Do the Numbers Really Speak for Themselves?

In my comment on the article by Dessewff y and Láng, I set out from some methodological problems which arise in connection with Big Social Data. Does N=all really hold? Is Big Data more objective than ordinary survey interviewing? Can we really understand data without their formation context? Can we really replace causation for correlation? Each question is illustrated with the example of hedonometer.org, an instrument that measures happiness of populations using Twitter as data source. Some of these methodological problems lead to exciting theoretical questions. I conclude that the fruitful approach is not to contrast Big Data and ordinary surveys but to combine them since they may complete each other’s shortcomings. Likewise, social theories and sociological methods should have a place in Big Data research to reach valid and meaningful results.