Charles

The wallet test is pretty crappy indeed. The wording is funny with "It was not the intention of this experiment to make any particular group look bad..." Why did they use 'was' instead of 'is.' I know it's a small thing, but it does raise the question of what the author thinks this all implies. Since it 'was' the intention, does he think that it now does make a particular group look bad even though an educated reader would recognize this is a pointless experiment for anything other than entertainment? And since it is only for entertainment, why go ahead and make graphs to seperate people by ethnicity and gender? Why not by height, hair length/color, or any other number of arbitrary things. The only possible link I can think of that would be relative is income level, testing if rich or poor people are sympathetic or greed.

This all just screams hidden agenda.

Rachel

The Wallet Test is horrible, and I'm surprised and disappointed that you of all people gave it this publicity. The only reason for its dramatic results is that it ignores statistical power. The non-significant results are reported in big flashy graphs right alongside the sole significant result (white females vs. everyone else).

In the forum under the thread "Cut the crap", someone has crunched the numbers (three cheers for "Darth Credence".)

Quote: "You do not have enough statistical evidence to state that whites are more likely to return a wallet than blacks - the power of your test based on the sample size is a paltry 0.355, and the p-value is 0.2 - not even close to enough to reject the null hypothesis that there is a difference. There is statistical significance to the difference between white females and every other group, but not between white males and black males or females."

To his credit, the filmmaker immediately asked the right question about what sample size he would need. NOT to his credit, he hasn't taken down any of his pretty graphs of what he now knows is bad information -- but of course it is very dramatic and he wouldn't get half the traffic if he confined his results to that which was actually demonstrated.

Read more...

paulwesterberg

Twelve years ago when I was a poor college student with negative net worth living from minimum wage paycheck to minimum wage paycheck with mounting student loan debt I found a wallet on the street with over $100 in walking around money.

I didn't know the guy had lost the wallet, but some of my friends said he was a fellow student and told me where he lived, I ended up returning the wallet and kept a 20 for myself as a finders fee. After that my friends joked that I was 80% honest.

Part of the problem with this study is that there isn't much money in the wallet, If you dont return it then the person is out $2 and an old wallet and conversely keeping the wallet will only net you $2 so keeping it may be more laziness than greed. I suppose that the fake gift cert is supposed to create the illusion that the wallet is worth something to its owner without costing the filmmaker, but I think the results would be more interesting with a more significant sum of cash involved.

Read more...

Lis Riba

Yes, but across all categories, individuals in the Wallet Test had a better rate of return than the MTA. At least according to the Inspector General's study reported here

Charles

The wallet test is pretty crappy indeed. The wording is funny with "It was not the intention of this experiment to make any particular group look bad..." Why did they use 'was' instead of 'is.' I know it's a small thing, but it does raise the question of what the author thinks this all implies. Since it 'was' the intention, does he think that it now does make a particular group look bad even though an educated reader would recognize this is a pointless experiment for anything other than entertainment? And since it is only for entertainment, why go ahead and make graphs to seperate people by ethnicity and gender? Why not by height, hair length/color, or any other number of arbitrary things. The only possible link I can think of that would be relative is income level, testing if rich or poor people are sympathetic or greed.

This all just screams hidden agenda.

Rachel

The Wallet Test is horrible, and I'm surprised and disappointed that you of all people gave it this publicity. The only reason for its dramatic results is that it ignores statistical power. The non-significant results are reported in big flashy graphs right alongside the sole significant result (white females vs. everyone else).

In the forum under the thread "Cut the crap", someone has crunched the numbers (three cheers for "Darth Credence".)

Quote: "You do not have enough statistical evidence to state that whites are more likely to return a wallet than blacks - the power of your test based on the sample size is a paltry 0.355, and the p-value is 0.2 - not even close to enough to reject the null hypothesis that there is a difference. There is statistical significance to the difference between white females and every other group, but not between white males and black males or females."

To his credit, the filmmaker immediately asked the right question about what sample size he would need. NOT to his credit, he hasn't taken down any of his pretty graphs of what he now knows is bad information -- but of course it is very dramatic and he wouldn't get half the traffic if he confined his results to that which was actually demonstrated.

Read more...

paulwesterberg

Twelve years ago when I was a poor college student with negative net worth living from minimum wage paycheck to minimum wage paycheck with mounting student loan debt I found a wallet on the street with over $100 in walking around money.

I didn't know the guy had lost the wallet, but some of my friends said he was a fellow student and told me where he lived, I ended up returning the wallet and kept a 20 for myself as a finders fee. After that my friends joked that I was 80% honest.

Part of the problem with this study is that there isn't much money in the wallet, If you dont return it then the person is out $2 and an old wallet and conversely keeping the wallet will only net you $2 so keeping it may be more laziness than greed. I suppose that the fake gift cert is supposed to create the illusion that the wallet is worth something to its owner without costing the filmmaker, but I think the results would be more interesting with a more significant sum of cash involved.

Read more...

Lis Riba

Yes, but across all categories, individuals in the Wallet Test had a better rate of return than the MTA. At least according to the Inspector General's study reported here