previous result pattern

just ran a simple logit model on last year level 1 google result sheet. found some interesting pattern so thought of sharing with all of us. only around 15% of 200 people failed so data is more of those people who passed. If you get more than 70% in fsa, it explains most of the stuff, so out of 200 available results 121 observations received more than 70% in fsa and that predicts their pass or fail probabilty quite close to 100% so other subjects appear rather insignificant. we need to exclude these people to get a better sample of people where people got between 50 to 70% or less than 50% in in FSA. stata results indicate number of obs …77 pseudo-R2…0.442 result = 1.15 (ai)+ 0.275 (derivatives) + 1.26 (economics) + 1.148 (equity) + 1.42 (ethics)+ 1.25(quant) + 0.84 (portfolio) + 0.90 (fixed income) - 17.8 derivatives, portfolio, and fixed income are insignificant variables and equity analysis is also weak but still significant. FSA is not in the results because it is too strong and resutls are valid only for those getting less than 70% in fsa

Iif you get above 70% in FSA, odds are you know your stuff and probably did well on other parts of the exam. Probably Pass. If you get between 50-70% in FSA, odds are you are borderline and probably were borderline on other parts of the exam. Borderline Pass or Fail. If you got below 50% in FSA , that is you failed, well you failed on the most important part of the exam and most probably didn’t pass. Probable Fail. Add in a pass or fail in Ethics and sometimes statistics are not needed.