A Date with Data — Issue 9
It’s Tuesday, and this is your weekly date with data! Here are the stories that grabbed our attention from the past week, Sept. 24 — Sept. 30.
Child marriage in Africa persists
Sept. 25, 2018
Every year 12 million girls marry before the age of 18, according to the UN Children’s Fund (UNICEF), and if no action is taken, by 2030 there’ll be over 150 million additional child brides. Last week, The Economist, presented the numbers behind this social crisis, showing that while in parts of Asia the percentage of girls married before 18 is dropping, Africa’s rates are not following the same trend. UNICEF predicts that with the current trend, half of the world’s child brides will be African by 2050. Niger has the world’s highest rate of child marriage, where three out of four girls are married before they turn 18. But it’s far from an unsolvable problem. For example, Ethiopia’s commitment to end child marriage by 2025 and invest more in education, paid off. During the past decade, this practice dropped by a third in Ethiopia. GRAPHIC DETAIL — THE ECONOMIST
Biased News Media or Biased Readers? An Experiment on Trust
Sept. 26, 2018
In a series of graphs, The New York Times presented the results of a social study that explored which news readers are at greatest risk of bias. In this 2017 study, by Knight Foundation and Gallup, the researchers invited a random sample of Americans to read news from different media sources and rate the trustworthiness of their content. To measure the level of bias, half of the participants weren’t allowed to see the source of the news, while the rest of them read the news as it would typically appear on the news outlet’s website. They discovered that the difference in trust could be up to 1.5 points (on a scale of 1 to 5 ) between the “blind” and “non-blind” reviewers. Interestingly, the readers that had less trust in media, were the more biased ones. And similarly, the ones with more extreme political views, gave more biased ratings of news content. THE UPSHOT — THE NEW YORK TIMES
Congratulations. Your Study Went Nowhere.
Sept. 25, 2018
The New York Times published yet another article on bias, but this time it covered something different: publication bias. Despite efforts of governmental organizations, such as the US’s Food and Drug Administration (F.D.A.), that oblige researchers to register their drug trials and report their results, there’s still a long way to go before this bias is eliminated, as professor of pediatrics Aaron E. Carroll explores in this article. There’s still a vast amount of negative results that are left unpublished, leaving doctors, patients, and healthcare authorities with only a partial picture of the true effects of medicines. Specifically, about half of the studies with negative findings are still missing from the peer-reviewed academic literature. However, there are ways to tackle the problem, from encouraging journals not to judge papers based on their results, to systematically identifying missing outcome data, as explained in a BMJ paper published last week. THE UPSHOT — THE NEW YORK TIMES
Dollars for Booze
Sept. 27, 2018
Do people spend more on alcohol today than 80 years ago? The graph-building tool, Datawrapper, visualized how much Americans spent on alcohol, at home and when out, every year between 1935 and 2014. While telling an interesting story about alcohol buying habits since the mid-1930s — two years after the end of Prohibition — this article also walks us through some common statistical mistakes to avoid when calculating and visualizing data over time. Using a series of heatmap charts they show us step by step how they eliminated “pollution” in the raw data caused by two common phenomena: price changes and increase in population. These adjusted data were finally presented on a simple line chart, which we can use to draw meaningful conclusions. CHARTABLE — DATAWRAPPER
That’s all for this week. Thanks for reading!
