I’m Hank Green, and this is SciShow. [Intro] First, let’s discuss how we ended up with this imagined link between vaccinations and autism in the first place. Autism diagnoses are DEFINITELY on the rise; now many scientists believe that this is largely or even completely because of more effective diagnosis, and changes in how the diagnosis is reported. So while diagnoses of autism are increasing, we can’t say for sure whether the incidence of autism is also increasing. If it is, it must be because of some environmental factor.
Now, when we talk about autism, we’re really referring to range of developmental disorders, which can affect a person’s ability to communicate or socialize, or cause them to develop patterns of behavior that become pretty specific and inflexible. The condition can manifest itself in a lot of different ways, but you’ve probably heard of them referred to together as autism spectrum disorder, or ASD. While ASD has been found to have some strong genetic components to it, there also seem to be environmental factors at work as well. And that’s really the root of this controversy — we simply don’t know precisely what causes autism.
And in the absence of an explanations, people try to make sense of it themselves. And the way our brains do that is almost entirely with cognitive bias. A cognitive bias is really just anything that skews how we process and interpret new information. There are tons of different kinds of bias — some biases cause us to ignore certain data; others lead us to put too much emphasis on certain data; they can even drive us to focus on facts that are actually irrelevant to what we’re observing. But essentially, when we hear a hypothesis and think, “Yeah, that ‘Makes Sense’” really what we’re saying is “Yeah, that fits with my cognitive biases.”
And so people blame all sorts of things for Autism…plastics, pesticides, the use of anti-depressants during pregnancy, GMOs, sugar, gut bacteria, and vaccines. Basically, you start with whatever makes the most sense to the person doing the hypothesizing. The onset of autism typically happens in one of two ways. Either parents notice a delay in language development, typically around the first birthday. Or they notice an apparently sudden loss of existing development, which might happen all the way up through the third birthday. Now, humans are pattern recognition machines.
We need to be able to figure out what behaviors and strategies lead to positive outcomes. But, even more than that, we’re on the lookout for things that lead to negative outcomes. This over-weighting of negative outcomes is a well known psychological effect called “negativity bias.” So imagine you wake up one morning and your car doesn’t work. Your brain is going to want to know what happened. Did you leave your lights on? Did you drive though a huge puddle yesterday that maybe shorted something out? There has to be SOME reason why it won’t start! On the other hand, if you get in a 15 year old car and it starts up just fine after having had a bad week of barely getting going, you tend to not wonder “What went right!?”
We spend far more cognitive resources attempting to figure out why a bad thing happened than we do trying to determine why something good happened. In psychology, the search for these explanations is called “Explanatory Attribution” and different people have different “explanatory styles”. Some people are more prone to blame themselves, while others search for an external event to blame. But one thing is clear: we are very bad at not blaming anything. It’s not surprising that parents of children with autism, especially parents who notice a sudden loss of previous development, will search for a possible cause. And when the most significant recent event in the health of the child was a vaccination, as can be said for many moments in the life of a young American, we might identify that as a potential cause and deem that link worthy of further examination. Now this, is completely logical.
The problem is that over a dozen peer-reviewed papers have found no correlation between autism and the MMR vaccine, or any other vaccine for that matter. And yet, when you Google vaccines and autism, a fair number of the results claim that there is a link between the two, and that that link is being covered up either by the government or by big corporations. A parent, already experiencing frustration with the medical community’s inability to tell them why this thing has happened to their child, will, on the internet, find a vibrant community of similarly frustrated people who share their values and experiences. These communities are full of anecdotes that draw connections between vaccines and autism. And so, unsurprisingly, some people become convinced that they have found the reason for their child’s disability. Once their mind has been made up, confirmation bias sets in.
Confirmation bias is simply our tendency to more readily, and with less scrutiny, accept information, anecdotes, and worldviews that confirm our existing beliefs. And, again, it is a completely normal thing that every person does. Indeed, trying to convince someone that a previously held belief is incorrect has been proven to actually increase their affinity for that idea. And so a community is born, and the safety of vaccines is called into question. And once the procedure for getting a vaccine goes from the doctor telling you that it is now time for a vaccine — and 99% of parents agreeing because that person went through medical school — to it being a question to ponder, vaccination rates will go down.
A 2011 study showed that parents who think about vaccines before their child is born are eight times less likely to vaccinate their children. Basically, when given an opportunity to research on their own, what they find is confusing. And when confused, the default choice is to simply take no action. This is an example of yet another bias, called omission bias. In effect, we judge harmful actions as less moral than harmful inactions, or omissions. In fact, a frequently cited study found that, when the choice to vaccinate is framed as an action, the average parent will only vaccinate their child if not vaccinating is at least TWO TIMES more dangerous than vaccinating.
This has to do with our perception of future regret. Parents report that they’ll feel worse if they take an action and it harms their child, than if they don’t act and the child is harmed by a failure to act. This perception of potential regret can be so strong that even bringing up the choice of acting versus not acting seems to be counter-productive. A 2013 study found that attempts to convince parents to vaccinate their children actually decreased the percentage who went on to choose vaccination. If vaccination is presented as a personal choice, instead of a necessity for good public health, then potentially harmful inaction can seem more moral than potentially harmful action, and vaccination rates go down.
Parents are choosing to “let nature take its course.” And as you might expect, this effect is much stronger in people with a measurable “naturalness bias.” This is just a tendency to perceive things that come from nature as being inherently less threatening than things that we invent ourselves. One way psychologists measure this bias is by asking a subject if they’d prefer a substance extracted from an herb or one synthesized in a lab, even if they’re chemically identical. And of course, others have biases against big government or big corporations, and these ideas about vaccinations fit well with those worldviews.
Confirmation bias at work again. But even people who don’t hold those biases end up being more likely not to vaccinate if they start doing research before their baby is born. This is because of another failure of the human brain. We are terrible at what psychologists call “Risk Perception.” Given the merest sliver of a possibility that vaccines will cause developmental disorders, parents are now weighing a disease they have seen, autism, against diseases they have never seen. Since the 1970s, measles has been pretty much unheard of.
Measles doesn’t scare people my age for the same reason a giant man-eating squirrel doesn’t scare us…we’ve never seen it. Risk perception is basically a science all on its own, and we have found that vague, future hazards, like the future probability of an illness, are far less frightening than immediate, specific hazards, like the sudden onset of autism. So, amazingly, the success of vaccines is one of the reasons that people are less likely to vaccinate their children. So yes, it turns out humans are complicated, and this is a complicated problem. Humans are inherently bad at understanding the effects of self-selecting samples — like online anti-vaccine forums — and often completely unable to accept that a negative outcome could really be the result of something that’s beyond their control — and still not very well understood.
This is not a “anti-vaxxer” problem; it’s a human problem. Those of us who trust science or have built an understanding of statistics and bias simply have had different lives than people who more heavily weight anecdotes or the opinions of their friends, or strangers they meet online who feel the same way. So next time you find yourself frustrated about the decline in vaccinations in America, remember that it’s only because of the dramatic success of vaccines that we could even think of having this debate, and that those anti-vaccine activists are being driven by the exact same logic traps and cognitive biases that every one of us suffers from. Only by understanding and accepting these psychological pitfalls that we’re all so susceptible to will we be able to solve this problem. And that’s what science is all about.