4 ms·

It's a common tactic, attack the source of data to avoid discussing the data.

by talmand 11y ago

afarrell 11y ago
In theory, we should all be able to evaluate the validity data on our own, so this tactic is dumb. In reality, we have no practical way of doing this. So, whenever we cite a source of data, we are implicitly saying "This is a credible source for the statement is making. You can trust that they have done the work to make sure they aren't just repeating their own biases." And in the case of some organizations and people, you actually cannot say that. If the American milk board makes a claim about the health benefits of milk, it is not a fallacy to be skeptical of that claim.
- yummyfajitas 11y ago
  In reality we have no practical way to take the DOJ's data table to find the number of black on white crimes? Let me introduce you to data science. First I'll read the numnbers in row 2 column 2 and row 2 column 5. Then I'll plug them into ipython: $ ipython ... In [4]: 4091971*0.137 Out[4]: 560600.027 Do the same for row 3 col 2 and row 3 col 4: In [6]: 955800 * 0.104 Out[6]: 99403.2 Looks like the article was right after all. Whew, that was hard. Took all the skills I learned in my math Ph.D. to finish the job. Just lucky I didn't have to break out Hadoop. If I wasn't a professional statistician I wouldn't have had a prayer.
  - ceejayoz 11y ago
    That's not data science, that's middle school multiplication. Data science would be trying to find out why and putting it into context, especially if there are variables in play that aren't shown by a cursory look. Which a white supremacist has little interest in.
    - Nadya 11y ago
      Nobody wants to study why because they'd be labeled as racists and lose their job the moment any group of crybullies discovered what they were researching. So it goes largely untouched, which only does the actual racists more benefit because they can use and represent the data to fit their narrative because there is no other narrative because nobody wants to touch the data. It's a self-feeding loop.
      - AnimalMuppet 11y ago
        I'm stealing the term "crybullies". It's beautiful.
- talmand 11y ago
  There's a difference between showing data or its interpretation is possibly incorrect versus just saying we don't need to inspect the data because of who referred to it. Being skeptical is one thing I can agree with, but not wholesale dismissal of the data because one is skeptical of the source. Besides, often times the source isn't really the source. For example, if a publicly devout racist quoted FBI crime statistics, does that mean we can ignore the data because we heard it from the racist and not the FBI?

It's a common tactic, attack the source of data to avoid discussing the data.

by talmand 11y ago

It's a common tactic, attack the source of data to avoid discussing the data.

In theory, we should all be able to evaluate the validity data on our own, so this tactic is dumb. In reality, we have no practical way of doing this. So, whenever we cite a source of data, we are implicitly saying "This is a credible source for the statement is making. You can trust that they have done the work to make sure they aren't just repeating their own biases." And in the case of some organizations and people, you actually cannot say that. If the American milk board makes a claim about the health benefits of milk, it is not a fallacy to be skeptical of that claim.

yummyfajitas 11y ago
In reality we have no practical way to take the DOJ's data table to find the number of black on white crimes? Let me introduce you to data science. First I'll read the numnbers in row 2 column 2 and row 2 column 5. Then I'll plug them into ipython: $ ipython ... In [4]: 4091971*0.137 Out[4]: 560600.027 Do the same for row 3 col 2 and row 3 col 4: In [6]: 955800 * 0.104 Out[6]: 99403.2 Looks like the article was right after all. Whew, that was hard. Took all the skills I learned in my math Ph.D. to finish the job. Just lucky I didn't have to break out Hadoop. If I wasn't a professional statistician I wouldn't have had a prayer.
- ceejayoz 11y ago
  That's not data science, that's middle school multiplication. Data science would be trying to find out why and putting it into context, especially if there are variables in play that aren't shown by a cursory look. Which a white supremacist has little interest in.
  - Nadya 11y ago
    Nobody wants to study why because they'd be labeled as racists and lose their job the moment any group of crybullies discovered what they were researching. So it goes largely untouched, which only does the actual racists more benefit because they can use and represent the data to fit their narrative because there is no other narrative because nobody wants to touch the data. It's a self-feeding loop.
    - AnimalMuppet 11y ago
      I'm stealing the term "crybullies". It's beautiful.
talmand 11y ago
There's a difference between showing data or its interpretation is possibly incorrect versus just saying we don't need to inspect the data because of who referred to it. Being skeptical is one thing I can agree with, but not wholesale dismissal of the data because one is skeptical of the source. Besides, often times the source isn't really the source. For example, if a publicly devout racist quoted FBI crime statistics, does that mean we can ignore the data because we heard it from the racist and not the FBI?