3 ms·

Which Words Matter the Most

9 points by ankitoberoi 13y ago

webmaven 13y ago
I wonder what choice people would make if both programs were described with complete information on both positive and negative consequences.
swombat 13y ago
Worth reading through to the end of the article, rather than just the point about Program A and B... a surprisingly good article. I use a lot of the techniques in there in my blogging, and have done for years... they're pretty much unconscious for me by now.
infinity0 13y ago
In both cases, Program A is NOT equivalent to Program B. Program A is missing the vital information that there are 60,000 people in total.
- gabemart 13y ago
  It's implied that the participants were told in advance there were 60,000 total predicted fatalities. Regardless, the programs are not equivalent. A: 100% chance of 20,000 saved, 100% chance of 40,000 dead B: 33% chance of 60,000 saved, 66% chance of 60,000 dead If we allow 33% to equal 1/3 and 66% to equal 2/3, we can see that the expected value of the options is the same, but that does not mean the options are equally attractive. In gambling terms, the variance [0] of B is much greater than that of A. Which option is more attractive depends on the risk appetite of the chooser, which in turn should depend on their circumstances. If, for example, there has been a great catastrophe and there are only 60,000 people left alive on Earth, and the government has to choose which option to pick with the goal of ensuring the survival of the human race, A is far more attractive, because it minimizes or eliminates (to borrow from gambling terminology again) the risk of ruin [1]. If, on the other hand, the 60,000 predicted fatalities constitute the standing army of a small country that is under threat of invasion and is protected only by the size of its military, B may be preferable, as a chance at preserving the entire army may be more attractive than the certainty of reducing it in size by 2/3 and being vulnerable to invasion. But this is all a little off-topic, as I don't believe options A and B were ever meant to be equivalent. The point is options A1 and A2 are equivalent, yet the change in wording between then leads to people choosing A1 over B1 yet B2 over A2. [0] https://en.wikipedia.org/wiki/Variance https://en.wikipedia.org/wiki/Variance [1] https://en.wikipedia.org/wiki/Risk_of_ruin https://en.wikipedia.org/wiki/Risk_of_ruin
  - Dylan16807 13y ago
    When infinity0 says 'equivalent' they are talking about the scenario in which the choice is made. A and B have to have the same preconditions or the choice is meaningless. Not that A and B have the same results, which is blatantly false.
    - infinity0 13y ago
      Exactly. The participants were given extra information that are subsequently NOT given within the statement of each "Program". gabemart's point about the probabilities is correct, but it's a much minor one compared to the omission of information from the statements.
vorg 13y ago
> A typical example of this is the sentence ‘place the box on the table by the window in the bedroom’. This sentence can be interpreted in three different ways: * Pick up the box from the table, which is standing next to the window and put it in the bedroom. * Pick up the box from the table and put it next to the window in the bedroom. * Pick up the box and put it on the table that is standing next to the window in the bedroom. Only in writing. When speaking, we'd use different intonation. We'd stress the capitalized words: * place the box on the table by the window in the BEDROOM * place the box on the table by the WINDOW in the bedroom * place the box on the TABLE by the window in the bedroom

9 points by ankitoberoi 13y ago

I wonder what choice people would make if both programs were described with complete information on both positive and negative consequences.

swombat 13y ago

Worth reading through to the end of the article, rather than just the point about Program A and B... a surprisingly good article. I use a lot of the techniques in there in my blogging, and have done for years... they're pretty much unconscious for me by now.

infinity0 13y ago

In both cases, Program A is NOT equivalent to Program B. Program A is missing the vital information that there are 60,000 people in total.

gabemart 13y ago
It's implied that the participants were told in advance there were 60,000 total predicted fatalities. Regardless, the programs are not equivalent. A: 100% chance of 20,000 saved, 100% chance of 40,000 dead B: 33% chance of 60,000 saved, 66% chance of 60,000 dead If we allow 33% to equal 1/3 and 66% to equal 2/3, we can see that the expected value of the options is the same, but that does not mean the options are equally attractive. In gambling terms, the variance [0] of B is much greater than that of A. Which option is more attractive depends on the risk appetite of the chooser, which in turn should depend on their circumstances. If, for example, there has been a great catastrophe and there are only 60,000 people left alive on Earth, and the government has to choose which option to pick with the goal of ensuring the survival of the human race, A is far more attractive, because it minimizes or eliminates (to borrow from gambling terminology again) the risk of ruin [1]. If, on the other hand, the 60,000 predicted fatalities constitute the standing army of a small country that is under threat of invasion and is protected only by the size of its military, B may be preferable, as a chance at preserving the entire army may be more attractive than the certainty of reducing it in size by 2/3 and being vulnerable to invasion. But this is all a little off-topic, as I don't believe options A and B were ever meant to be equivalent. The point is options A1 and A2 are equivalent, yet the change in wording between then leads to people choosing A1 over B1 yet B2 over A2. [0] https://en.wikipedia.org/wiki/Variance https://en.wikipedia.org/wiki/Variance [1] https://en.wikipedia.org/wiki/Risk_of_ruin https://en.wikipedia.org/wiki/Risk_of_ruin
- Dylan16807 13y ago
  When infinity0 says 'equivalent' they are talking about the scenario in which the choice is made. A and B have to have the same preconditions or the choice is meaningless. Not that A and B have the same results, which is blatantly false.
  - infinity0 13y ago
    Exactly. The participants were given extra information that are subsequently NOT given within the statement of each "Program". gabemart's point about the probabilities is correct, but it's a much minor one compared to the omission of information from the statements.

vorg 13y ago

> A typical example of this is the sentence ‘place the box on the table by the window in the bedroom’. This sentence can be interpreted in three different ways: * Pick up the box from the table, which is standing next to the window and put it in the bedroom. * Pick up the box from the table and put it next to the window in the bedroom. * Pick up the box and put it on the table that is standing next to the window in the bedroom. Only in writing. When speaking, we'd use different intonation. We'd stress the capitalized words: * place the box on the table by the window in the BEDROOM * place the box on the table by the WINDOW in the bedroom * place the box on the TABLE by the window in the bedroom