3 ms·

Getting income data from specific people can be hard...

by JimtheCoder 3y ago

lm28469 3y ago
Harder than publishing bogus studies for sure
- cykotic 3y ago
  We live in a world of imperfect knowledge. We look for patterns to try to determine cause/effect. The beginning process for this is to point out correlations and then further investigate. It’s not possible to account for all variables when it comes to things like suicide. It is possible and worthwhile to try to study the phenomenon as best we can.
  - lm28469 3y ago
    These thing end up on wikipedia and in dubious news sites as "Suicide is caused by high altitude", so yes it does matter a lot actually. It's poor science at best, if you can't control variables then the whole thing is next to useless, which happens a lot these days. At the end you have correlations, not causations
    - Tomte 3y ago
      You never get causation from statistical procedures, regardless how well you control for variables. You'll always have to augment statistical associations with further knowledge of the world, so you can get to causes.
    - cykotic 3y ago
      If we did science as you seem to think it should be done then very little progress will be made. You should read up on the history of progress in neuroscience. Metaphorically speaking; when you walk into a dark room one has to reach out and try to make sense of things before stumbling upon where the lamp is.
      - foxyv 3y ago
        You may want to consider, how much that bogus science has set us back. How long have we removed tonsils, ate the food pyramid into Type 2 Diabetes, measured people's skulls for criminality, and convicted people on the results of polygraph tests? Bogus science can move us forward, but also backwards.
        cykotic 3y ago
        On the whole though you would agree progress has been made, right? In a world with imperfect knowledge you do the best you can with the information you have. The response to a paper ought not be, “This paper is bogus because they didn’t control for all variables.”. If the paper is bogus then give reasons for why this is the case. Was the statistical analysis wrong? Are there reasons why this can’t be true? Is this new information worthy of further investigation? It’s sloppy thinking to say it’s a bad paper without having studied the issue and the data.
        foxyv 3y ago
        > This retrospective study used mortality data assembled by the U.S. Centers for Disease Control and Prevention (CDC). Retrospective studies are fairly rife with poor science. They are heavily limited by the data collected and can be easily manipulated by changing different controlling factors. While I wouldn't describe the paper itself as bogus. The title certainly takes a lot of liberties.

Getting income data from specific people can be hard...

lm28469 3y ago

Harder than publishing bogus studies for sure

cykotic 3y ago
We live in a world of imperfect knowledge. We look for patterns to try to determine cause/effect. The beginning process for this is to point out correlations and then further investigate. It’s not possible to account for all variables when it comes to things like suicide. It is possible and worthwhile to try to study the phenomenon as best we can.
- lm28469 3y ago
  These thing end up on wikipedia and in dubious news sites as "Suicide is caused by high altitude", so yes it does matter a lot actually. It's poor science at best, if you can't control variables then the whole thing is next to useless, which happens a lot these days. At the end you have correlations, not causations
  - Tomte 3y ago
    You never get causation from statistical procedures, regardless how well you control for variables. You'll always have to augment statistical associations with further knowledge of the world, so you can get to causes.
  - cykotic 3y ago
    If we did science as you seem to think it should be done then very little progress will be made. You should read up on the history of progress in neuroscience. Metaphorically speaking; when you walk into a dark room one has to reach out and try to make sense of things before stumbling upon where the lamp is.
    - foxyv 3y ago
      You may want to consider, how much that bogus science has set us back. How long have we removed tonsils, ate the food pyramid into Type 2 Diabetes, measured people's skulls for criminality, and convicted people on the results of polygraph tests? Bogus science can move us forward, but also backwards.
      - cykotic 3y ago
        On the whole though you would agree progress has been made, right? In a world with imperfect knowledge you do the best you can with the information you have. The response to a paper ought not be, “This paper is bogus because they didn’t control for all variables.”. If the paper is bogus then give reasons for why this is the case. Was the statistical analysis wrong? Are there reasons why this can’t be true? Is this new information worthy of further investigation? It’s sloppy thinking to say it’s a bad paper without having studied the issue and the data.
        foxyv 3y ago
        > This retrospective study used mortality data assembled by the U.S. Centers for Disease Control and Prevention (CDC). Retrospective studies are fairly rife with poor science. They are heavily limited by the data collected and can be easily manipulated by changing different controlling factors. While I wouldn't describe the paper itself as bogus. The title certainly takes a lot of liberties.