If I ever find the person who thought a statistical evaluation of this data was appropriate, I'm going to kick them in the shins until I wear holes in the toes of my steel toed boots. Every time a sample is added, removed, or modified I have to start over and it doesn't have a significant impact on the final result. Time waste. GRRRRRR 😠
@Foxthorn
If the records have time stamps, maybe you can make a teachable moment out of it.
Also, is there a quick and dirty statistical test you can run to evaluate whether adding a new record is likely to make any difference?
Last thing - can you bar them from adding more data while you are working?
@Helical_Code Unfortunately, the records already exist and I have to include them. It's more a matter of people making adjustments to how samples are classified based on interpretation and engineering judgement. Some of it is fixing actual errors made by other engineers that weren't caught previously. A lot of this stems from having too many engineers working on one design.
@Foxthorn Well, that does sound quite aggravating! I guess the liberal use of profanity may help. 🍀
@Helical_Code 😂 The joke around the office is that all of my passwords are probably some combination of swear words. My use of profanity is on par with my contractors.
@Foxthorn
Not that I doubted that you are a real pro, but this would have clinched it!
😂
@Foxthorn
Totally serious BTW. I used to work with radioactive compounds and the occasional burst of profanity would help to steady my hands.
@Helical_Code I completely understand. I find that a little bit of profanity helps me get my point across to the client in a clear and concise manner. For example, "your site is sh!t" is much more impactful than "your site has issues," and "absof*ckinglutely do not do what you are planning" is a pretty clear indication of whether or not I'm willing to sign off on something. There's usually some shock when the tiny middle aged woman starts belting out cuss words like a sailor, but they adapt.
@Helical_Code I think they also appreciate knowing that I'm not sugar coating any issues that they might discover later on in the process.
@Helical_Code Normally, these small adjustments would be easily handled because we don't do this level of mathematical analyses for parameters that are based largely on empirical correlations, and changing one sample at one depth does not significantly affect an entire analysis location. In this case, I am required to do the analysis whether or not I agree with the approach. My hands were tied before I started this design.