In the natural sciences, “Raw Data” is always considered sacred. When a result is questionable, the first question is always: “Where is the raw data?”. In cases where the first to discover is questionable, lab notebooks and raw data are the deciders.

As somebody who worked on instruments and measurement for a living, I always questioned what raw data is. In the words of Max Yaffe, (cofounder of Gamry Instruments) “Every data is cooked in some way. If nothing else, it is filtered and A/D converted”. Though I fully agree with Max on this issue, there is the question of reproducibility. A given instrument running a given version of the software should yield the same result on the same sample no matter who measures it in whatever part of the world. Yes, both data are cooked, but they are cooked by the same hardware/software the same way.

Therefore the sacred “Raw” data is really data that is not irreproducibly manipulated by a human. Even when the manipulations are documented, human beings are more motivated by their own dopamine release than the truth. As shown below(from a paper which I will not identify), arbitrarily manipulating data is routine when it is left in the hands of a person.

For direct human observations, the issue is made worse. Ultimately, a pair of human eyes and a human brain processes the data before noting it down somewhere.

Ultimately, the “Raw Data” that should be considered sacred is the machine-generated data that is well documented as to the measurement procedures, parameters along with software and hardware versions for reproducibility purposes.

January 3rd, 2025