This chapter discusses the importance of theory in focusing and organizing analyses of Big Data. Causal relationships have to conform to certain requirements to qualify as such. The chapter explores these since they set a framework within which quantitative analyses of political relationships can be carried through. A good theoretical explanation is required because two events could always occur together, and one could constantly precede the other, without actually causing it – particularly in political science, where correlations between all manners of variables are so common. Most political science journals require articles, particularly statistically based ones looking at data-based patterns and relationships, to specify their research hypotheses at the beginning of the article and to say at the end how far these have been upheld by the data analyses. Reliability and other checks on statistical formulations and measures, given that error affects all of them, are useful as an element in systematic political research.