PSA vs. SGC - A look at the T206 set

[frankbmd]

If all the cards were of equal value theoretically, lumping all the grades and cards together would have more statistical validity.

This is a nice overview and interesting, but may be flawed from a pure statistical point of view. Without getting into p values and questioning the percentage of and which hall of famers are in each group, incorrect conclusions could be inferred from the numbers.

In other words there could be some apples and oranges issues contained within the data presented. For example, if the higher valued cards are more likely to have been graded by PSA, then comparing average card prices by grade between SGC and PSA could be due to either higher prices for one company's slabs for the same card or a different distribution of cards within the subset and not a real difference in price for the same card.

In no way should this be interpreted as criticism of the work in terms of accuracy, but merely a caution to members in terms of drawing too many conclusions from the data presented. I would add that I am not a statistician capable of validating presentations of this type, but merely a math guy who knows that numbers can be manipulated to validate assumed conclusions.