Questions in statistics class

When it gets too quiet, here are some standard questions to ask:

• What is the performance of this technique.
• decompose
• cost and risk
• bias, variance
• avoidable/unavoidable
• linear/nonlinear
• industrial/professional "standard" procedures and their weaknesses.
• SAS command for it?
• R/S+
• Other related tools (for instance, devices for taking measurements, and such)
  
• Is this same as ANOVA?
• Does this have a max-margin interpretation?
• Max-ent?
  
• Min-max?
• Is there a semi-supervised version?
• What if we add malicious data?
• What if we add uncorrelated data?
  
• How does the new thing (regularisation, re-parametrization, inference, model) do to practical modeling?
  
• What new insight does it provide?
  
• What does it do better at?
  
• Can it be explained on a classical data set?
• Does it have new applications?
  
• So... now that you've done this great thing, what's left for us?
  

;-)