Posted by: maroonmaurader | October 28, 2009

Statistics Abuse and the Gender Gap Report

So the latest Global Gender Gap Report is out. I was looking into how they generated their rankings out of curiosity. To summarize briefly: they gather a number of fairly clearly defined and quantitative statistics about each country which reflect aspects of gender (in)equality (e.g., M/F ratio in the workplace, ratio in high office, life expectancy, birth ratio…).

They opted to not adjust the “gender gap” scales negatively if the ratio was more favorable to women then men (leading to the improbable ad absurdum that a country in which men are forbidden to work or hold office would be given a remarkably small gender gap). However, I’m sure they have their reasons and I doubt it really effects the overall results so I’d be fine with accepting that.

It’s in how they create the overall ranking that I’m truly disappointed. They create four sub-indices: economic, education, health, and political equality. Each of these sub-indices considers a number of different factors. As they note when talking about assigning weighted values to subindex scores, “Averaging the different variables would implicitly give more weight to the measure that exhibits the largest variability or standard deviation.” So instead they take a weighted average, to remove that effect. However, when they go to create the overall ranking they simply take a flat average of the four main sub-index scores.

Health scores vary between 0.98 and 0.93. Education scores, except for the bottom 30 countries, vary between 0.99 and 0.93. It should be noted that a small gap between high scores doesn’t necessarily indicate that both countries are very good – it frequently just means that their metric is scoring everything artificially high. For example, consider the sex ratio at birth. They examine this because it can reflect abortions due to a preference for male children. If one out of five female fetuses is aborted specifically for being female, I would consider that an incredibly high rate. By their metric, however, that would score a 0.8; when weighted and averaged with life expectancy ratios, a country could still end up with a score of almost 0.9 in health – which means it would have relatively little impact on the overall rankings, when compared to such massive influences as number of years with a female executive out of the past 50. Aborting one out of five female fetuses is weighted equivalently equivalent to the difference between having a female candidate win half the head-of-state elections over the past 50 years, and win one fewer than half (on a 4 or 5 year election cycle).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: