What are the odds the Supreme Court will overturn the Court of Appeal in any randomly selected case coming before it?  The Justices may just be more able to provide that probability analysis after its review of Duran v. U.S. National Bank Association 203 Cal.App.4th 212.  This case, now pending before the Supreme Court, will decide the way proof is admitted to prove class members properly belong in the class. 
In Duran, the issue was the scope of proof needed to establish an employer’s liability to a class for non-payment of overtime.   The Court devised its own method of getting to an “efficient” trial method to admit evidence applicable to all 260 employees:  it randomly selected 21 individuals out of the total population of 260.  Based on this random sample, Court ruled the entire population of employees was misclassified as salaried exempt.    
The trial court refused to allow evidence obtained by the defense from 70 employees that they were not properly in the class overtime eligible employees.  At least these 70, argued the defense, should have been excluded from the “presumed” class of all 260 employees.  
Our conservative Supreme Court may track the thinking of the U.S. Supreme Court in Wal-Mart Stores, Inc. v. Dukes (2011) 131 S.Ct. 2541 (2011) that found serious due process concerns with statistical methods that overlooked significant individual differences among a large part of the population of all “class” members.  But, our Court, applying our overtime laws, and following class action law for California employees, will be free to fashion their own class action procedures. 
The strange part of the trial court’s decision in the Duran case was it’s finding of a 95% certainty that each member of the class worked 11.86 overtime hours per week, subject to a relative margin of error of 43%.  The logic of this statement is bizarre:  “I am  95% certain that my conclusion is true about one-half the time.”    I predict the CA Supreme Court will not let this kind of statistical analysis result in liability for the all class members, especially where the defense sought to present 70 declarations from purported class members showing they were not properly in the class. 
Here is Wikiepedia’s short version of the major decision points in crafting a survey sample to produce a reliable statistical analysis:
§  Defining the population of concern
§  Specifying a sampling frame, a set of items or events possible to measure
§  Specifying a sampling method for selecting items or events from the frame
§  Determining the sample size
§  Implementing the sampling plan
§  Sampling and data collecting
In conclusion, I’m 95% certain that I have about a 50% chance of picking the outcome of the Duranappeal. 

“If the pink slip doesn’t fit, get redressed!”
Social Media to see my complete social “pink slip” wardrobe. “Fighting for the Little Guy”