This is a great post, especially all of the technique suggestions.
Socially Required Token Disagreement: I’m especially surprised by the “drive safely” study—and it’s especially weird since “keeping America beautiful” would seem to contradict putting a big ugly sign on your front lawn. Maybe the effect wasn’t through the person’s support for vague feel-good propositions, but through their changed attitude to following requests by strangers knocking on their door.
Were all the signs placed in one neighbourhood? This could partly invalidate the test, since there could be an effect of discussing the issue with neighbours and changing opinions as a result of discussions, not solely as a result of a consistency effect. The study should have avoided this, choosing houseowners far from each other. Was it the case?
The result is surprising to me, so, like you, I looked to see if there were any theories that required less buy-in from me (in an attempt to minimize complexity).
Do these studies make sure that the actions are desirable? I can easily fit the data into a model in which people consider “work” to be a separate category of action in their head which they neither enjoy nor question.
In the sign example, a large sign could be reputational cost, or a future work to make your yard look as nice as it did, while a small and then large sign is two chances to enjoy supporting a cause. In the payment study, it could be true that people who were paid larger sums went into job-mode and performed services for money, while people paid a lesser amount looked to see if the action was actually enjoyable or not and learned they did.
This could easily be shown to be false if any of these studies pay varied sum for something very unpleasant, with the group that received the lesser amount reporting higher satisfaction.
This is a great post, especially all of the technique suggestions.
Socially Required Token Disagreement: I’m especially surprised by the “drive safely” study—and it’s especially weird since “keeping America beautiful” would seem to contradict putting a big ugly sign on your front lawn. Maybe the effect wasn’t through the person’s support for vague feel-good propositions, but through their changed attitude to following requests by strangers knocking on their door.
Maybe seeing the little signs on their neighbors’ yards made them believe signs were acceptable in their neighborhood.
Were all the signs placed in one neighbourhood? This could partly invalidate the test, since there could be an effect of discussing the issue with neighbours and changing opinions as a result of discussions, not solely as a result of a consistency effect. The study should have avoided this, choosing houseowners far from each other. Was it the case?
The result is surprising to me, so, like you, I looked to see if there were any theories that required less buy-in from me (in an attempt to minimize complexity).
Do these studies make sure that the actions are desirable? I can easily fit the data into a model in which people consider “work” to be a separate category of action in their head which they neither enjoy nor question.
In the sign example, a large sign could be reputational cost, or a future work to make your yard look as nice as it did, while a small and then large sign is two chances to enjoy supporting a cause. In the payment study, it could be true that people who were paid larger sums went into job-mode and performed services for money, while people paid a lesser amount looked to see if the action was actually enjoyable or not and learned they did.
This could easily be shown to be false if any of these studies pay varied sum for something very unpleasant, with the group that received the lesser amount reporting higher satisfaction.