Originally Posted by
sgopal2
Sorry for being unclear. I didn't mean to say that 3 year spend is unimportant. It is clearly important, for both first time GS and re-qualifiers. However the data shows that the 3 year avg PQD is considerably lower for new GS invites as compared to requalifiers:
New GS (3 year Running Avg) PQD: $36,347
Requalifier GS (3 year Running Avg) PQD: $54,324
This is statistically significant at the 5% level. And the difference is huge, almost $20K. So this clearly means that they are more forgiving for first time GS in the 3 year spend. But it also means that once you are invited to GS, they have higher expectations on maintaining that PQD spend to requalify.
I appreciate your efforts and I don't mean to be blunt, but despite references to P values, t-statistics &etc., your method is statistically invalid. The averages you are using
are not the average of what it took to get in to GS; they are the averages of people who happened to get in -- which is a very different thing. Just a few people with big surpluses over what was actually required to get in (say $80K or 100K PQD), easily skews the average higher. Likewise, if there is a large contingent who just barely made the requirement, the average would be skewed lower. Either way, they are useless averages.
This is akin to trying to compare the freezing points for different liquids by averaging the temperatures of different samples you happen to find in a liquid state. All you know is that they are above the freezing point; but you have no idea by how much.