“Spurious correlations” is the name of a website I came across recently. There you can see plenty of cases where correlation may not imply causation. What does it mean in terms of research methods in the social sciences? It means that whenever our research approach is uniquely quantitative, we take the risk of come up with wrong or non-consistent findings. In other words, a certain doses of qualitative interpretation is always required in order to avoid wrong and sometimes ridiculous depictions of the social reality, like in the bellow example. It is true that advance and multivariable analysis allow having a more accurate understanding of certain social phenomena nowadays. However such analysis should always be preceded by a good interpretation of such phenomenon. What does “good” mean? Well, first of all take into consideration whether two or more variable are susceptible to be part of the same social reality. It doesn’t seem that “US spending on science, space, and technology” and “Suicides by hanging, strangulation and suffocation” has anything to do one each other. Secondly, even when two variables may potentially explain certain realities we should try to include more than one explicative variable. For instance, when studying the reality of Polish labor market, we may find correlation between education and employment. However, an accurate analysis should also include such other variables as age, income or social class. Such analysis are possible via the so called multivariate statistical analysis to be hopefully covered in future posts. The key of this sort of analysis is considering the right variables. To do so, we can use previous theories, where certain authors (based on previous empirical studies) suggest a number of a priori relevant and explicative variables.
I like the title. False cause is one of the names of a type of logical fallacy. Maybe we should call some of them spurious correlations.
(Just saw the chart–could have done without it–it was a bit graphic and sad for me).