Archive for the ‘statistics’ Category

For teens, digital technology is good. Or bad. Or maybe neutral?

January 18th, 2019
For teens, digital technology is good. Or bad. Or maybe neutral?

Enlarge (credit: SimpleTexting.com)

In South Korea, people under the age of 16 can’t play online games between midnight and 6am. The UK Parliament has launched an official inquiry into “the impact of social media and screen use on young people’s health.” Meanwhile in the United States, the Wait Until 8th campaign asks parents to delay giving their children a smartphone until they’re in eighth grade. Worry about kids and technology is rampant—so have smartphones, in fact, destroyed a generation?

A paper published in Nature Human Behaviour this week answers that question, often differently, thousands and thousands of times. Researchers Amy Orben and Andrew Przybylski took three huge datasets and threw every possible meaningful question at them. In part, their analysis is an illustration of how different researchers can get wildly different answers from the same data. But cumulatively, the answers they came up with indicate that tech use correlates with a teeny-tiny dent in adolescent well-being—and that there’s a big problem with big data.

High numbers don’t necessarily mean high quality

Studying small numbers of people, or rats, or trees can be a problem for scientists. Comparisons between small groups of subjects might miss a real finding or luck out and find something that looks like a pattern but is actually just noise. And it’s always tricky to generalize from a small group to a whole population. Sometimes small is the only sort of data that’s available, but some research disciplines have had the recent(-ish) boon of gigantic, rich datasets to work with.

Read 15 remaining paragraphs | Comments

Posted in Behavioral science, Biology, science, statistics | Comments (0)

There’s a debate raging in science about what should count as “significant”

August 4th, 2017

Enlarge (credit: flickr user: Artiom Gorgan)

Psychology and many related fields are in the midst of what can be viewed as a coming-of-age crisis. Following a stream of depressing revelations about a lack of reliability in the field, lots of researchers are dedicating themselves to improving the discipline’s rigor.

The latest proposal to up that rigor is a big one: 72 researchers from a range of disciplines have drafted a manuscript arguing that the threshold for claiming “statistical significance” should become much stricter. There’s often a fair amount of consensus on how science could be improved, but this suggestion has stimulated some intense debate.

Statistical significance in a very small nutshell

Statistical significance is a concept underlying a huge amount of science—not just psychology or social sciences, but medicine, life sciences, and physical sciences, too. “Significance” used in this way doesn’t mean the importance or size of a finding, rather it’s the probability of that finding showing up in your data even though your hypothesis turns out to be wrong.

Read 14 remaining paragraphs | Comments

Posted in Behavioral science, Biology, science, statistics | Comments (0)

There’s a debate raging in science about what should count as “significant”

August 4th, 2017

Enlarge (credit: flickr user: Artiom Gorgan)

Psychology and many related fields are in the midst of what can be viewed as a coming-of-age crisis. Following a stream of depressing revelations about a lack of reliability in the field, lots of researchers are dedicating themselves to improving the discipline’s rigor.

The latest proposal to up that rigor is a big one: 72 researchers from a range of disciplines have drafted a manuscript arguing that the threshold for claiming “statistical significance” should become much stricter. There’s often a fair amount of consensus on how science could be improved, but this suggestion has stimulated some intense debate.

Statistical significance in a very small nutshell

Statistical significance is a concept underlying a huge amount of science—not just psychology or social sciences, but medicine, life sciences, and physical sciences, too. “Significance” used in this way doesn’t mean the importance or size of a finding, rather it’s the probability of that finding showing up in your data even though your hypothesis turns out to be wrong.

Read 14 remaining paragraphs | Comments

Posted in Behavioral science, Biology, science, statistics | Comments (0)