Sgc

[darwinbulldog]

It's an easy enough experiment to do. Just get a few hundred raw cards, randomly assign each of them to SGC or PSA, wait for them to come back, enter each of the individual grades, compare the SGC mean to the PSA mean (indepedent samples t-test), voila, you have your answer.

In practice it may reliably go in one direction for some issues and the other way for others, but I'd be interested in the main effect (i.e., who is more lenient on average).

[frankbmd]

Originally Posted by darwinbulldog

It's an easy enough experiment to do. Just get a few hundred raw cards, randomly assign each of them to SGC or PSA, wait for them to come back, enter each of the individual grades, compare the SGC mean to the PSA mean (indepedent samples t-test), voila, you have your answer.

In practice it may reliably go in one direction for some issues and the other way for others, but I'd be interested in the main effect (i.e., who is more lenient on average).

You should be able to get a federal grant to fund this experiment.

[darwinbulldog]

Originally Posted by frankbmd

You should be able to get a federal grant to fund this experiment.

Bah, I can't get federal grant money for my real research. Most of us probably already have hundreds of pre-war cards in our collections, but even if you were starting with no cards at all you could get enough together for adequate statistical power for a few thousand dollars. The average sale price of the last 25 raw pre-war cards on eBay was about $7.50.

[iowadoc77]

I think Frank should fund it

[frankbmd]

Originally Posted by iowadoc77

I think Frank should fund it

Not a problem. Send me the cards and I'll submit them with my return address.

[Dan Carson]

I understand PSA and SGC are both out sourcing now!!!

[pokerplyr80]

Originally Posted by darwinbulldog

It's an easy enough experiment to do. Just get a few hundred raw cards, randomly assign each of them to SGC or PSA, wait for them to come back, enter each of the individual grades, compare the SGC mean to the PSA mean (indepedent samples t-test), voila, you have your answer.

In practice it may reliably go in one direction for some issues and the other way for others, but I'd be interested in the main effect (i.e., who is more lenient on average).

And even better way would be to submit that same group to one, note the grades, crack out and send to the other. Assuming none were damaged in the process you would have a definitive answer.

Finding some one to pay for and take the time to do all of this would be the tough part.

[darwinbulldog]

Originally Posted by pokerplyr80

And even better way would be to submit that same group to one, note the grades, crack out and send to the other. Assuming none were damaged in the process you would have a definitive answer.

Finding some one to pay for and take the time to do all of this would be the tough part.

I was thinking it would be better to split them because they would necessarily be in worse (true) condition the 2nd time around, but the best thing would probably be to combine the two methods, so that half of the cards are graded first by PSA, half are graded first by SGC, then all are switched.

[h2oya311]

make sure the submitter doesn't have any "friends" at either PSA or SGC, although I suspect this issue might be evident at only one of the TPG's listed. It would be cool to see this same experiment done first by the average joe and then secondly with a big dealer/AH submitting the cards to see if there is any validity to the argument that the big guys get preferential grades.

Also, crossovers aren't unbiased because the one TPG can already see what the other one thought about the card (good or bad).

[pokerplyr80]

Originally Posted by darwinbulldog

I was thinking it would be better to split them because they would necessarily be in worse (true) condition the 2nd time around, but the best thing would probably be to combine the two methods, so that half of the cards are graded first by PSA, half are graded first by SGC, then all are switched.

Sounds like a plan. I'm sure we all have raw cards we could use for a group submission. Who feels like paying double grading fees for this experiment?

[bnorth]

In my opinion the best way would be buy the 10 cheapest PSA 10 cards and the cheapest 10 SGC 98 cards off eBay and crack them out and spend them back to the same company. That would show how consistent they really are.