4 ms·

From an amazon review: This book presents the thesis that many ills of today's society (obesity, mental illness, rates of drug abuse, ...) can be attributed to

From an amazon review: This book presents the thesis that many ills of today's society (obesity, mental illness, rates of drug abuse, ...) can be attributed to large income inequality. The authors make this point using two-dimensional scatter plots, with income inequality on the x-axis, the prevalence of some form of social ill on the y-axis, and dots in the plot representing individual countries. These plots generally show a positive correlation between income inequality and various social ills. As a statistician, I would like to comment on the soundness such argumentation: unfortunately virtually all graphs are plagued be a confusion of correlation with causation. Reviewer then explains correlation and causation. I could claim "inequality is good" using the same methods. I'd argue inequality is immaterial but does tend to be correlated with poverty, which is what we should focus on. Not to be outright dismissive (I should read the book), but this is a problem in social sciences in general. I think software simulations would be a lot more useful at this point than scatterplots comparing two variables used by so many economists today. I think I've read about people starting to do that in various areas, not sure about how prevalent it is in economic modelling. AFAIK not very.

mattgibson 11y ago
Yeah that's a fair criticism, although as someone who has spent 10 years teaching social science, I have to say that was at the forefront pf my mind when reading it. I was still mightily impressed with their data. The range of studies is extremely broad (into the hundreds), and there is a very wide range of experimental procedures in use. Not all are correlations. >inequality is immaterial but does tend to be correlated with poverty This point, they explicitly refute with a vast array of data. In cases where there is both poverty an inequality, there is a problem. When poverty alone, the problems are massively reduced. I read the book with a critical eye. I've taught social science for the best part of 10 years and this was not lacking in statistical rigour. Other academics in the field who I've discussed it with agree. I think you'll enjoy the book. It challenged a lot of my preconceptions to the point where I changed my mind, which is was unexpectedly refreshing.
wfo 11y ago
Causation does not exist in science. Causation only exists in theories that reside in our head. If you want to rely on evidence, and you want the notion of causality, you're in trouble because you can't have both, see the entire history of western philosophy in particular David Hume.
- cbd1984 11y ago
  > Causation does not exist in science. This is so over-simplified it's wrong. If I drop a weight, what happens? Is there a causal chain there? Would an animal unable to perceive a causal chain be able to live very long?
  - wfo 11y ago
    Okay, show me an experiment that verifies our intuitive notion of causation. Cause is a philosophical principle that exists like I said only in our heads and it's a question about the future. If I do this, this will happen. Science is about evidence, experiments, and the past: we did this and this happened. If you are going to be doing science, you will not prove causal links. You will develop a theory which explains some phenomenon and implies some predictive assertions and then provide evidence for that theory and test the assertions. Cause is much deeper than science can currently tread.
    - cbd1984 11y ago
      You've defined causation in a very weird way that excludes what you think science covers, and I don't know why you did that.

mattgibson 11y ago

Yeah that's a fair criticism, although as someone who has spent 10 years teaching social science, I have to say that was at the forefront pf my mind when reading it. I was still mightily impressed with their data. The range of studies is extremely broad (into the hundreds), and there is a very wide range of experimental procedures in use. Not all are correlations. >inequality is immaterial but does tend to be correlated with poverty This point, they explicitly refute with a vast array of data. In cases where there is both poverty an inequality, there is a problem. When poverty alone, the problems are massively reduced. I read the book with a critical eye. I've taught social science for the best part of 10 years and this was not lacking in statistical rigour. Other academics in the field who I've discussed it with agree. I think you'll enjoy the book. It challenged a lot of my preconceptions to the point where I changed my mind, which is was unexpectedly refreshing.

wfo 11y ago

Causation does not exist in science. Causation only exists in theories that reside in our head. If you want to rely on evidence, and you want the notion of causality, you're in trouble because you can't have both, see the entire history of western philosophy in particular David Hume.

cbd1984 11y ago
> Causation does not exist in science. This is so over-simplified it's wrong. If I drop a weight, what happens? Is there a causal chain there? Would an animal unable to perceive a causal chain be able to live very long?
- wfo 11y ago
  Okay, show me an experiment that verifies our intuitive notion of causation. Cause is a philosophical principle that exists like I said only in our heads and it's a question about the future. If I do this, this will happen. Science is about evidence, experiments, and the past: we did this and this happened. If you are going to be doing science, you will not prove causal links. You will develop a theory which explains some phenomenon and implies some predictive assertions and then provide evidence for that theory and test the assertions. Cause is much deeper than science can currently tread.
  - cbd1984 11y ago
    You've defined causation in a very weird way that excludes what you think science covers, and I don't know why you did that.