Statistics


For Today’s Graduate, Just One Word: Statistics

Statistics, like calculus, is one of the key and integral languages of science. Every American ideally should have a rudimentary understanding of statistics or at least statistically-oriented economics coursework by the time they enter the working world.  It is statistics that makes one a smarter shopper, consumer, driver, reader and thinker. Understanding confidence intervals, p-values, bias, trends, hidden variables and sampling methods are key to cutting through the lies dished out to us daily through the media.   Essentially every "study" presented to us with buzzworthy findings on health, medicine, economics and public perception is riddled with holes in each's statistical makeup. Completion of only an elementary college-level course in my mind gives every American the ability to question a study's findings, the awareness to avoid the lottery, scams, any casino game with a house edge for life, as well as the ability to effortlessly question the claims of scam artists worldwide.  Instead of simply accepting information, Americans must routinely question everything read with statistics as the interrogator. Was the sample size adequate? N=10,100,1000, 1,000,000? P-value? Outliers? Interference? Correlation coefficient? Standard deviation? Mean? Lurking variables?  Look at the application of statistics in this article--a perfect case in point. What is the probability of your completing a PhD programme and subsequently landing a fantastic job at Google? Not good, not good at all.
Alan, Pacific Beach, CA
August 6th, 2009

I see a lot of confusion over what statistics actually IS. There are a lot of jobs that only require the skills taught in an introductory statistics course, and due to the availability of data, the number of these jobs might be growing.  However, companies like Google, Microsoft and Netflix are asking statistical questions that are way beyond anything that is even seen in a first year grad course. They range from discerning topics in online articles to predicting user preferences given past behaviors to modeling language structure. At this level, statisticians need not only the mathematical tools taught at a PhD level, but they also need a tremendous amount of creativity and insight to find workable solutions.  Statisticians are not just "number crunchers."  The best ones are closer to artists, trying to find elegant summaries of their world (and trying to make their computers do all of the hard work).
LH, New Jersey
August 6th, 2009

_______________________________________________________________________________

Lies, Damned Lies, and Medical Science

_______________________________________________________________________________

Op/ed:  Why everyone should take Statistics

_______________________________________________________________________________