmation bias (on both sides of any debate). Those biases do have an effect, and 'social scientists' do tend to be left-leaning in my experience.

But, those people have done the work, usually had their protocol evaluated, written up their paper, had it reviewed, and had it published (under their real name).

Hacker News critics on the other hand have (typically) not done the work, and have not tried to get their experiment funded and written up, and make sweeping judgements like "methodologically bankrupt" based on unqualified criticisms like "poorly designed" (no specifics or evidence of understanding), usually under an alias with effectively no repurcussions.

You yourself claimed that designing convincing studies without shortcomings is "straightforward", implying that the lack of convincing (in your eyes) evidence is therefore proof that the evidence base can be broadly ignored without actual study. That's rationalising, not logic.

Similarly, when challenged to provide a design yourself you change the subject. My offer is genuine by the way... our investors include some of the most influential organisations in this space. If you have a great experiment design I'm happy to put it to them and give you credit.

[sheepmullet]

> You yourself claimed that designing convincing studies without shortcomings is "straightforward"

Yes because - like I said in my other comment - studies on effective teamwork have been done for decades.

Point me towards a well cited meta-analysis on the impacts of race and gender diversity where the studies are repeatable experiments involving complex problem solving in a small team.

> Similarly, when challenged to provide a design yourself you change the subject.

Because you are not acting in good faith.

We both know designing a proper study would take at least a couple of months of work.

What I have asked from you on the other hand is to provide a link to an existing meta-analyses that supports your claims.

For someone who works in this space and claims there is significant evidence this should be a 5-10 minute exercise.

[sheepmullet]

> then where are all the experiments

That is exactly what I am asking you.

Where are all the small scale experiments that show diversity in race/gender in a team lead to better outcomes?

There are a lot of small scale studies on effective teamwork. I went through them a long time ago when I did my business degree.

I don't recall any of them suggesting diversity of race / gender were large contributors.

Business schools have been running small scale studies on effective teamwork for decades.

[richmarr]

> That is what happens when you present weak evidence.

In a perfectly rational world you would be right. That's not where we live though, is it.

[wanderer2323]

>Too often it descends into methodology theatre and too rarely does it result in useful dialogue.

That must be because your studies have glaring methodology defects, like most of the social studies these days. (It's easy to predict the things you'll see: vanishingly small sample sizes, no preregistration, data massaging, etc.)

You however are trying to get other people to update on methodologically bankrupt information and, when asked for proof, are taking the high ground "I won't link to studies, it leads to methodology theater and I'm above that". With my apologies, your opinion that 'diversity leads to improved productivity' is unsubstantiated, likely false and you delude yourself if you think you are right.

[richmarr]

> That must be because your studies have glaring methodology defects...

How delightful it would be to live in such a simple world.

Here's a study for you:

Handley, Browna, Moss-Racusinc, and Smith (2015) found that men are more likely to judge evidence of gender biases as low quality, and that effect is particularly pronounced in STEM fields

[wanderer2323]

must be good to live in such a simple world where you can expect to bring a study about gender biases into a discussion about diversity studies and have it prove or explain anything. anyway, the problems with the study you linked are glaringly obvious -- poorly designed study, exceptionally poorly designed controls, no preregistration (that I can see mentioned at least), sample sizes that make me really want to check their math -- and these are just from the abstract.

If this is what you have in mind when you speak of research, then you can believe whatever you want, really. false -> true === true after all

[richmarr]

> must be good to live in such a simple world where you can expect to bring a study about gender biases into a discussion about diversity studies and have it prove or explain anything

As I'm sure would be clear to you if you stopped to think about it... you strongly implied that studies are evaluated based on merit alone. This shows that they are not.

Does the study look at precisely this situation? No, that would be an unlikely coincidence.

Does the study demonstrate a gender bias in the evaluation of gender-related research, even amongst those knowledgable in the subjects? Yes.

Does the study suggest that STEM fields could be more susceptible to that? Yes.

> If this is what you have in mind when you speak of research...

As I've been saying from the beginning, this is about the balance of probabilities. By all means create your own study (and get funding, and get it published) that counters the findings. Until you do that, or someone else does that, or someone finds that the studies in question are fraudlent, unreproducible, or so flawed as to have literally zero value, then the balance of probability is against you.

You're welcome to subscribe to the view that any evidence less than a large-scale randomised controlled trial is worthless, but there aren't the funds for that, and may never be.

So your choice is between learning nothing, or learning 'maybe something'.

If you choose to learn nothing, or more accurately to deny 'maybe something' then I'm curious to know why.