4 ms·
> We show that this bias holds across multiple populations (college students, MTurk respondents, shoppers at a local mall, and participants from a large, income
by ghughes 9y ago
> We show that this bias holds across multiple populations (college students, MTurk respondents, shoppers at a local mall, and participants from a large, income-stratified online panel)
MTurk workers are incentivized to answer quickly, not truthfully.
Internet polls are bullshit.
- frereubu 9y agoI get really tired of this knee-jerk "X is bullshit" response without even bothering to see whether researchers are aware of potential issues. A simple search on Google Scholar would have shown that they are, and want to figure out what the methodological implications are. https://scholar.google.co.uk/scholar?q=mechanical+turk https://scholar.google.co.uk/scholar?q=mechanical+turk It's just as easy to argue - when you're not citing any evidence - that online respondents are more likely to be truthful because they're anonymous, and therefore their answer doesn't have an impact on their social reputation.
- matthewmacleod 9y agoKnee-jerk dismissals are more bullshit. It’s not a fucking Facebook poll, it’s a peer-reviewed multi-population study.
- coldtea 9y ago>MTurk workers are incentivized to answer quickly, not truthfully. Which doesn't prevent truthfulness if it doesn't take more time. Besides, they didn't say that they're super popular and cool -- so why'd they lie about not having a rich social life? >Internet polls are bullshit. Not really.