Comment by mykowebhn | Hacker News Reader

Comment by mykowebhn | Hacker News Reader

I might be the extreme sort of skeptic, but I can't help but roll my eyes whenever I see the published results for any experiment conducted in the social sciences. What worries me even more is that when published the public can take the results and treat them as fact.

So let's take a look at what we can about this experiment:

1) A sample size of 87: Okay, better than many studies, but I still have my doubts about the statistical power for this study.

2) The adults used for this study were recruited from Amazon's Mechanical Turk: Okay, already we have a selection bias. At this pointed I'm strongly tempted to read no further.

3) I can't get any more details about the experiment itself since I don't want to pay $US40 to access the PDF.

http://sci-hub.tw/downloads/2019-05-14/30/10.3758@s13423-019...

Thanks! Here's the relevant part:

> We recruited 114 participants on Amazon Mechanical Turk based on the expected effect size for curiosity on trivia questions (as estimated via Kang et al., 2009 and Gruber et al., 2014) and how many participants we estimated we would need to exclude (e.g., due to inattention) based on previous laboratory experiments on Amazon Mechanical Turk. From this sample, we excluded 16 participants for providing the same curiosity rating on at least 90% of trials in the learning phase, and 11 for skipping 20% or more of trials.1 We included the resulting 87 participants in the analyses reported below. Participants received a monetary payment at a rate of US$10.00 hourly in exchange for their participation, and the task lasted approximately 50 min.

Take from it what you will, but I feel the title should be changed to "What curious attentive Amazon Mechanical Turk users whose curiosity rating tends to increase, who don't skip tasks, who need to earn some money, and who have 50 minutes to spare think they know--but might not--pushes them to learn more..." But I jest...

We need more abstracts like this in the social sciences.

Titles can be as long as you want. We need more titles like this in all of science.

Reply on news.ycombinator.com