Which one dominates the performance upward bias?
Survivorship bias for sure (by 1.5% - 3.0% a year) but backfill bias can also do it too bc in HFs, it’s voluntary reporting and you’re not going to report your old performance numbers unless they are good! In magnitute, backfill bias I *think* would be less as a percentage contribution to the industry published performance data.
Thanks…does L3 pass rate have survivorship bias…?
YES bc all the Crazy people come back to take the exam again…they push the scores up and then CFAI has to come in to revise it down to continue their “gate keeping” reputation.
The low band people are more likely to give up, Band 9-10 are certain to retake, otherwise the pass rate would be even lower if everybody had to retake by force No opt-out, no break year, no mercy for anybody 
^Good analysis!
If you mean the exam results, then the answer is I dont know.
If you mean the candidates themselves, then yes. Most of whom fail the exam, passes away within 5 to 7 years. This has been reported by the reputable Graveyards of International Pathetic Society; aka GIPS