Sunday, July 25, 2010

Cause and Effect

Last night I finished reading "Freakonomics"  by Steven D. Levitt and Stephen J. Dubner.  This is another provocative book that examines the underlying elements that influence economic and other outcomes.  They write about the hidden side of everything.   They have an interesting take on relating cause and effect as they analyze some fascinating data bases.    I like the way they probe the underlying data to mine elements of causality that seem to make sense after the fact.   However, I am skeptical of their procedures as being successful in predicting outcomes for the future.   And some of their correlations are hard to accept.   Their final epilogue chapter throws the reader a curve.   In all the preceding chapters they connect data in  a logical manner to the outcomes of people and systems behavior.   However in the final chapter they tell of two individuals whose success or failure contradict all elements that should control their behavior.   In effect, serendipitous events seem to trump all the other predictors.

The chapters on parenting are especially challenging to understand and accept.   These chapters wrestle with the nature versus nurture issue in a child's development.    They basically conclude from a California data base that the child's success is more correlated with who the parents are rather than what they do to nurture the child.  That is, for example, socio-economic status of the parents has a major correlation with success.   I wrote about this earlier as I pondered the phenomena of physician families producing physicians and health care workers and attorney families producing attorneys etc.    Although in the context of massive data banks these type of correlations exist, it is also true there are individuals that are mavericks succeed in spite of the negatives in their lives.   As one who has parented seven children in one way or another, I still like to think that nurture does count for something.  In a way good nurturing is no guarantee for success but it does give one comfort that you at least have set up the conditions for success.   After that, serendipitous events may well create some havoc or a boost.  Who knows?

Finally, "Freakonomics"  has made be recall my days in working with the Engineering Admissions process at Cornell.   Through our entrance criteria we attempted to select the student applicants most likely to succeed.   Logically we should see a correlation of combined SAT scores with the Grade Point Average achieved in the program.    In the aggregate is was true that on the average, there was a strong correlation between SAT scores and GPA.   However, if one broke out individuals from the masses it was common to find highly successful students with low SAT scores and also the reverse would happen.   I can only conclude that there were many hidden aspects of the students behavior and preparation that we did not know or understand.  I guess that bottom line is that humans and political and economic systems are so complex that we are unable to comprehend all the aspects of cause and effect.   This demands that we should always test the conventional wisdom.  There may well be something we missed.

No comments: