Lurker > COVxy

Kazi1212 posted...

What was the context? Maybe thats what Im missing. I dont know about you, but I personally prefer to give people the benefit of the doubt of when interpreting what they say unless theres a clear pattern of saying inappropriate things. Also, as I said earlier, he didnt say women are troublesome, if he literally said that I would be with you. But like I said, it seems as if hes more saying that he had trouble with some women in the workplace because he fell in love with some of them, thats something I think a lot people who have colleagues can attest too, 15% of all married couple worked with one another at some point I believe after all.

He's saying this in relation to his earlier statements about his beliefs about women in the lab.

He's not just recalling a personal experience, he's reflecting on how his personal experience has affected his beliefs.

But yeah, totally, he's just saying that he had this tough time once.
---

=E[(x-E[x])(y-E[y])]

Kazi1212 posted...

COVxy posted...

Kazi1212 posted...

COVxy posted...

...lol.

What, you think he actually wants separate labs?

I believe his clarifications are an accurate reflection of his beliefs. He believes women in the lab are trouble because you can fall in love with them or they will fall in love with him.

Are you sure youre not being unfair in your interpretation of his wording? I didnt read it as he thinks women are trouble per se, but that either falling in love with them or a woman falling in love with him in the lab can lead to problems in the workplace, which is a true statement. It also didnt come off as him saying this is always bound to happen if men and women work together, he repeatedly uses the word I to designate its his personal experience. Seems to me he personally fell in love with some and others fell in love with him, its not all that uncommon in the workplace I would imagine.

"I did mean the part about having trouble with girls. It is true that I have fallen in love with people in the lab, and that people in the lab have fallen in love with me, and its very disruptive to the science. Its terribly important that, in the lab, people are on a level playing field. And I found these emotional entanglements made life very difficult. I mean, Im really, really sorry that I caused any offence thats awful. I certainly didnt mean I just meant to be honest, actually."

If we eliminate all the context, sure. Everything is innocuous.

If I said "I think some people are just less intelligent than other people, that's just the way the world is!"

No problem, right?

But what if that statement was preceded by a question: "do you think black people are less intelligent than other people?"

Perhaps the true underlying meaning of that sentence changes, no? Like wise, saying that love is complicated is innocuous, but when it comes to thinking about women being troublesome in the lab, then it becomes an issue.
---

=E[(x-E[x])(y-E[y])]

Kazi1212 posted...

COVxy posted...

...lol.

What, you think he actually wants separate labs?

I believe his clarifications are an accurate reflection of his beliefs. He believes women in the lab are trouble because you can fall in love with them or they will fall in love with him.
---

=E[(x-E[x])(y-E[y])]

...lol.
---

=E[(x-E[x])(y-E[y])]

Alphamon posted...

it was a joke.

Except that it wasn't, as he confirmed.

It was said in a joking manner, but it literally was he true viewpoint. He sees women in the lab as troublesome.
---

=E[(x-E[x])(y-E[y])]

Alphamon posted...

and did he really deserve the hysterical, emotional, nagging he got online?

Do you really want the message to the scientific community to be: "it's okay as a PI to discriminate against women in the selection of students because they are harder to handle?"

When i high ranking person puts forth a viewpoint that the community as a whole do not agree with, it's important for the community to make that known. Scientists did that.
---

=E[(x-E[x])(y-E[y])]

Alphamon posted...

Crucially, Hunt said, he then added the words, now seriously before going on to praise the role of women in science and in Korean society. The words now seriously make it very clear that I was making a joke, albeit a very bad one, but they were not mentioned in the first reports and I was deluged with hate mail, Hunt said.

It was a disproportionate reaction and you know it.

He was making light of his own stances given the occasion, but they were still his stances.
---

=E[(x-E[x])(y-E[y])]

Indeed.

"I did mean the part about having trouble with girls. It is true that I have fallen in love with people in the lab, and that people in the lab have fallen in love with me, and its very disruptive to the science. Its terribly important that, in the lab, people are on a level playing field. And I found these emotional entanglements made life very difficult. I mean, Im really, really sorry that I caused any offence thats awful. I certainly didnt mean I just meant to be honest, actually."

---

=E[(x-E[x])(y-E[y])]

Alphamon posted...

COVxy posted...

Alphamon posted...

hey guys remember what feminists did to tim hunt.

toxic feminism is real

I mean, he said some pretty stupid things at an event meant to honor female scientists lol.

you do realize that it was a joke that was applauded? and connie st louis lied?

toxic. feminism.

Preeetty sure he said he was being honest in later interviews.
---

=E[(x-E[x])(y-E[y])]

Alphamon posted...

hey guys remember what feminists did to tim hunt.

toxic feminism is real

I mean, he said some pretty stupid things at an event meant to honor female scientists lol.
---

=E[(x-E[x])(y-E[y])]

Anarchy_Juiblex posted...

principle of charity

What are you, a low-T cuck?
---

=E[(x-E[x])(y-E[y])]

Antifar posted...

pikachupwnage posted...

But yeah its true that male victim sexual assault is often not treated as it should be by society and that is is a larger problem than your average person likely thinks it is.

Oh definitely. See the way Bill Maher treats prison rape as a punchline, for example.

Important to note that this is feminist rhetoric, though.
---

=E[(x-E[x])(y-E[y])]

darkjedilink posted...

COVxy posted...

Mostly likely a troll made that. And even if they didn't, why are you pretending this is representative?

The worst thing is that you probably know this, but are spreading it around for "lulz" , despite the real damage this shit does.

How is it not representative, when toxic feminism outright states men can't be raped, or be the victim of sexism?

Who, exactly, has said this?

EDIT: I should clarify: besides the feminist within your mind that you angrily smash down with your righteous logic in order to make yourself feel better.
---

=E[(x-E[x])(y-E[y])]

Mostly likely a troll made that. And even if they didn't, why are you pretending this is representative?

The worst thing is that you probably know this, but are spreading it around for "lulz" , despite the real damage this shit does.
---

=E[(x-E[x])(y-E[y])]

Imagine taking Omega Hunter seriously as a poster.
---

=E[(x-E[x])(y-E[y])]

Although. real JP quote here:


lol.
"Hm. Are we sure this feminism thing isn't just PMS?"
---

=E[(x-E[x])(y-E[y])]

*any speech that isn't blatantly conservative or anti-SJW*
"This is why Trump is going to win!"
---

=E[(x-E[x])(y-E[y])]

vocedelmorte posted...

Why did you close the poll after only 35 votes anyway, whats the point ??

That's when they saw the results they wanted.
#p-hackinggamefaqssurveys
---

=E[(x-E[x])(y-E[y])]

frozenshock posted...

ledbowman posted...

Trump has admitted to obstruction of justice at least once. Junior has admitted to conspiracy. They're guilty.

These days guilt or innocence depends solely on whether you have a D or an R next to your name.

I'm pretty sure that if it had been any other president he would have been brought up on impeachment by now. It's insane how much political standards changed in 2016.
---

=E[(x-E[x])(y-E[y])]

Vyrulisse posted...

LUL. As if this is something new...

I mean, it actually is relatively new due to the large scale ability to reach people using social media.

Still, new or not, it should concern you.
---

=E[(x-E[x])(y-E[y])]

Darksaber310 posted...

This is the kind of shit that makes me laugh about the whole Russia bs. At the end of the day, it's facebook ads and we've reached the apex of panic, but when lamestream media does it for twenty years, meh, at least it's not Russia, right guys?

Let's all pretend this is all that it was, everyone!

Regardless, it absolutely should terrify you that another nation can do such a successful job in a widespread propaganda campaign to influence the election.
---

=E[(x-E[x])(y-E[y])]

DarthGravid posted...

If it happens, you won't be able to find a single person that voted for him anywhere.

P.much this.
---

=E[(x-E[x])(y-E[y])]

Balrog0 posted...

averagejoel posted...

there's a much bigger gap between far left and moderate left than there is between moderate left and far right

it's sad that you actually have convinced yourself this is true

To be fair, there's been a large scale propaganda effort to convince people of such things.
---

=E[(x-E[x])(y-E[y])]

Last bump for discussion.
---

=E[(x-E[x])(y-E[y])]

"Wheres muh vidya gayme sales? Obvs the only object that matters!"

Prime day is good if you are an adult.
---

=E[(x-E[x])(y-E[y])]

Up.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

How would there be a ceiling effect if there isn't a good standard on what a max out score is?

Many of them have definitive boundaries, and if they didn't there would be an effective boundary. If a test score doesn't discriminate between high performers, you get a heavy skew, because by definition all high performers have a similar score.
---

=E[(x-E[x])(y-E[y])]

IllegalAlien posted...

Sample statistics give different information about the sample. IDK why anyone wants the variance over say the median, but w.e.

Yeah, but not more information than seeing the entire distribution.
---

=E[(x-E[x])(y-E[y])]

I'm saying that there's no evidence that their results are due to the test being too easy because this would cause an obvious ceiling effect if this were true, but this largely isn't the case. You can assert that the tests wouldn't be predictive of later mathematical ability, but that is an entirely different argument. One that also requires evidence (and this likely exists in the literature if you wanted to look).
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

COVxy posted...

Squall28 posted...

In a class? 90%

What does a distribution of grades look if they are bunched up around 90%?

Probably normalish just like any other %. The main thing is that the distribution would be narrow.

Your intuition is incorrect, this creates an extreme skew in the data, obvious to spot.

Here's a quick example, subjects who are normally distributed around 90 with a standard deviation of 20, but scores are clipped with random error at 100, the bound. This produces data like:

---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

In a class? 90%

What does a distribution of grades look if they are bunched up around 90%?
---

=E[(x-E[x])(y-E[y])]

Sure, explain it to me.
---

=E[(x-E[x])(y-E[y])]

CreekCo posted...

COVxy posted...

thelovefist posted...

COVxy posted...

Squall28 posted...

If you are taking an easy course, almost everyone is going to get the same score.

And what would that score be?

Irrelevant.

Lol. Absolutely not irrelevant, in fact is the entire point.

Read what you wrote again, slowly.

You can take that advice, if you'd like.
---

=E[(x-E[x])(y-E[y])]

thelovefist posted...

COVxy posted...

Squall28 posted...

If you are taking an easy course, almost everyone is going to get the same score.

And what would that score be?

Irrelevant.

Lol. Absolutely not irrelevant, in fact is the entire point.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

If you are taking an easy course, almost everyone is going to get the same score.

And what would that score be?
---

=E[(x-E[x])(y-E[y])]

The_Ivory_Man posted...

JACKBUTTMOMMY posted...

Why would someone's gender determine their math ability? Or sex for that matter? I'm not sure a study was necessary.

Basically while men and women average at the same point, men have far greater spread.

This means more stupid men, as well as more smart men.

This is verifiable.

Is it? There's no evidence of tail effects in this study.

This argument has been floating around a lot more often since Peterson spouted it. Small mean differences do not imply "whopping" differences in the tails, as he argues.
---

=E[(x-E[x])(y-E[y])]

scar the 1 posted...

Ah yes, the classic "sample size is too small" argument. I wonder if this is so common because it's the only statistics term that the regular CEman knows?

To be fair, it's a nonintiutive concept that you can sample an extraordinarily small number of people in the scope of the target population and accurately estimate parameters. That said, it's clear that this is the go to: "i don't have any real criticism of the study, but I don't like it's conclusion" response.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

COVxy posted...

The standard deviation in comparison to the mean doesn't give you the information you seek, and certainly wouldn't be more informative than just looking at the distributions.

Why does it not? If there's a test, and the average is 70, and everyone scores within 5 points if that it doesn't mean much. Contrast that to an average of 70 but people fluctuate like 20 points. That's a test where you can differentiate high and low performers.

The plot distribution gives you an overall look at the data which I generally like, but I want to pinpoint something.

Simple version: the variance around the mean can be caused by many things. Essentially, this is usually a metric used for signal quality, and in fact you want the variance to be small in relation to the mean. But it's kinda nonsensical here.

There's no simple answer to what you're looking for, other than there's a clear age effect across most of these measures, meaning that the variance is indeed segregating individuals on the basis of certain factors, just not gender. If you are looking for an effect where the task is too easy or too hard, pushing everyone down or up to the same performance, this is called a floor or a ceiling effect, and it is easily observable on the basis of the distribution being skewed against a boundary. You can see this in a couple of their measures, but not most, and in those that it does, the ceiling tends to be driven by age, where the task is too easy for older kids.
---

=E[(x-E[x])(y-E[y])]

The standard deviation in comparison to the mean doesn't give you the information you seek, and certainly wouldn't be more informative than just looking at the distributions.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

I've said I already seen the plots and the deviation looks small there. I want to see what the actual numbers are per test.

And you expect that the standard deviation will be a better descriptor than just looking at the distribution? Lol.

In what way does the variance look 'small'? In comparison to what?

We know they aren't at ceiling. We know they aren't performing the same (age predicts). In many of these measures, they are getting scores the constitute a large proportion of the scale.

Idk what you mean by "variance seems small".
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

Standard deviation in the data set for boys and in the data set for girls. I want to see how much variation is actually in the data, or is everyone more or less the same as I expect.

...they plot the distributions in several ways, you can literally look at the plots and see.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

I'm having trouble finding the actual standard deviation for each data set, but based on the errors and gray areas on the graphs, the variance is low so the tests don't really discriminate between children.

In comparison to what, exactly?

In only a couple plots does a ceiling problem even rear its head a little, and even within those analyses it doesn't seem like an issue since if you mentally split the data prior to when kids are hitting ceiling, there are still no apparent gender differences.
---

=E[(x-E[x])(y-E[y])]

BignutzisBack posted...

COVxy posted...

Honestly didn't expect people to be this bent out of shape about this lol.

"I don't like people disagreeing with me"

Has nothing to do with disagreeing with me. Has everything to do with irrationally arguing agaisnt data with no real basis.
---

=E[(x-E[x])(y-E[y])]

Honestly didn't expect people to be this bent out of shape about this lol.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

One test was choose the side with more dots. Another test was count as high as you can.

And these were really young children.

For what you say to be an actual reason for the null effects, you'd see it in the data. The data would be pushing up against the upper bound (statisticese: at ceiling). The data show that these tests discriminate between children.
---

=E[(x-E[x])(y-E[y])]

Idk, the Russia scandal is pretty insane, in general. 5-10 years ago, if someone told you this story, you would have told them to pick another profession because writing realistic fiction isn't their strong suit.
---

=E[(x-E[x])(y-E[y])]

CreekCo posted...

By that logic, they could sample 10 kids and it be valid

No, no they couldn't.
---

=E[(x-E[x])(y-E[y])]

Squall28 posted...

tl;Dr It's easy to have equal performance when the test is easy

There's no evidence here that the tests were "easy".

Dragonblade01 posted...

Serious question, do we see natural discrepancies in certain abilities develop over time between genders, or is it impossible to isolate "intrinsic" discrepancies from those influenced heavily by environment?

I think the wording "intrinsic" is too strong in their title.
---

=E[(x-E[x])(y-E[y])]

CreekCo posted...

COVxy posted...

CreekCo posted...

The main graph given has a n-value of like 241. That's not even half of that.

The collected data from 500 children in total, though various analyses have differing numbers of subjects.

Regardless, each individual analysis here is actually really well powered. Your complaint about sample size is entirely unfounded and improper.

Yeah, it's maybe 500 if you add together all the data they borrowed from other people's work. There's n-values in the single digits for their work. Single digits. The headline you posted was based on a study of 241 kids and a lot of them were 4 year olds. This is no basis for any definitive conclusion other than they were all ready for snacktime after the survey. Everybody hurried up to finish so they could eat their snacks for completing the survey and watch Peppa Pig, lol.

Again, the sample sizes here are all appropriate, and you have done nothing to counter their argumentation:

The absence of statistical differences across the major developmental milestones of early mathematical cognition are unlikely to be due to sample size. Power analyses suggest that given the sizes of the samples analyzed here, we should have been able to detect small to medium effect sizes ranging from Cohens d = 0.34 to 0.65 (80% power, p = 0.05; Infant Numerosity Comparison (looking time): d = 0.65; Early Childhood Numerosity Comparison (w): d = 0.37; Recitation of Count List (How High? task): d = 0.47; Counting Principles (Give-N task): d = 0.52; Math Concepts (TEMA): d = 0.34, (Formal/Informal Math Scores): d = 0.40). Importantly, even if smaller effects do exist, they are unlikely to reliably, meaningfully, or consistently manifest in children. Caution should be taken when interpreting any small effects in large sample to ensure that their importance is not over-exaggerated.[13,75]

They are well within power to detect any effects that would matter. And their power analyses seem to be overly conservative (they seem to be on the basis of a two tailed test, when really they should be looking for a one tailed effect).

And then the only other thing you have to offer is rambling about Peppa The Pig.
---

=E[(x-E[x])(y-E[y])]

CreekCo posted...

The main graph given has a n-value of like 241. That's not even half of that.

The collected data from 500 children in total, though various analyses have differing numbers of subjects.

Regardless, each individual analysis here is actually really well powered. Your complaint about sample size is entirely unfounded and improper.
---

=E[(x-E[x])(y-E[y])]