Technology

Randomised controlled trials are the gold standard of scientific proof. 

Much of what we now believe is the result of experiments in which people are randomly assigned, like lottery, to receive a treatment and other people are randomly allocated to be in the control group. Researchers then look at the outcomes of intervention. Unlike the observational studies, this tool can get rid of selection bias, when groups being studied are already significantly different after they are "selected" to receive treatment.

Observational studies often fall prey to "reverse causality", in which the outcome drives the "exposure", rather than the other way around. When we observe that people with obesity survive better, it doesn't mean obesity causes longevity. One explanation for the observation or correlation is that people with deteriorating health, such as advanced cancer, often lose weight. When we aren't careful, as it turns out, we could have assumed obesity is the cause of living longer.

All of which is to say that we should put more emphasis on randomised controlled trials. 

This much we know. Yet when it comes to social science, randomisation is not often practical. The idea that social media have detrimental effect on adolescents have been around long enough to become cause of concerns. Jonathan Haidt, the American social psychologist who wrote The Anxious Generation, found 20 randomised controlled trials on social media, most of them randomly assigning students to limit the use of social media platforms (or not reduce the use in the control group). More than two-thirds of these experiments found evidence of harm.

Quasi-experiment study design to measure the effect at community level, on the other hand, is more practical (or realistic) than randomly assigning all students in 20 middle schools to put their phones in a phone locker for a year. One good example of such studies takes advantage of the rollout of high-speed internet. In a "natural experiment" in Spain, researchers made note of the fibre-optic cables and high-speed internet, which came to different regions at different times. Drawing on data that are proxy of internet penetration in Spanish households, the study examined the group-level effects rather than the individual-level effects. As it turns out, there was a significant effect of access to high-speed internet on hospital discharge diagnoses of behavioural and mental health cases among adolescents.

This natural experiment study spoke volumes to the potential epidemic of teen mental illness after the arrival of world-changing smartphone technology.