In its « 12h15 » newsletter dated October 6th, and its web edition on the 7th, le Monde blisfully reported about a study by the Boston Consulting Group. 64% of interviewed job seekers are ready to consider a job abroad. And 94% of French ones, it was underlined. The “12h15” insisted that more than 200 000 respondents in 189 countries participated to the survey. This is what is called Big Data, was the implicit subtitle.
94%. Well, well. A recent week-end in London showed me again that my obstinate love for my country, its innovative businesses, but also it quality of life, its culture, its restaurants and its subsidies for small businesses like mine, seemed a bit offbeat to non-French people. But by so much? I thus diligently went to the BCG website to understand what the vast majority of my fellow citizens think.
A methodological summary taught me that « BCG and The Network believe that taking a data-driven approach can help solve [HR businesses] challenges ». The Network is the organisation which provided the email addresses used for the survey. According to its website, The Network is « your partner of choice for international recruitment ». The sample is thus only representative of The Network’s clients: their profile is way off the generic profile of job seekers, even high income ones.
The bias in the sample can be seen immediately from the available profiling data. Data for France are not shown, but those for Germany are. In the German sample, only 30% of respondents have dependent children. A few clicks away, the data available on Destatis’ website tell us that the true proportion is over 50%.
The reference population has nothing to do with job seekers, nor high income ones. The title of the article in Le Monde needs to be reworded: 94% of French high income job seekers who contact a head hunter specialised in international placement want to work abroad. I know, less sexy. But then, I am worried for countries where the figure is below France’s….
Nothing new with all this, unfortunately. Everybody knows the story. In 1936, Gallup correctly predicts Roosevelt’s re-election thanks to a well calibrated sample, when the Literary Digest, with more than two millions answers from its readers, was anticipating the election of its Republican opponent, Landon. Businesses which bet on the election of a president closer to their interests could only kick themselves. Relatively speaking, businesses which would use the BCG survey to understand the yearnings of their employees would be in the same situation.
Big Data – the capacity to collect tons of data at a fraction of the cost of a few years ago – is an opportunity that all businesses looking for profitable growth should seize. But throwing away at the same time basic statistical concepts only can lead to bad decisions. By combining statistical expertise with the knowledge of handling vast amount of data, SLPV analytics can help you make the right choices. Really “data driven” ones.