Database Reference
In-Depth Information
how we allocate work—maybe to a state of permanent, temporary work, for
the mostly nontechnical ranks of the work force” (Hardy 2013c).
The cloud and big data come with the vision of perfecting our knowl-
edge of the world if we can collect more information, improve the sifting
for correlations, and come up with just the right reinements in models
and algorithms. But it may be that the world is so complex that the lofty
aspirations of big-data enthusiasts are out of reach. Perhaps it would be
better to at least supplement big-data studies with old-fashioned depth
interviews on a carefully selected sample. But diversifying methods is pos-
sible only when analysts approach the problem with open minds and the
skill to carry out research using multiple approaches, rather than with the
view that we have discovered the key to a revolutionary transformation in
how we acquire knowledge.
Given his considerable success in forecasting election results, one would
not expect Nate Silver to take a critical view of big data. However, this is
precisely the position he supports in all of his writing, but especially in The
Signal and the Noise (2012), a carefully written overview of the potential
and the problems of large-scale statistical analysis and prediction. For
Silver, devotion to the statistical techniques and values of Bayesian analysis
means committing to probabilities over certainties and recognizing that
all research is infused with biases that we can recognize, if not eliminate,
and then account for them. Assume, he maintains, that the complexity
of the world puts certainty out of reach and one is likely to do a better
job of approximating an accurate conclusion and make reasonable, if not
always accurate, predictions. It is not the size of the data set, but, as has
been the case for as long as people have carried out social research, the
skill and humility of the researcher that most often determine success.
A good example of this point arose in 2013 when a doctoral student
uncovered signiicant errors in an academic paper that has been used by
government policy makers and corporate decision makers to support strong
economic-austerity measures by public authorities around the world. The
article “Growth in a Time of Debt” drew on several large data sets to
ostensibly demonstrate that when the ratio of government debt to gross
domestic product (GDP) exceeds 90 percent, the median rate of economic
growth drops by 1 percent and the average growth rate by considerably
more. The 90 percent threshold applied to both developed and emerging
economies (Reinhart and Rogoff 2010). If one can speak of an academic
inding going viral, this paper is a prime case in point. The authors, one
Search WWH ::




Custom Search