Quote:
Originally Posted by Spuds
Ok quick look through the cited study and it's references and have some interesting observations. See study here:
https://www.pnas.org/content/118/20/e2013637118
It looks like this study makes use of data that other studies have performed which are based on a PM2.5 model that was created by through yet another study. I haven't looked into the model much, though it is open source, but I found that at least one creator of the model shows up as an author in all layers of aforementioned studies. While not unusual to see somebody participating in follow-on work, it is important to note creator's bias here.
The question becomes how accurate is the model, and is it backed up by empirical evidence. Judging by the rate of development and amount of data outputs it claims to support, I would SWAG that this model is based entirely on outside research. Meaning based on research others have done for their own purpose, reused to meet the needs here. With how all the covid-19 data repackaging and assessments have led to misinformation, we should all be aware of the potential for that here, though I am certain it is unintentional on the author's part.
The model in question:
http://spatialmodel.com/inmap/
As an engineer, I would not feel comfortable using this study as a sole source of evidence of the implied relationship. I would want to see hard data to back up the simulated findings for at least half of the steps it takes to connect them. I don't think I have seen any test data in any of the references I have access to. As such, this is an interesting mathematical exercise, but not much more. |
The authors mention that this study is novel, and as such is the case, there is the implication that more research is needed, and I would imagine that even the authors are skeptical of the results in light of that. With that said, this type of research isn’t novel, and it is backed by many supportive studies addressing respiratory problems in farm workers.
Models are best when they are predictive and well vetted. I don’t know how this one ranks, but these types of models and comparative studies have been done before for related things like fire related particulates or smog related particulates and deaths. I recall that LA’s smog is attributed from similar studies to a certain number of annual deaths. This article addresses the topic well:
https://www.google.com/amp/s/www.new...431692%3famp=1
It is also worth noting that many studies use data from other sources like from other studies or from private or government public databases. In fact, there are scientists that specialize in statistical analysis that collaborate with authors or who just reanalyze studies where light statistical analysis was used. These scientists will find relationships from the data that previous authors did not find because those authors weren’t skilled in statistical analysis or didn’t invest the time to do enough analysis. I’ve read studies where the statisticians came to more profound and glaring conclusions using the data from other authors’ studies without having to ever go into the field or pick up a test tube or a pipet. Talk about stealing someone’s thunder.