Just because something is complex and difficult does not make it equal to religion. And I don’t actually see this as the key difference in how we understand this issue.

Quite the contrary I see you as having the religious approach and me taking the scientific one. Science is built on models and predictions. Just because you find or don’t find some kind of correlation between some variables you picked doesn’t mean you got any understanding.

I like to quote this from lifehacker as I think it is pretty good:

  • Fact: Observations about the world around us. Example: “It’s bright outside.”
  • Hypothesis: A proposed explanation for a phenomenon made as a starting point for further investigation. Example: “It’s bright outside because the sun is probably out.”
  • Theory: A well-substantiated explanation acquired through the scientific method and repeatedly tested and confirmed through observation and experimentation. Example: “When the sun is out, it tends to make it bright outside.”
  • Law: A statement based on repeated experimental observations that describes some phenomenon of nature. Proof that something happens and how it happens, but not why it happens. Example: Newton’s Law of Universal Gravitation.

In science we combine facts, formulate a hypothesis and use that to build a theory. He theory gives some sort of rational or logical explanation that ties together the facts we observe. I just don’t see you attempt to provide some kind of coherent explanation for anything. You just hack statistics.

I don’t know if you have noticed. But I go into great depth in explaining my mental model of how I think things work. I have a hypthesis and a model in mind. I state something about what I expect to see. I just don’t see that kind of analysis from you. It is all rather shallow. Which is an unrealistic characterization of how human society works. Human society simply isn’t as simple as you try to make it.

The biggest issue I’m having here is not with bias, but with your selective choice of when logic or correlation can be applied. You have some opaque ideas as to when countries can be compared to each other and when they can not.

It is better if you simply criticize our point out assumptions or ideas you disagree with. I mean your whole criticism here is kind of vague.

Homicide Rates

You suggested that the murder rates among Black Americans (~20/100,000) is higher than the average murder rates in Africa (~12/100,000). Both are higher than the white murder rates in the US (~2/100,000) and Europe (~1/100,000).

You say I am opaque, but you don’t spell out what you believe this implies or why this matters?

I can try to walk you through my thinking. I believe it is natural to see higher rate or crime among a low socio economic class anywhere regardless or skin color. Poverty and inequality are driving factors of crime. It thus makes sense that blacks are overrepresented relative to whites on the crime statistics. However what looks quite peculiar is that despite America being significantly richer than any African country, the homicide rate is quite a lot higher. Not only that but we can easily find multiple African countries with significantly lower rate.

I think it should be perfectly reasonable to compare the best of Africa against America. Why not? Why should African-Americans not be as well off on every metric as the very best of Africa? America has after all been a democracy much longer and has more stable institutions and is vastly richer.

That Black America cannot handily beat every country in Africa should set the alarm clocks off and ringing. It points to something being profoundly wrong about American society with respect to the African-American population.

You are trying to suggest there is something wrong with black people. I am saying the evidence points to something very wrong with America.

You try to make this into a relative thing. Because blacks are worse relative to whites in Britain the problem is somehow the same. By doing that you mask the enormous difference in homicide rate which exists in the US relative to the UK.

And I have noticed people like you often end up with a circular argument.

“Why is the homicide rate so high among African-Americans,” I ask.

Your response is, “The relative difference is the same as in the UK.”

So I ask, “Why is the total homicide rate so high in the US?”

To which you have seemed to imply it is because of blacks. But that is circular argumentation. You have not explanation for why the total rate is so high in the US. You flat out reject guns has anything to do with it. We are left with a mystery.

It’s quite possible that there’s something about America (either prevalance of guns or some other cultural or socioeconomic factor)

Strange concision given that have argued quite strongly repeatedly that there is no proof that guns lead to higher homicide rate. I seem more like you shift your stance depending on what you argue to make your argument stick together. Your claims about guns and black cannot easily be true at the same time.

Racism in Health Care System

And you concluded that this international comparison proves systemic racism in American hospitals.

A scientific approach means you try to combine multiple observations into a coherent whole. Stuff has to fit together. And we find just such a share amount of concrete stories that point to racism in the health care system.

This has also been studied extensively. Blind tests have been performed. Here is on particularly telling example:

When you got a whole house full of health professionals and police, who all treat a black man like this. Nobody putting their foot down or objecting. Then this isn’t just a random fluke. It points to a deeper systemic issue. The probability of everybody in that apartment just randomly acting racist is rather slim. And the statistics correlate these stories.

You compared infant mortality among black Americans (~14/100,000) against Sri Lanka and Cuba (~5/100,000).

And you seem to think this is weird and arbitrary, so let me explain the rational. Both Sri Lanka and Cuba are significantly poorer countries than the US. The US has the worlds most advance research hospitals, major drug companies, medical research etc. Thus we should under normal circumstances expect that the US should be able to handily beat these countries on health care.

Yet what we see is that African-Americans get worse care that citizens of very poor countries. Again the alarm bells should be ringing. Something is obviously wrong.

You could have compared infant morality in Africa, which is 10 times higher than among black Americans, and concluded that America does relatively well for black mothers.

No, why should you? That is a completely arbitrary choice. I am giving you a rational for my choices. What on earth would be the rational to expect African countries to do better than American on health care? These are dirt poor countries, with typically underfunded health care system, poor infrastructure. They have sever lack of medical equipment and medication. It is not without reason that in Rwanda they experiment with drone based delivery of medication, because infrastructure simply isn’t good enough to get medication out to people.

I think what you fail to see in how I make comparisons, is that I start with a model. I make a prediction A > B, based on a number of criteria such as income, education etc. When instead we find that A < B, then we know something is off.

What you so is take a situation where you would expect A > B, and then you find A > B and you conclude: Look everything in A is totally fine. Duh! That was as expected.

As far as infant mortality is concerned, the white vs. native gaps in Canada are similar in size to the white/black gaps in America.

And that suggests a serious problem which should be looked into. That in no way gets America off the hook. There are a number of questions I would naturally ask from this:

  1. Thus this include natives on reservations? Do reservations have access to proper health care?
  2. What is the health situation for indigenous people in other countries. How does it compare.

Basically, I have no idea what your rules are. You think that comparing health care between black Americans and Cubans and Sri Lankans is a valid argument.

Because you can ahead of time make a prediction of what you should expect given wealth and development level. What is interesting here is that outcomes differ from what simple predictions tied to wealth and development should suggest.

But if I present a graph of murder rates that includes Bulgaria, Moldova, and Iceland, among many other European data points, my argument is invalid.

The issue isn’t what countries you include, but how you compare them. You attempt a linear regression. I did not do that with Sri Lanka, Cubans and Americans. See the big difference?

Image for post
Image for post

However you can make a prediction about Bulgaria, Moldova and Iceland. Then you can test that prediction. E.g. I can make a prediction that because Iceland is richer, is more developed, has longer democratic traditions, has higher levels of equality and has low population density, it should have low homicide rate than Moldova and Bulgaria, because the latter are poorer, has more inequality, higher population density etc.

Looking at the graph, I see nothing unusual in that regard. My prediction checks out. Things are as the should be expected between those countries.

However if I am to do a linear regression against two variables, I naturally cannot have ton of other variable that vary too much. The effects of gun ownership could be lost because other factors dominate more. Hence we must minimize the difference between these other factors to tease out whether number of guns matter.

On the graph you objected to, Bulgaria’s data point was close to the Netherlands. Moldova was close to Portugal.

Again you should make some predictions about what you would normally expect. The Netherlands should most likely be higher than Iceland due to denser population, larger cities etc.

However we should expect the Netherlands to be lower than Bulgaria and Moldova as it is more developed. Moldova has GDP at about 1/3rd of Bulgaria we should expect them to have worse statistics than Bulgaria. That also checks out.

While e.g. Portugal is a bit of a mystery. It is higher than the Netherlands, which makes sense since the Netherlands is richer. Yet it is much higher than Bulgaria despite being more developed. That is naturally something that requires some kind of explanation and further investigation.

Is it valid to compare the Netherlands to Iceland? Portugal to Iceland? Portugal to Spain?

As I said, it depends on what sort of comparison you do. Notice how you just throw out countries, without any kind of analysis. While I actually provide a long list of reasons for why I expect or don’t expect what I see?

I don’t understand your rules here. And I don’t know how I’m supposed to have a logical discussion when you arbitrarily choose which international comparisons are allowed.

As I hope I have explained with numerous examples and by going through my reasoning in detail, it should be clear that the comparisons are not arbitrary. It is not merely a question of which countries you compare but also what sort of comparison you attempt.

Correct Use of Linear Models

You’ve likewise decided that linear models are an unacceptable means of measuring which variables affect educational outcomes.

No, my argument was more nuanced than that. My argument is that educational outcomes is a complex model. Not a linear one.

However if one factor has a very strong influence overshadowing other factors then a linear model will be able to pick up a correlation.

My issue is that you assume that just because you cannot pick up something, there is no relation. All that tells you is that either, there is no correlation as you assume, or the factor you are investigating is simply not powerful enough to overshadow other factors.

Basically you don’t distinguish between the logical expression A ⇒ B and A ⇔ B. You can use a linear regression to discover a relation, but you cannot use it to disprove a relation when you deal with a complex model.

And… I really don’t know where to go from there.

Real scientists do far more complicated and time consuming analysis. You cannot complain because you cannot whip up a comparison in an hour.

Looking for correlations is a standard technique in sociology and criminology, you start by finding correlations to find which variables are important and then work towards proving causation.

Yes, it is a useful starting point to begin an investgation. I never said anything different. My issue is that you use what is basically a preliminary investigation to jump to conclusions. Since you find no correlation it would be natural to assume that many factors influence the outcome and hence to tease out whether one particular factor has any influence, you must even out the other factors somehow. Something you never did.

One correlation alone does not prove causation. But lack of correlation is a pretty good hint that two variables are not related.

Not for a complex model. That simply suggest further examination must be done.

As I’ve said before, this feels very much like having a discussion about religion.

To me this feel like having a discussion with somebody who is playing scientist, but who doesn’t appreciate that there is a reason why it is a profession and why studies require lots of funding. Doing proper analysis is hard and difficult work that takes time.

Can You Measure Racism?

This reminds of your argument — “we can not measure racism in simple linear models, it is a mysterious force, only detectable by more complex arguments”.

But that isn’t even true. I have provided numerous examples through over back and forth that backs up my claims with statistics.

  1. Guns and drugs are more likely to be found among whites searched by police, pointing to racial profiling.
  2. There is a clear statistical anomaly in the amount of drugs reported on black drug offenders, which don’t exist to the same degree for whites.
  3. Cops don’t stop blacks at the same rate when they cannot tell their skin color. This has been thoroughly examined.
  4. When people guess whether a person holds a gun, cell phone or something else, studies show people are more likely to assume the black person holds a gun.
  5. Test with identical resumes where race is swapped on applications show that a white person is more likely to be called for an interview. In fact they can even add a criminal record to the white person and they will still be more likely to be called for an interview.

There is a multitude of statistics like that and we have decades worth of stories about racism that all African-Americans can share with us. All this data reinforce each other and paints a clear picture.

Yet you seem to think you can settle the issue with simple linear regression you whip up in an afternoon. You are the one doing voodoo here, not me.

“Racism of the gaps” is the argument to explain the differences that remain. Whatever gaps exist must be there because of racism, because you’ve adopted a faith based position that there are no significant differences in genes.

No, again you turn it on its head. Basically your article of faith is that racism doesn’t really exist and that the left’s only argument for racism is by pointing at difference of outcomes.

But I have repeatedly in this discussion listed countless examples of how there are real measurable differences in starting conditions and how racism is very much a real thing.

Gaps could come from some unmeasured form of systemic racism, like stereotype threat. But they could also come from genes. Or from the culture of each racial group.

Yes, and my issue with your position is that you are too dismissive of anything that isn’t based on genetics. I don’t rule out genetics, I just find it implausible. I don’t think that that is an unreasonable assumption, especially when we know their is a strong political agenda to prove such a connection. And it has been that way for over 200 years.

Written by

Geek dad, living in Oslo, Norway with passion for UX, Julia programming, science, teaching, reading and writing.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store