Wednesday, April 24, 2019

Contest Idea: Can You Write an Philosophical Argument That Convinces Research Participants to Give Some of Their Bonus Money to Charity?

In a series of studies supported by The Life You Can Save, Chris McVey and I have been showing research participants (mTurk workers) philosophical arguments for charitable giving. Other participants read narratives about children who were helped by charitable donations or (as a control condition) they read a middle-school physics textbook discussion of energy.

We then ask participants their attitudes about charitable giving and follow up with this question:

Upon completion of this study, 10% of participants will receive an additional $10. You have the option to donate some portion of this $10 to your choice among six well-known charities that have been shown to effectively fight suffering due to extreme poverty. If you are one of the recipients of the additional $10, the portion you decide to keep will appear as a bonus credited to your Mechanical Turk worker account, and the portion you decide to donate will be given to the charity you pick from the list below.

Note: You must pass the comprehension questions and show no signs of suspicious responding to receive the $10.  Receipt of the $10 is NOT conditional, however, on your attitudes toward charity, expressed on the previous page, nor on how much you choose to donate if you receive the $10.

If you are one of the recipients of the additional $10, how much of your additional $10 would you like to donate?

[response options are in dollar intervals from $0 to $10, followed by a list of six charities to choose among]

Our November 21 blog post "Narrative but Not Philosophical Argument Motivates Giving to Charity" describes some of our results. Short version: When presented with the narratives, participants choose to donate on average about $4.50 of their possible bonus. When presented with the physics text or the argument, they donate about a dollar less. We've tried varying the argument, to see if we can find a variation that statistically beats the control (with 100-200 participants per condition), but so far no luck.

This is where you come in. Maybe Chris and I are bad at writing convincing arguments! (Well, one argument we adapted from Matthew Lindauer and collaborators, in consultation with Peter Singer.) The philosophical community might be able to help us create a more effective argument.

So -- is this too goofy? -- I'm thinking that a contest might be fun. Write a philosophical argument (300-400 words) that actually leads mTurk participants to donate more of their bonus to charity than they do in the control condition. The prize might be $500 outright plus $500 to the winner's choice of an effective charity. If no one can create an argument that can beat the control condition, no winner; otherwise the winner is the author of the argument that generates the highest mean donation.

There would need to be some constraints: no use of narrative (personal or historical), no discussion of individual people who might be helped, no pictures, no highly emotionally charged content or vivid sensory detail. The argument shouldn't be obviously fallacious, foolish, or absurd. It ought to be something that a thoughtful philosopher could get behind as a reasonable argument. Statistics, empirical details, evidence of overall effectiveness, etc., are fine.

I'm open to suggestions about how best to administer such a contest, if I can find funding for it -- including thoughts about rules, parameters, the best statistical approach, what the prize should be, what to do if we receive too many submissions to run them all, etc. (I'm also open for funders to volunteer.)

Also, I'm definitely open to ideas about what features of an argument might make it effective or ineffective among ordinary readers, if you have thoughts about that but don't feel game to write up an argument.

[image source]


Drake Thomas said...

This sounds very fun (and useful)! Providing exact text of the arguments used in the previous study would probably be a good idea, so competitors can see what was and wasn’t effective.

(Sadly, I don’t think I would be a very good competitor personally - as a person who was very quickly convinced by philosophical arguments for effective altruism, I don’t think I model folks who do not think like this very well.)

Callan said...

I'm inclined to think people working on mturk aught to be the recipients of charity more than the ones giving. I'm not sure of the ethics of the matter as this money probably actually matters to them.

Eric Schwitzgebel said...

Thanks for the comments, folks!

Drake: I'm debating about that. One disadvantage would be that it might create and anchoring effect or a comparison that others were reacting to, so we wouldn't get as natural a spread of contributions.

Callan: mTurk participants vary substantially in income. (So far, we have limited participants to the U.S., for a couple of reasons.) Since we are giving 10% of them a $10 bonus and the average participant is keeping more than half, arguably they are receiving something analogous to charity from us.

Unknown said...

If you are not a mother fucker then you will donate all 10$ to charities.
I am not a motherfucker
Therefore, I will donate all 10$ to charities. There is your winner of the contest. I need 10,000$. I am thankful in anticipation. Regards qulli

Unknown said...

On a serious note:My argument for the contest. And I will donate if I win.
I am not such a person who believes in inequality between human beings; like, blacks are inferior to whites, women are inferior to men,slavery is natural etc. And who never helps other people through charity.
If I am not such a person who believes in inequality between human beings; like, blacks are inferior to whites, women are inferior to men,slavery is natural etc., who never helps other people even when he get extra money then I am such a person who believes in equality of all human beings and help my fellows beings who are in need of charity, so, I will donate half my extra money this time.

Ramiro said...

Why to proscribe fallacies? Either a fallacy tends to score worse than sound arguments, or it'd prove that soundness doesn't correlate with persuasion.

Anonymous said...

Wow this Is really coil!
How can I submit the argument?

And yes, I also think that provide a list of used arguments could be useful. True that the list could be and anchor, but It also could push to "keep thinking" till something original Will appear. Scraps are essential for creative process!

Thank you.

Anonymous said...

You have been randomly select for an additonal survey. We have set aside $10 for charity. Here is recipient list (which includes self). For each dollar you choose not to donate please list a charity you would donate to that is not listed on our recipient list. Now that you have selected preferred destinations for those dollars you have set aside for your self, would you like to reallocate those dollars to those charities?

Tom Burns said...

The winner shouldn't be the one who writes the argument which generates the highest _mean_ donation. This is because without actual statistical analysis a noisy sample could lead to a less reliably higher-donation-generating argument being chosen as the winner. At the very least, you should run some basic t-tests to compare those who read an argument and the control. Happy to provide any further technical advice on this if the project goes ahead and you would like advice/help.

Eric Schwitzgebel said...

Thanks for the continuing comments, folks!

Unknown: interesting argument! I haven’t found funding or decided the exact rules yet, but that’s a different approach than the arguments I’ve tried so far.

Ramiro: That would be interesting to know. Maybe fallacious arguments should be permitted. On the other hand, part of the idea is to find an argument that a philosopher could present to people without embarrassment.

Anon1: I’ll post on this blog when and if it happens. I’ll think more about whether the post the old arguments.

Anon2: Interesting start but not yet a fully structured argument.

Tom: Yes, of course we will run stats. The previous arguments were tested with ANOVA and Tukey posthoc and/or with two-sample t.

Unknown said...

Hi Eric,

I am fascinated by this. My initial thoughts within your constraints were what would happen if you merely quoted the results of the prior study back prior to running it again with no other differences. My guess is that anchoring would occur at $4.50 with few donations falling "below average" and that a new, higher average would then result. The technique could then be repeated on the next group. I'd be interested to know the final stable value.

What would happen if you asked three groups of participants to give Group 1: their best philosophical arguments; Group 2: their best narrative arguments; and Group 3: both, and then presented them with the same choice and asked how much of the bonus money they, themselves, would give? What is these groups formed small groups and shared their thoughts with others?

As for my own argument, I have a hard time distinguishing narrative from non-narrative arguments, possibly because of my profession as a lawyer, where I both have to convince judges, convince clients (to hire me), and convince employees of my sound judgment as a leader. However, this is the best I have: I'd ask people to imagine what they would spend a bonus $10 on, fast forward 10 days, and then ask themselves if they would be any happier from that purchase. Then I would say "Imagine you are starving and don't have any money. It only takes $1/day to feed you. Now fast forward 10 days. How do you feel after 10 days of not eating?"