I’ve been reading the book Best Evidence by David S. Lifton on the assassination of John F. Kennedy. Lifton related how, in 1965, he used his tax refund check to buy one of the original prints of all twenty-six volumes of the Warren Commission investigation. He described their disorder, stating that the original version didn’t even include an index:

No verbal description really captures the voluminous and disorganized nature of the twenty-six volumes. Numerous side trips were necessary to relate the approximately eight thousand pages of documentary exhibits in the last eleven volumes to the five thousand pages of testimony in the first fifteen. Given a normal schedule and a reasonable curiosity, that task could easily stretch to a year.

In 1965 the twenty-six volumes lacked an index. They were like a library without a card catalog. One had to plow through everything to find anything. The chaos seemed planned. In 1965, the prevailing climate was one of confusion and a sense of mystery, an uneasy feeling the government was deliberately hiding something.

The sentiment that the “chaos seemed planned” resonated with me because I had the same feeling when I was updating the website Sum COVID with Vaccine Averse Event Reporting System (VAERS) data on the covid vaccines.

Initially, all tables and graphs I made for Sum COVID only included data from the CDC’s Open Data API. The CDC’s Open Data API is nice. You can browse over a thousand different datasets, like the COVID-19 Case Surveillance Public Use Data. You can register for an API key to fetch JSON responses programmatically from an application, or experiment with specific endpoints right in the browser. You can even pass “SoSQL statements,” kind of like SQL statements in the form of a query string, to have data filtered and summarized on the CDC’s own servers, before it ever reaches you.

In contrast, the VAERS data was hard to access. For whatever reason, it was a separate system from the CDC’s Open Data API. For VAERS, there was no API. The site looked like it hadn’t been updated in years. To get any data, you needed to manually complete a form with many fields and run a report that was slow relative to the Open Data API. The form was only accessible after acknowledging a long disclaimer in the form of a pop-up. JSON output wasn’t available. If you navigated the sub-menu of the results page and found the button labeled “Export API,” it would download an XML file without telling you the file type. If you just clicked the button that said “Export” on the default screen, it would download a TXT file. The bottom of the TXT file was crammed with disclaimers and other information that you needed to either delete or filter out if you wanted to programmatically read the data.

The CDC’s Open Data API showed they had the knowledge, budget, and intent to make data convenient to use by the public, and they did it well. With the Vaccine Adverse Event Reporting System, it was like the CDC decided to make all of the opposite decisions that had made the Open Data API such a pleasant experience. Of course, the natural explanation was that there was nothing sinister about it — that VAERS was just never prioritized or would have been difficult to upgrade, while the 1,618 other datasets were considered more important or easier to include for whatever reason. However, given the climate in which I was updating Sum COVID, of government officials and corporate journalists attempting to demonize those who chose not to get a covid vaccine, to destroy their ability to make a living and participate in society, the little decisions that had been made by the CDC that resulted in making it harder to access VAERS data felt conspiratorial. I remember thinking at the time, “This shit is intentional.”