Antivaxx
We live in a world where data surrounds us. Collected on almost any possible occasion, data has become a critical resource for almost every company. New jobs are created to process, analyze, and store data, as companies value information drawn from data. The inferences drawn matter too, as they decide how the company, consumers, and other stakeholders can and will act.
Even though a big burden lies on the process of analyzing information, sometimes people don’t do their duty. They might be ignorant about a certain aspect of the data (intentionally or unintentionally), they fail to consider external sources, they might tell subtle white lies, or other nuances such as these things. All of this feeds into bullshit. Bullshit isn’t strictly defined, but any time one fumbles or their way through something, misrepresents a concept or explanation, or things talked about above, they haven’t lied per se, but they haven’t made an explicit effort to tell the truth in its entirety.
Bullshit may not stem from an direct attempt to misinform, which separates it from lying and propaganda. These direct try to refrain from the truth or cherry pick information respectively, the conscious effort separating them from the bullshit. While these could be classified as evil, bullshit could simply be a result of laziness: it becomes a matter of virtue ethics.
Since intent matters, misinformation is placed in a category of its own as well. The author/storyteller could try to tell the truth, yet due to circumstances out of their control, they end up spreading false things. While unfortunate and bad, it could happen to everyone, but this assumes the analysis was done through due process.
Unfortunately common, ignorance can stem from not doing a proper analysis. Conducting the right analysis can mitigate unintended consequences. The step by step instructions in exploratory data analyses force the research to go under a sort of procedural review, which we know as it has been tried and tested, refined, and covers many important bases.
Many research cases have failed to follow this process correctly, which has lead into bullshit with massive unintended consequences that turn into ripple effects. Examples of these can be found in the media, one in particular has made parents all over the country afraid of vaccinating their children.
In 1998, Dr. Andrew Wakefield conducted an experiment in which eight children were given vaccines to measles, mumps, and rubella, and several days later, showed signs of autism. He published his findings before doing further research, and thus an unfortunate era was born.
This experiment didn’t remotely follow due process. The number of subjects was too low, he failed to consider alternate sources, he published too early, etc. He lost his medical license over this incident.
At this point, there’s enough evidence to sufficiently disprove Wakefield’s claim, but the damage was dealt. A movement of anti-vaccines was born. Other researchers have gone to extensive lengths to try to replicate Wakefield’s research, but all of them have degrees of bullshit in them. Take the figure below of reported measles incidence in Canada as an example.
The figure above shows that vaccines seem to not help combat Measles very much, yet it cleverly masks two elements. First, although the data looks continuous, there are only 5 data points represented. Secondly, it conveniently fills in missing data to help show the progression it wants. The bullshit factor is highlighted in this example, as the ignorance to consider other data and the laziness in filling in the data shouldn’t have been ignored. Take a look below at the real incidence, reported by the Canadian government, that shows a different side.
The true story is that the vaccine clearly did reduce incidence rates. This is only one example in the anti-vaccine movement that has probably tricked many people, unfortunately. Since many people have started thinking that vaccines are ‘bad’ and have no merit, I wanted to learn how it has affected people, so I did some research of my own.
I found a dataset on Kaggle that shows data in California on the incidence of preventable diseases between 2000–2017. I wanted to do proper data analysis, so I started with a research question of: “With the increase of people in anti-vaxx movement, has the rates of preventable diseases gone up in California?” I then read in my dataset into a Jupyter Notebook (using pandas), and checked the packaging. It looked fine, so I looked at the top and bottom of my data, and it seemed like everything was there. I checked the rows for any missing or mutated data, and it was all proper.
To validate my assumption that the anti-vaxx movement was growing, I also brought in an additional dataset of vaccination rates of kindergarten students in California, and checked that the anti-vaxx movement was increasing. The graph below shows the percentages of unvaccinated kindergarten students over time starting in 2000, about two years after Wakefield’s paper.
The movement is growing, but how has it affected people? The below graph shows my findings from the deaths dataset. Its a graph of the count of preventable deaths over time, starting in 2001.
It sure seems like counts have gone up a lot, but I wanted to incorporate some statistics, so I ran a t-test to check. I compared the increasing rates of unvaccinated kids in California to deaths, and it gave me a p-value of .0027, which means that I fail to reject my null-hypothesis that there is no correlation. In other words, there very much seems like there’s a correlation between using vaccines and preventing deadly diseases in children.
Its unfortunate that people are heads-down with the idea that vaccines = bad, but the results from my experiment lead me to believe that they save lives.
