Scan barcode
A review by daeus
Calling Bullshit: The Art of Skepticism in a Data-Driven World by Jevin D. West, Carl T. Bergstrom
4.0
Really glad it was written, so moving from 3 stars to 4. A good overview, but a lot of the second half was just going over data science, math, and statistics basics. Still, fun to read and a solid refresh (as a DS).
Quotes/notes
- Animals often deceive to survive. But humans (and maybe Ravens) have a mental model of the person they are trying to deceive, making it more complex.
- Sophists just try to win an argument and aren't looking for truth.
- "We impose strong social sanctions on liars... with all of these potential penalties, its often better to mislead without lying outright, this is called paultering. If I deliberately lead you to draw the wrong conclusions by saying things that are technically not untrue, I am paultering." Bill Clinton: 'it depends upon what the meaning of the word 'is' is' since there 'was' a sexual relationship.
- Implicature provides wiggle room. Implying without being explicit. "People all too often use this golf between literal meaning and implicature to bullshit."
- 'The start was a lie, the rest was bullshit.'
- "Coen suggests a rest for unclarity, if you can negate a sentence and its meaning doesn't change, its bullshit."
- "A primary objective of nonfiction authors is to appear authoritative, one good way to do this is to be correct, but that is neither necessary nor sufficient."
- "Aspects of technology or experimental procedure are blackboxed when they are so thoroughly accepted by the community that they become generally agreed upon standards and are no longer open to question.... when a machine runs efficiently, when a matter of fact is settled, one need only focus on its inputs and outputs and not on its internal complexity. Thus, paradoxically, the more science and technologies succeed, the more obscure and opaque they become."
- "A lie becomes bullshit when the speaker attempts to conceal it using various rhetorical artifices."
- 'length in feet is 100% correlated to length in inches.' 'Time into a hockey game is -100% correlated with time remaining in a hockey game.' I like those examples. .... "Identifying correlations can generate important hypothesis among other things."
- An ML model is only as good as its training data chapter. Mathiness
- Correlation is not causation chapter.
- Selection bias chapter.
- "Changing denominators wreak havoc on percentages."
- "This is Berkson's Paradox in operation, by selecting for both niceness and attractiveness, you have created a negative correlation between niceness and attractiveness among the people whom you would be willing to date."
- "We don't like being presented with graphs in which some or even most of the pattern is due to statistical artifact, and then being told by the author that some of the pattern is real, trust me."
- "Without theory, predictions based on data rely on mere correlations."
- "The more variables you add, the more training data you need.... this problem with adding additional variables is referred to as the curse of dimensionality. If you add enough variables into your black box, you will eventually find a combination of variables that performs well, but it may do so by chance. As you increase the number of variables you use to make your predictions, you need exponentially more data to distinguish true predictive capacity from luck."
- Science hypothesis testing shows if a null hypothesis is false, sometimes it can be like playing battleship. No effect is boring, but it tells you important information.
- "A single study doesn't tell you much about what the world is like, it has little value at all unless you know the rest of the literature and have a sense about how to integrate these findings with previous ones. Practicing scientists understand this, researchers don't make up their minds about a complex question based on an individual research report, but rather weigh the evidence across multiple studies and strive to understand why multiple studies often produce seemingly inconsistent results. But the popular press almost never reports the story in this way, it's boring."
- "scholars are rewarded for the reputations they accrue, and publishing is how reputations are made."
- "any scientific paper can be wrong. That is the nature of science."
- "By legitimate we mean that a paper is (1) written in good faith, (2) carried out using appropriate methodologies, and (3) taken seriously by the relevant scientific community."
- "'Lies are designed to lead away from the truth. Bullshit is produced with a gross indifference to the truth.'"
- "'At any given time, the chief source of bullshit with which you have to contend is yourself.'"
- "Keep it simple. One advantage that falsehood has over truth is that the truth is often complicated. Whereas falsehoods can be crafted to be simple."
- "Take it offline.... privately people are more open to critique."
- "Fill in knowledge gaps with alternative explanations. It is not enough to simply debunk myth, you need to replace it with an alternative account. People don't like incomplete stories in which events lack explanations."
Quotes/notes
- Animals often deceive to survive. But humans (and maybe Ravens) have a mental model of the person they are trying to deceive, making it more complex.
- Sophists just try to win an argument and aren't looking for truth.
- "We impose strong social sanctions on liars... with all of these potential penalties, its often better to mislead without lying outright, this is called paultering. If I deliberately lead you to draw the wrong conclusions by saying things that are technically not untrue, I am paultering." Bill Clinton: 'it depends upon what the meaning of the word 'is' is' since there 'was' a sexual relationship.
- Implicature provides wiggle room. Implying without being explicit. "People all too often use this golf between literal meaning and implicature to bullshit."
- 'The start was a lie, the rest was bullshit.'
- "Coen suggests a rest for unclarity, if you can negate a sentence and its meaning doesn't change, its bullshit."
- "A primary objective of nonfiction authors is to appear authoritative, one good way to do this is to be correct, but that is neither necessary nor sufficient."
- "Aspects of technology or experimental procedure are blackboxed when they are so thoroughly accepted by the community that they become generally agreed upon standards and are no longer open to question.... when a machine runs efficiently, when a matter of fact is settled, one need only focus on its inputs and outputs and not on its internal complexity. Thus, paradoxically, the more science and technologies succeed, the more obscure and opaque they become."
- "A lie becomes bullshit when the speaker attempts to conceal it using various rhetorical artifices."
- 'length in feet is 100% correlated to length in inches.' 'Time into a hockey game is -100% correlated with time remaining in a hockey game.' I like those examples. .... "Identifying correlations can generate important hypothesis among other things."
- An ML model is only as good as its training data chapter. Mathiness
- Correlation is not causation chapter.
- Selection bias chapter.
- "Changing denominators wreak havoc on percentages."
- "This is Berkson's Paradox in operation, by selecting for both niceness and attractiveness, you have created a negative correlation between niceness and attractiveness among the people whom you would be willing to date."
- "We don't like being presented with graphs in which some or even most of the pattern is due to statistical artifact, and then being told by the author that some of the pattern is real, trust me."
- "Without theory, predictions based on data rely on mere correlations."
- "The more variables you add, the more training data you need.... this problem with adding additional variables is referred to as the curse of dimensionality. If you add enough variables into your black box, you will eventually find a combination of variables that performs well, but it may do so by chance. As you increase the number of variables you use to make your predictions, you need exponentially more data to distinguish true predictive capacity from luck."
- Science hypothesis testing shows if a null hypothesis is false, sometimes it can be like playing battleship. No effect is boring, but it tells you important information.
- "A single study doesn't tell you much about what the world is like, it has little value at all unless you know the rest of the literature and have a sense about how to integrate these findings with previous ones. Practicing scientists understand this, researchers don't make up their minds about a complex question based on an individual research report, but rather weigh the evidence across multiple studies and strive to understand why multiple studies often produce seemingly inconsistent results. But the popular press almost never reports the story in this way, it's boring."
- "scholars are rewarded for the reputations they accrue, and publishing is how reputations are made."
- "any scientific paper can be wrong. That is the nature of science."
- "By legitimate we mean that a paper is (1) written in good faith, (2) carried out using appropriate methodologies, and (3) taken seriously by the relevant scientific community."
- "'Lies are designed to lead away from the truth. Bullshit is produced with a gross indifference to the truth.'"
- "'At any given time, the chief source of bullshit with which you have to contend is yourself.'"
- "Keep it simple. One advantage that falsehood has over truth is that the truth is often complicated. Whereas falsehoods can be crafted to be simple."
- "Take it offline.... privately people are more open to critique."
- "Fill in knowledge gaps with alternative explanations. It is not enough to simply debunk myth, you need to replace it with an alternative account. People don't like incomplete stories in which events lack explanations."