Last year, exactly on this date, I posted an article about women on this site ( ). Today I have decided to look at things from a somewhat light-hearted point of view. Even though women’s plight in the world has not changed since last year, and that it may even be worse, I do not feel like dramatizing things, so let’s keep on reading..

As it would be expected in any developing economy, the number of motor vehicles in Turkey has been steadily increasing. The Statistical Institute of Turkey reports that the numbers went up from 8,521,956 in 2001 to 19,994,472 in 2015. This is not only due to the expansion of the economy, but also because of the increase in the population, which, according to the World Bank, grew from about 64 million in 2001 to almost 76 million in 2014. The number of divorces also increased, parallel to the population increase during this period. In the face of such sharp increase in numbers, and out of curiosity, one might very well be interested in looking into this. The following plot displays what is explained above. The correlation coefficient r = 0.9359 as calculated by R, is highly impressive. However, let’s face it, despite the rationale one might think of regarding the increase in individual variables and their covariation, this is one of those cases of spuriousness.

Motor vehicles and Divorces in Turkey (2001-2015)
Motor vehicles and Divorces in Turkey (2001-2015)

It should be obvious to anyone with a minimal understanding of socioeconomic phenomena and statistics, that it just does not make any sense to think of divorces somehow being linked to the number of motor vehicles. One cannot possibly cause the other, so we will have to think of confounding factors (variables in technical lingo) such as development of economy, emancipation of women, economic crises causing unresolvable disputes between spouses, etc. that actually play a role in the background, affecting our variables of interest. As this exercise illustrates, one can find a relationship and calculate a correlation coefficient between almost any two variables (factors or phenomena in ordinary language) in life. However, the question is whether or not that seeming relationship will be meaningful.

Briefly, the moral of the story for the newly initiated is that looks can be deceptive, we would be well advised to delve into the matter and look for hidden factors.


The Human Development Index (HDI)

The Index is generally viewed numerically, and countries are compared on the basis of their numeric distance from each other (UNDP Data ). While this is usually sufficient to get an idea about the differential levels of development, one might also be interested in a visual inspection of the data. Here’s a nifty study on the HDI of the United nations: Clusters in the Human Development Index

As I have indicated in a previous post, cluster analysis comes in handy in many situations where categorical differences are of interest.

The Rising Tide of R

I do not know to what extent R, the Statistical Computing Environment,  is being used in the social science community (or by fellow political scientists for that matter) but in a variety of disciplines R is becoming the tool for data analysis.  Despite the fact that it has a steep learning curve, it is certainly worth diving into.  It is a very powerful language, and as such it poses a real challenge to the commercially available software such as SPSS or SAS.  Although commercial software has its own advantages, R is superior in many ways.  Not only is it freely available, it also has a wide and growing community of users that keep adding to its arsenal.  Even though a user doesn’t need all the features R offers, it has something for everyone who deals with data.  The magazine Nature has published an article that should give you a good idea about R.  Click on the following link to read it: