2 ms·

> We recruited 180 university students who participated for extra credit (60.6% female, Mage = 25.71, SD = 6.40). Just like we have cured pretty much every dis

> We recruited 180 university students who participated for extra credit (60.6% female, Mage = 25.71, SD = 6.40). Just like we have cured pretty much every disease in studies in mice, I am sure that in the social studies college demographic (who participate in studies for extra credit) we have decoded the human psyche. Unfortunately, often the mouse studies and college student social studies translate poorly to the real world.

notafraudster 7y ago
I would think a stronger indictment of the external validity of the study is not the sample (I have no reason to believe people slightly older/younger are materially different than college students, and to the extent that wealth is a confounder I would expect the direction of confounding to attenuate the study as presented), but rather the design: the way people perceive "prizes" won from participating in a study is materially different than the way they appreciate gifts given to them by people they know and have a bond with. There's also an internal validity argument to be made that the effect size is trivial in substantive terms. 7.06 versus 7.82 on a 9 point scale of satisfaction for a mug? Who gives a shit? I would interpret both 7.06 and 7.82 as "I like this gift, it is a good gift". And also, who cares about getting a mug? If a mug is anchored at a 7-7.8 for me, then every gift I actually get from loved ones is going to be about a 500 out of 9. Study 3 is meant to deal with the social-emotional element of giving a gift, but that's the only study where people don't actually get a gift. They use MTurkers and ask them to imagine. I would completely ignore this part of the study. The more interesting takeaway here is the 1.5 point spread in anticipation between neat and sloppily wrapped gifts in Study 2. That's a way bigger effect than I'd think.

notafraudster 7y ago

I would think a stronger indictment of the external validity of the study is not the sample (I have no reason to believe people slightly older/younger are materially different than college students, and to the extent that wealth is a confounder I would expect the direction of confounding to attenuate the study as presented), but rather the design: the way people perceive "prizes" won from participating in a study is materially different than the way they appreciate gifts given to them by people they know and have a bond with. There's also an internal validity argument to be made that the effect size is trivial in substantive terms. 7.06 versus 7.82 on a 9 point scale of satisfaction for a mug? Who gives a shit? I would interpret both 7.06 and 7.82 as "I like this gift, it is a good gift". And also, who cares about getting a mug? If a mug is anchored at a 7-7.8 for me, then every gift I actually get from loved ones is going to be about a 500 out of 9. Study 3 is meant to deal with the social-emotional element of giving a gift, but that's the only study where people don't actually get a gift. They use MTurkers and ask them to imagine. I would completely ignore this part of the study. The more interesting takeaway here is the 1.5 point spread in anticipation between neat and sloppily wrapped gifts in Study 2. That's a way bigger effect than I'd think.