Group items tagged

Interesting discussion of why headlines like this one "85% accurate" for the detection of stomach cancer" about an experimental breath test are problematic (because some people who don't have the condition get diagnosed with it, and they can miss people who genuinely do have the condition!). Good example using pregnancy tests as an infographic.

As an epidemiologist, I am interested in disease - and more specifically, who in a population currently has or might get that disease. What is their age, sex, or socioeconomic status? Where do they live? What can people do to limit their chances of getting sick? Questions exploring whether something is likely to happen or not can be answered with quantitative research. By counting and measuring, we quantify (measure) a phenomenon in our world, and present the results through percentages and averages. We use statistics to help interpret the significance of the results. While this approach is very important, it can't tell us everything about a disease and peoples' experiences of it. That's where qualitative data becomes important.

Interesting discussion of data on vote recounts and using electronic or hand counting methods (in America where they use electronic voting machines quite commonly). These numbers represent three main kinds of disputes, Foley told me. First, candidates (and their lawyers) argue over what ballots should be counted and which should be thrown out as ineligible. Then, they argue over which candidate specific ballots should count for. Finally, they argue over whether all the eligible votes were counted correctly - the actual recount. Humans are much better than machines at making decisions around the first two kinds of ambiguous disputes, Stewart said, but evidence suggests that the computers are better at counting. Michael Byrne, a psychology professor at Rice University who studies human-computer interaction, agreed. "That's kind of what they're for," he said.

examples of data journalism in audio / podcast form - including: Right To Remain Silent is one particularly good example, because it's about bad data: specifically. police who manipulated official statistics. You might also listen to Choosing Wrong, which includes a section about polling. Another favourite of mine is an audio story by The Economist about the prostitution industry, based on data scraped from sex trade websites: More bang for your buck (there are even worse puns in the charts). David Rhodes, a BBC data journalist, has a range of stories on his Audioboom account, including pieces on Radio 4, Radio 5 Live, and a piece discussing "Did Greece really not pay 89.5% of their taxes in 2010" from the excellent factchecking radio programme, More or Less.

When creating a machine-learning algorithm, designers have to make many choices: what data to train it on, what specific questions to ask, how to use predictions that the algorithm produces. These choices leave room for discrimination, particularly against people who have been discriminated against in the past. For example, training an algorithm to select potential medical students on a data set that reflects longtime biases against women and people of color may make these groups less likely to be admitted. In computing, the phrase "garbage in, garbage out" describes how poor-quality input leads to poor-quality output. In this case we might say, "White male doctors in, white male doctors out."