Comparison of NPC Crew Skill Level and their Gender... Male Bias perhaps?

[Susanna]

Warning: pointless nerdery incoming!

Six tests are done. One to see if the general ratio of men vs women being generated differs from 50%, and one for each specific skill level. To control for the 'lucky shot' (the more testst you do, the higher the odds one will be significant by sheer chance) a primitive bonferoni correction is applied to standard binomial tests. As such the alpha is .05/6=.0083. In other words, if the p-value of any of these tests is smaller than .0083, there is evidence 'something is up'. Lets go!

Q: Is there evidence that that the percentage of women being generated over all levels higher or lower than 50%?
A: No (p=.088).

Q: Is there evidence that that the percentage of women being generated in the harmless category different than 50%?
A: No (p=.672).

Q: Is there evidence that that the percentage of women being generated in the mostly harmless category different than 50%?
A: No (p=.672).

Q: Is there evidence that that the percentage of women being generated in the novice category different than 50%?
A: No (p=.480).

Q: Is there evidence that that the percentage of women being generated in the competent category different than 50%?
A: No (p=.672).

Q: Is there evidence that that the percentage of women being generated in the expert category different than 50%?
A: No (p=.672).

Q: Is there evidence that that the percentage of women being generated in the expert category different than 50%?
A: No (p=.015).

Conclusion: based on this data there is no evidence of gender differences in the generation of NPCs in EDs crew lounge. It is recommended not to get outraged over simple plots. More funding is needed.

Well review after I've gathered more data and retested the Expert category ... I'll let you know in a week or two.

 

[Ian Skippy]

Because stuff like this absolutely needs to be ridiculed.

The OP flat out states in the first post that their data is corrupted by factors they didn't account for such as 'male/androgynous' women and instead of taking steps to reduce that data corruption, went ahead and 'assumed' their gender into the male catagory.

This is not good science. This is someone looking for confirmation bias.

If they were REALLY interested in proving that there is a bias in favor of male crew, they'd take great care to reduce that kind of corruption and prove there were more confirmed male crew then confirmed female crew. Instead we get a "well I had trouble telling so I just went ahead and put them in the male catagory instead of owning up to my inability to confirm their gender".

And you know what? No one would have faulted them for having an 'undetermined' catagory. It would cut out data corruption. It would be more legitimate because there's ALWAYS going to be some sort of error that needs to be looked into for better accurate understanding of what you're looking at. This is why there are a phase 2, 3, and so on to experiments.

The more you reduce data corruption, the less you have to prove. This way people can't look at it and go, 'yeah, there's a bias here'.

The OP has tried to prove that there is not an even split between genders and wound up proving their own bias. I've already said that a kid's show proved that to get an even split in a coin toss, the toss had to be preformed over 2,500 times before they got anywhere NEAR the even split. Their sample size is only 50 systems strong. Nowhere NEAR the 2,500 minimum required to get anywhere CLOSE to proving there may or may not be a bias. If they can reach 3K and prove there is STILL an uneven bias, THEN I'll be convinced.

But posting a full on spreadsheet doesn't look like an observation to me. OP was trying to post proof they were right.

OP didnt try to prove anything, didnt claim to perform science (there wasnt even any statistics at all!). OP just presented a series of observations for FD to consider. You dont need to ridicule anything, you can get of your high horse, calm down, grab some tea and then talk to others in a more respectful manner.

 

[VR Murgle]

Oddly I was seeing the exact opposite tonight admittedly my sample size was even smaller but there was a clear bias towards female pilots across all ranks. Now this is probable just a coincidence but I do employ two female pilots, in fact I've only every hired female pilots (killed a few when I pancaked my Cutter and fired a few others for...reasons)

 

[Ziljan]

Hey!

This is interesting.

Just wanted to let you know that I've had a word with the server developers... they told me when generating the list of choices there is an equal chance that the NPC Crew member will be male or female.
There is no purposeful gender bias in 3304.

If there is a problem... then it's a bug. They said they'll continue their investigation on this one tomorrow, but they also believe the sample size is too small, as some people have pointed out above in the thread.

Ed

Thanks for the reply Ed. Can you verify whether the "mundanity" is a based on rank as well? Seems that the Expert seem to look a bit more polished and unique, both men and women. Whereas the lower ranking ones tend to be more homely and with the same smaller subset of generic voicings. Do the higher ranks just have more money for plastic surgery and cooler spacesuits?

 

[Susanna]

Unless it's randomly generated..

Pseudo-code:

procedure FetchCrew(ACount: Integer; ACrewList: TCrewList);
var
  Crew: TCrewMember;
  I: Integer;
begin
  // ACount represents a number of NPC crew to generate
  // MAX_SKILL_LEVEL is a global constant, represents the max skill level an NPC can be hired at. Can be changed if Frontier wanted to lift the highest skill level.

  // Clear ACrewList
  ACrewList.Clear;
  for I := 0 to ACount-1 do
  begin
      Randomize; //Initialize the random number generator on each iteration
      Crew := TCrewMember.Create;

      Crew.SkillLevel := Rand(MAX_SKILL_LEVEL );
      Crew.Sex := Rand(2); // Male or Female
      Crew.Name := GenerateName;
      Crew.Biography := GenerateBiography(Crew.Name, Crew.Sex);                  
      {...}

      ACrewList.Add(Crew);
  end;
end;

procedure LoadCrewLounge;
var  
   CrewList: TCrewList;
begin
  // CREW_LOUNGE_LIMIT is a constant (eg: 5)
  CrewList := TCrewList.Create;
  try
    FetchCrew(CREW_LOUNGE_LIMIT, CrewList);
    DisplayCrew(CrewList);
  finally
    CrewList.Free;
  end;
end;

... Yeah.. got carrried away; I enjoy pseudo-coding.. lol.

Edit: Writing a quick app to generate 10,000 crew, sex will be randomly selected; will then calculate distribution..

My extremely thorough research, without bias, shows:

10,000 Crew

First Pass:
Male: 4924 - 49.24%
Female: 5076 - 50.78%

Second Pass:
Male: 4969 - 49.69%
Female: 5031 - 50.31%

Third Pass:
Male: 4889 - 48.89%
Female: 5111 - 51.11%

Fourth Pass:
Male: 4922 - 49.22%
Female: 5078 - 50.78%

5 crew
First Pass:
Male: 1- 25%
Female: 4- 75%

Second Pass:
Male: 4 - 75%
Female: 1- 25%

Third Pass:
Male: 5 - 100%
Female: 0 - 0%

Fourth Pass:
Male: 1 - 25%
Female: 4- 75%

100,000 Crew (One pass cos it takes a while)
Male: 50709 - 50.709%
Female: 49625 - 49.291%

Op's Crew Count (50)
First Pass:
Male: 27 - 54%
Female: 23 - 46%

Second Pass:
Male: 20 - 40%
Female: 30 - 60%

Third Pass:
Male: 26 - 52%
Female: 24 - 48%

Fourth Pass:
Male: 21 - 42%
Female: 29 - 58%


What does this show?
I have no idea.
But it was fun to do.

Repped and most amusing!

I think it proves more data points are needed ... roll on 100. I'll gather more tomorrow.

 

[Susanna]

Maybe of interest. This is Foekje Dillema:

Spoiler

https://upload.wikimedia.org/wikipedia/commons/6/6e/Foekje_Dillema.jpg

She was a female athlete. She was 'physically' female, was registered female, raised as a girl and lived her live as a woman. She was banned for life when refusing 'gender inspection' in 1950. After her death

Her body looked female. Her face is 'mixed'. Her DNA is impossible to qualify as either male or female. The notion that gender or sex is binary is really not very well supported by any kind of evidence. We have treated it as such for cultural reasons, and it manifests itself in our language, but that doesnt mean it is true. As for gender identity, there is strong evidence in neuroscience that people who identify as belonging to the other sex show neurological responses to pheromones belonging to the sex they identify with, not the sex they are registered with. You cant fake this.

Oh - more interesting stuff! I'd rep you again but sadly it wont until I give some to others <sad> [where is it]

 

[Susanna]

RNGs can be tricky, e.g., it's surprisingly easy to screw one up that's supposed to produce non-power-of-two ranges "quickly", and if you do it in some naive ways, you can end up with more or less subtle bias.

If the crew members are down to some "procedural" rules, it may even be that there's a station somewhere producing only female crew.

Note to self - gather data from a range of stations... it might be worth starting again and logging their location this time!

 

[Rafe Zetter]

I hope you apply this same level of passion to every RNG in Elite if that is how you feel.

Come a little closer and I'll show you ;P - It's how I am all the time - some call it crazy, others; well... crazy; but the ladies who brave the crazy get a different experience.

The issue isn't the observations, but rather the conclusion drawn. An assumption made without regard for statistics is naturally going to draw criticism. Making a claim of bias with the sample sizes presented has no factual basis.

Very true - but I was referring to the FIVE PAGES of posts that got moderated after I flagged one - with some pretty personal remarks.

Back on topic (just in case Slopey is watching) Whatever you personally feel about the thread or it's conclusions - I found it quite interesting that FDev also found it worth checking out.

 

[Adhock]

OP just presented a series of observations...

A series of observations would have been "I went to X, Y, Z, and Q and noticed A, B, C, and K."

Not creating and posting a full on spreadsheet.

Worse still the OP went on to final stage of compiling all the data from the spreadsheet and consolidated it into a graph. That's the final step in an experiment. Consolidate your findings and make them clear to anyone you present your experiment to so that they can understand your findings and how you reached them.

You can sit there and lecture me on respect all you want to but riddle me this.

I am at least acknowledging the work the OP presented. Even if it's shoddy I can tell effort was put into it.

You're just dismissing it as "observations".

Which one of us is on the high horse?

 

[Ian Skippy]

A series of observations would have been "I went to X, Y, Z, and Q and noticed A, B, C, and K."

Not creating and posting a full on spreadsheet.

Worse still the OP went on to final stage of compiling all the data from the spreadsheet and consolidated it into a graph. That's the final step in an experiment. Consolidate your findings and make them clear to anyone you present your experiment to so that they can understand your findings and how you reached them.

You can sit there and lecture me on respect all you want to but riddle me this.

I am at least acknowledging the work the OP presented. Even if it's shoddy I can tell effort was put into it.

You're just dismissing it as "observations".

Which one of us is on the high horse?

Not sure what experiments you're involved with, but making a graph isn't the 'final step' where I'm from. And I added to OPs work by adding the statistics, for which she repped me. Can you stop this nonsense now, its getting a bit too silly.

 

[Susanna]

Curious how OP had trouble distinguishing some of the males from females. Even when they have a disastrously hideous potato face, they're still clearly male or female cos they use the same male or female [face-body] mesh assets as any others? Any examples? I've literally never seen one that couldn't be identified by that alone, regardless of if they had visible boobies or not...

The NPC crew images are all linked in the spreadsheet - take a look and see if you agree with my gender allocation...

 

[DNA-Decay]

LONG HAIR NOW!

Just wanted to let you know that I've had a word with the server developers... they told me when generating the list of choices there is an equal chance that the NPC Crew member will be male or female.
There is no purposeful gender bias in 3304.

If there is a problem... then it's a bug. They said they'll continue their investigation on this one tomorrow, but they also believe the sample size is too small, as some people have pointed out above in the thread.

Ed

Actually I believe Susanna has already identified in her original post where the gender bias has crept in.

Secondly, many of the female characters are so androgynous (dare I say 'blokish') that its pretty hard to tell whether they are even female at all (unless they have a certain chest characteristic which seems to be largely universal in the female character avatars)!

And that is that the strongest signal for femininity has been left out of the game.

LONG HAIR NOW!.

I've said it many times before. https://inara.cz/cmdr-logbook/3414/17739/

But in the real world - female astronauts HAVE LONG HAIR.
And also in the real world male astronauts don't have beards. (well two did, but there are also a couple of female astronauts with short hair).

Frontier made a design choice against developing long hair. Gender bias is the outcome.

LONG HAIR NOW!

Judith Resnik - sadly lost on the Challenger.

 

[CMDR Corlas]

Not sure what experiments you're involved with, but making a graph isn't the 'final step' where I'm from.

If I could rep that comment a dozen times, I would, lol.

 

[Rafe Zetter]

Oddly I was seeing the exact opposite tonight admittedly my sample size was even smaller but there was a clear bias towards female pilots across all ranks. Now this is probable just a coincidence but I do employ two female pilots, in fact I've only every hired female pilots (killed a few when I pancaked my Cutter and fired a few others for...reasons)

get a bit "handsy" did they? lol

"favoritism eh? What I wouldn't give for a "handsy" co-pilot, I've been piloting for 30 years and not once have I had a female co-pilot get "handsy" with me, not even a male pilot either! chance would be a fine thing...."

/life of Brian parody (guy hanging upside down in jail).

 

[GunnerBill]

Personally (again), I find this an annoying... [mad] Can we have a fair distribution between the male and female NPCs please and remove this blatant gender bias which is just more subliminal programming that men don't need any more of!

I find this astonishing. You are annoyed?

Personally I hadn't even noticed, maybe my male brain is too dull to be subliminally programmed. But in any case how is it now up to a 'woman' (sorry I assumed your gender) to decide how a mans brain needs to be programmed?

 

[Marklar_]

LONG HAIR NOW!

I agree. Also, I now want a zero-g fro.

 

[Rafe Zetter]

I find this astonishing. You are annoyed?

Personally I hadn't even noticed, maybe my male brain is too dull to be subliminally programmed. But in any case how is it now up to a 'woman' (sorry I assumed your gender) to decide how a mans brain needs to be programmed?

You assumed correctly.

As per a t-shirt I saw: "It's a womans world, men just live here".

Yes I'm a feminist, yes it made me laugh. In most walks of life there's usually a woman "picking up" after men, but then I understand this because I'm an elightened metrosexual.

 

[Vandaahl]

A read somewhere that random generators behave like photons in quantum mechanics, observing them changes their behavior.
Ignoring them would be the best way to deal with the problem.
Or not.

 

[SystematicChaos]

Computer pseudo-random number generators are not random anyway. They are always going to exhibit patterns, that's why they're not used for things like lotteries.

https://www.random.org/randomness/

 

[CMDR Corlas]

When people are so used to near-complete representation in media, it's very hard (maybe even impossible) to understand how under-representation, much less non-representation, can feel.

I tick basically all the boxes for "majority representation" in media. And even though intellectually I think I understood that "representation matters", I didn't really "get it" - it wasn't real as a thing - until I saw a sci-fi movie where one of the heroes was the same ethnicity as a much-loved and recently deceased-far-too-early uncle of mine. It was shocking to me how much more of an emotional impact that gave that aspect of the film.

It wasn't just another cool character in a space movie. He was more relatable. More likable. Less a film character and more... real. I could see my uncle in him. It was unexpectedly impactful, and frankly kind of a weird understanding. And that's for me as a (relatively) mature and well-adjusted adult. I've always had hero figures that I could look at think "Yeah, that could be me, in every way". I never really understood until then that I didn't really have any that were "Yeah, that could be HIM, in every way".

I can't imagine what it's like for a kid, growing up where no, or few, media hero figures that are really "them" and they're able to relate to, and look up to, as someone they could really grow into and follow.

So yeah, I'm never going to mock someone for questioning representation in media. It matters. And until you've been on the other end of it, I don't think you can fully understand why. I don't think I can fully understand it, but I appreciate the glimpse I got.

/you may now mock me for being honest and open on the internet