Hacker News Reader

Comment by poemxo | Hacker News Reader

I don't appreciate reading anecdata in response to cited findings. It cheapens the discussion. Now everyone is going to spend time writing knee-jerk responses to you.

At least the parent commenter had the grace to reply with another source instead of falling for it.

View on HN

I disagree. Were this an academic symposium I would agree. But this is the internet where folks who know how to establish causality and understand research methods and proper citing are uncommon. Fortunately, I do appreciate the author's thought and contribtions to observation data, and, tongue in cheek, as a utility monster my appreciation more than negates your lack of appreciation.

There is nothing hn doesn’t understand. There are just a lot of juniors, which is important.

Was your comment meant for me or someone else? I don't follow.

Just responding to: “folks who know how to establish causality and understand research methods and proper citing are uncommon”

There may be more folks on hn with said skills than you presume.

This may well be true in terms of absolute rate, but still satisfy "uncommon."

Were this an academic symposium on lung cancer I would agree. But this is the internet where folks who know how to establish causality and understand research methods and proper citing are uncommon. As a utility monster my appreciation reading about the author's grandma who lived 'til 90 smoking a pack a day more than negates your obsession with medical data.

The irony of this being every discussion of education on HN.

Must consume more utility, rawr.

Whose to say "cited findings" have any more value than "anecdata".

The institutions that build these national and international statistics do so with bias and goals, or without complete data. For example, how can a bureau make a national statistics on crime accurate when cities intentional report crime incorrectly to look better in statistics.

To think "cited findings" is gospel truth is naive. I know it's highly desired here, but I stand by what I'm saying. Data is lovely, but garbage in, garbage out, and most national-level data is complete garbage with an agenda or bias or naivety.

Anecdotes are not a very useful tool in discussions about generalisations. They provide little evidence aside from saying that it's a category of event that can exist. No one at any point has said citations are gospel. Just that anecdotes aren't adding much to the discussion at hand. If you've got issues with the cited data, be precise instead of casting general aspersions on academia.

Given that this is just a discussion between random strangers on an internet forum, I personally find both statistics and clear anecdotes, which GP provided, valuable in creating the richest perspective.

This isn't Proceedings of Hacker News or parliament: we're writing ephemeral internet words and trying to enrich each other.

Cited data on broad human population shows correlation at best, not causation.

I posited a cause based on the lived experiences people shared with me.

You're free to disagree with my conclusion, or to suggest an alternative cause. None of the cited data has actually done either of those things.

Then argue the methodology and data; anecdotes are great tools for sharing narratives, but a narrative based on bad data doesn't help anyone achieve good outcomes.

> Whose to say "cited findings" have any more value than "anecdata".

The history of human civilization.

> I don't appreciate reading anecdata in response to cited findings. It cheapens the discussion.

So does the linked PDF address this proposed "hopelessness" factor, or is it that once someone cites something the discussion becomes restricted to only things that have published study results?

Also, if someone were to cite https://doi.org/10.1136/bmj.k5094 on the effectiveness of parachutes, are other commenters then forbidden from citing anecdata that disagree with the findings?

The cited findings, don't refute the culture argument though, and maybe even reenforces it. In US culture being illiterate is bad, being bad at math is not looked down up on. The stats show good at reading and bad at math.

While there are STEM (Science Tech Engineering and Math) initiatives, I have yet to see one that actually includes math. You see results in science but not math.

The U.S. isn’t “bad” at math. I found some updated charts that break down the data for 2022 (the underlying data is all from OECD and NCES). White kids in the U.S. outperform non-immigrant kids from every European country except Switzerland, the Netherlands, and Estonia: https://x.com/cremieuxrecueil/status/1732244690408214720?s=4.... Though to be fair, the US and Europe are all notable behind the developed asian countries (Singapore, Japan, Korea) which may indicate a lesser cultural focus on STEM.

The parent commenter’s “source” makes no claims about race related performance whatsoever - it measures by just about everything but that, and then sorts by country. So maybe this is one of those darned reflexive knee jerk responses.

Pages 16, 32, 50 and 62 have breakdowns of mean score by ethnicity actually.

Ethnicity is way too coarse-grained to answer questions about culture and family wealth/connections. That’s lumping together a kid from an old-money family in New Haven with a kid from a trailer park in Virginia.

I'd say it's an incomplete measure, but it's far from useless given the US' continued statistically-significant disparities between ethnicity outcomes. The first step of solving a problem, etc. etc.

I do agree with the general sentiment though and think that too much research/news over the last couple decades has been exclusively ethnically segmented, given the economic segmentation that should always also be involved.

They're perpendicular questions and best triangulate the American experience in tandem.

E.g. what are outcomes for wealthy members of disadvantaged ethnicities? What are outcomes for poor members of advantaged ethnicities?

Those are interesting socioeconomic questions!

> They're perpendicular questions and best triangulate the American experience in tandem.

They are not perpendicular, which is why it's difficult to separate them. You even seem to know this intuitively, that's why they're "interesting socioeconomic questions".

They are perpendicular, in that one does not determine the other.

They may influence each other, but it's obviously physically possible to be a wealthy minority in the US.

It shows correlation, not causation. In the vein of this thread, it's not more or less useful than other anecdotes.

I posited a causation based on how the anecdotes countered the general trends in the data. I welcome counter arguments better than "I'm ignoring you because you don't have numbers".

Were I being paid to research this more deeply, I would. I'm not, and if someone doesn't like my argument, they're free to find one of their own.

Test scores improve with income among all groups, but the gaps between groups remain relatively similar at each income level.

Reply on news.ycombinator.com