A new study shows that even good health communications programes never change the behaviour of parents who deny their children vaccines. It is fascinating and worrying just because it fits so well into the literature of social psychology.

An international group of child health experts and political scientists decided to conduct a rigorous evaluation of four different positive messages about vaccines. The messages were designed to convince parents who refused to immunise their children that vaccines were both safe and necessary. All failed to increase immunisation rates and some made it more likely that the parents would continue to expose their children to deadly childhood illnesses. Read the Slate version of the story here and the full paper from Pediatrics here.

The communications interventions sound as if they were good. Each did what it set out to do. For example, the message that vaccines are not linked to autism reduced belief amongst parents that the link might exist. However, those same parents became even less likely to immunise their children.

Convincing people has little to do with changing their behaviour. This is nothing new: what you label the effect is the cause of argument amongst sociologists and social psychologists but, we have known for years that showing smokers videos of blackened lungs makes them less likely to give up smoking (and showing parents a terrifying case study of a kid with a vaccine-preventable disease in this study made those parents less likely to protect their own children).

The counterintuitive effect often seems worse when linked to high-fear messages but it exists even when you use softer methods: when parents found out that their earlier fears about vaccines were misplaced, they were still less likely to vaccinate. When cult members wait for the end of the world and it fails to arrive, they become even more faithful to the cult.

It is in the interests of behavioural scientists to pretend that these phenomena don’t exist: for example, hundreds of millions of dollars have been frittered away on behaviour change programmes to reduce the risk of transmitting HIV with very little evidence that any work and a lot of evidence that they don’t — this PDF is the long story of what happened when the Global Fund tried to cut off funding for a South African HIV programme which appeared not to be having any impact on how much HIV was actually transmitted. Most of that money ends up in the pockets of psychologists, sociologists and doctors who are dabbling in making videos.

Funders (usually governments) keep paying because they often ask the wrong questions: did my health communications campaign raise awareness? (Probably if you hired a good creative and spent a lot) Did it change opinions? (Quite possibly if you spent even more) Did it change what people actually did? (Almost certainly not)

Health behaviour typically changes because societal norms change. This is something that health promotion campaigns rarely do in the short term but, for example, TV soap operas do well — a landmark study charts shifting sexual mores in Brazil as Telenovelas spread across the country.

Health behaviour also changes when it becomes easier to do the right thing than the wrong thing: banning smoking indoors drives down smoking and drives it down fastest in places with lots of freezing weather or rain. Vaccination rates go up when parents have to jump through a series of bureaucratic hoops to get their children exempted from a legal requirement to be protected. People should be allowed to have self-destructive ideas; we should make it inconvenient to act those ideas out.