Big Data and Cloud Culture - To the Cloud: Big Data in a Turbulent World - page 198

2013). Because quantitative research works best on data embodying little
in the way of subjectivity, researchers tend to neglect questions that require
their careful consideration. It is far easier to go for the low-hanging fruit
of voter analysis (there is little subjectivity in the determination of whom
one votes for), or of counting the frequency of search terms, than to
examine, for example, how a young person becomes a racist. The latter
involves an altogether different kind of methodology, which might make
use of some quantitative data but also requires close observation and
depth interviews—in other words, a careful qualitative study that aims
to comprehend the rich subjectivity that makes up personal and inter-
personal experiences. Big data deals with subjectivity to the extent that
analysts can do the impossible—i.e., assign a precise numerical value to
its various states. This is inherently lawed because subjective states such
as happiness, depression, or satisfaction mean different things to different
people, and assigning the same numerical value to the choice of this term
simpliies to the point of absurdity. The same goes for other attitudinal
terms such as like and dislike, agree and disagree, and their ampliiers,
such as “strongly.” What is the meaning of a number associated with these
terms? How can one assign any meaning worth taking seriously to the
numerical difference between disagree and strongly disagree?
It is uncertain which is worse: that big data treats problems through
oversimpliication or that it ignores those that require a careful treatment
of subjectivity, including lengthy observation, depth interviews, and an
appreciation for the social production of meaning. There is a difference,
as the computer pioneer Jaron Lanier notes, between using big data to
analyze weather or galaxy formation and using it to examine the emotional
states of human beings, which are often contradictory and unreliable
(Lanier 2013). Such an approach only feeds what Roman Kudryashov,
drawing on Roland Barthes, refers to as the myth of the quantiication
of quality: “When language cannot handle the complexities of reality, it
strives to economize the world: qualities become quantities, and once
again, language goes beyond reality to judge it. Though language tries
to be scientiic about its descriptions here, it has attributed properties
not belonging to the original object, and thus does not judge the object,
but its properties” (Kudryashov 2010). As Barthes himself asserted,
“A whole circuit of computable appearances establishes a quantitative
equality between the cost of the ticket and the tears of an actor” (1982,