I miss my old job measuring precision optics. Sometimes you get the measurement results and know it can't possibly be right because it doesn't make sense. Kind of the same thing. Don't trust the people doing the measurements.
I've been in this business for nearly 40 years. One thing I can say with absolute certainty, raw data has no agenda, those wishing to gain from data manipulation are found in nearly every company I've worked.
One company (I worked for) paid big bucks for a 3rd party analysis of our power grid and where "unreliability" will go in the next 20 years (keep in mind this was based on a small data set). My department was tasked to preempt power grid failures by being proactive and replacing aging infrastructure before unplanned power outages occurred. The data gathered showed the largest culprits of failure, and, to
@FractalAudio 's point, over 30% of all grid failures were due to cables breaking down (so it just just doesn't apply to guitars). We analyzed the the most at risk cables and proactively replaced or repaired them. Once this work was well underway, it was now my job to report future unplanned outages to the Public Utilities Commission (PUC). Problem is, if I underestimated (or overestimated) the "unreliability," we were fined for not performing due diligence. In order to accurately estimate unreliability out for a year required some serious analytics to model reliability predictions. My model, using ARIMA techniques, was accurate to ±1.8% of the predicted value, IOW quite accurate as I tested it on previous true data as well.
Now, this is where it got interesting, since we were being proactive in replacing aging infrastructure, our unreliability went down and our reliability went up. Sounds logical right? Nope. My boss wanted me to falsify reliability data reported to the PUC because he sold the rate case dollars to our customers claiming he needed vast sums of capital for years going forward in order to maintain reliability. He stated my reliability numbers are going in the wrong direction, the reliability should be going down and unreliability should be increasing....
So I asked him "are you suggesting I falsify reliability data?" He said "No. Of course not. I'm just saying the trend needs to go in the other direction."
I resigned shortly after and cited the reason for my resignation was due to corporate fraud.... yeah, nothing happened either...corruption is everywhere.