No, but not reading a paper doesn't mean you have to trust it; you can simply default to a skeptical stance of "I am not sure of the quality of this paper, so I

No, but not reading a paper doesn't mean you have to trust it; you can simply default to a skeptical stance of "I am not sure of the quality of this paper, so I can't really trust its results". Given the amount of crap already put out, I think that's a reasonable stance. The common faults include: - paper is good quality, but results are inflated/misunderstood by reporter - paper examines correlation, misses (seemingly obvious) other explanations. This can get so bad that you get "wet streets cause rain" results. - paper uses statistics incorrectly or in a misleading manner(example: Data says A is true 8% of the time, false 2% of the time, unknown 90% of the time. Paper concludes that "A is only determined false 2% of the time, which is technically true, but misleading") - paper uses unexpected definitions - words in common parlance but redefined for its field, definition is localized(legal, cultural), etc. A prime example is rape statistics, where the legal status of rape is different between countries, and many papers examining rape rates also use their own definition. To go back to the original point - if someone is telling you certain people are inferior because of their genetic makeup, as proven by paper X, this can be wrong on all 4 counts I mentioned - the definition of "inferior" might not match mine, the statistics might be misleading, the correlation may be spurious(especially when generalizing such complex phenomenon!), or the reports might be wildly inflated. Why would you even begin to trust such a paper, before verifying it?