4 ms·

What do we think of the experiment design? Good example? Anyone think of any obvious flaws?

nknighthb 13y ago
My first thought was local weather -- the exclusion zone is a pretty big place. The two pages of their paper I have access to, however, suggests they were quite conscious of that. If they accounted for it properly, the methodology is probably valid. I'd very much like to see the raw data, of course.
- keithpeter 13y ago
  Raw data with locations &c would make a really good interesting case study for teaching.
Someone 13y ago
As nknighthb says, it's hard to say without access to the full paper. I googled for it, but couldn't find it. There are two things I noticed in the web page (not the paper) that triggered questions immediately. Firstly the claim "Because they had so many bags placed in so many different locations, they were able to statistically control for outside factors such as humidity, temperature and forest and soil type". The map only has 19 or so locations. Seems a bit few to me to control for so many variables. Or am I mistaken? Secondly, the map has areas colored green, but the legend runs from yellow via orange to red. What does green mean? Radiation at normal levels?
- maxerickson 13y ago
  Ecologists tend to squeeze their data good and hard (you see them building a lot of the statistical modeling tools in R or whatever). Facile analysis ( http://en.wikipedia.org/wiki/Oecologia http://en.wikipedia.org/wiki/Oecologia ) says that it isn't the very best journal but probably isn't a joke either.
- thisisdave 13y ago
  I believe this[1] is the paper (paywalled). From the abstract: >we deposited 572 bags with uncontaminated dry leaf litter from four species of trees in the leaf litter layer at 20 forest sites around Chernobyl that varied in background radiation by more than a factor 2,600 Depending on how much variation there is in microclimate and soil type within their 20 locations, I could believe that they could get some pretty convincing results from 572 bags. Especially if, as it appears from their claims, the effect of radiation is very large relative to background noise. On the other hand, the senior author: * is not an expert in any related field as far as I can tell * has been known to fabricate his data [2] and engage in borderline misconduct during the data analysis and writing phases as well [3]. I'm agnostic. [1]: http://link.springer.com/article/10.1007%2Fs00442-014-2908-8 http://link.springer.com/article/10.1007%2Fs00442-014-2908-8 [2]: http://news.sciencemag.org/2004/01/ecologists-rocked-misconduct-finding http://news.sciencemag.org/2004/01/ecologists-rocked-miscond... [3]: http://bio.fsu.edu/~dhoule/Publications/mollerreview.pdf http://bio.fsu.edu/~dhoule/Publications/mollerreview.pdf (large PDF)
  - keithpeter 13y ago
    Latin square type design I presume. Thanks for posting this and spending the time. Given the record in your second reference, I'd imagine the lead author would want to have the data out there and eye-balled by as many as possible to improve reputation.
    - dalke 13y ago
      Some of the data is in link [1] as "Supplementary Material". This can be downloaded without a fee. It isn't much though.

My first thought was local weather -- the exclusion zone is a pretty big place. The two pages of their paper I have access to, however, suggests they were quite conscious of that. If they accounted for it properly, the methodology is probably valid. I'd very much like to see the raw data, of course.

keithpeter 13y ago
Raw data with locations &c would make a really good interesting case study for teaching.

Someone 13y ago

As nknighthb says, it's hard to say without access to the full paper. I googled for it, but couldn't find it. There are two things I noticed in the web page (not the paper) that triggered questions immediately. Firstly the claim "Because they had so many bags placed in so many different locations, they were able to statistically control for outside factors such as humidity, temperature and forest and soil type". The map only has 19 or so locations. Seems a bit few to me to control for so many variables. Or am I mistaken? Secondly, the map has areas colored green, but the legend runs from yellow via orange to red. What does green mean? Radiation at normal levels?

maxerickson 13y ago
Ecologists tend to squeeze their data good and hard (you see them building a lot of the statistical modeling tools in R or whatever). Facile analysis ( http://en.wikipedia.org/wiki/Oecologia http://en.wikipedia.org/wiki/Oecologia ) says that it isn't the very best journal but probably isn't a joke either.
thisisdave 13y ago
I believe this[1] is the paper (paywalled). From the abstract: >we deposited 572 bags with uncontaminated dry leaf litter from four species of trees in the leaf litter layer at 20 forest sites around Chernobyl that varied in background radiation by more than a factor 2,600 Depending on how much variation there is in microclimate and soil type within their 20 locations, I could believe that they could get some pretty convincing results from 572 bags. Especially if, as it appears from their claims, the effect of radiation is very large relative to background noise. On the other hand, the senior author: * is not an expert in any related field as far as I can tell * has been known to fabricate his data [2] and engage in borderline misconduct during the data analysis and writing phases as well [3]. I'm agnostic. [1]: http://link.springer.com/article/10.1007%2Fs00442-014-2908-8 http://link.springer.com/article/10.1007%2Fs00442-014-2908-8 [2]: http://news.sciencemag.org/2004/01/ecologists-rocked-misconduct-finding http://news.sciencemag.org/2004/01/ecologists-rocked-miscond... [3]: http://bio.fsu.edu/~dhoule/Publications/mollerreview.pdf http://bio.fsu.edu/~dhoule/Publications/mollerreview.pdf (large PDF)
- keithpeter 13y ago
  Latin square type design I presume. Thanks for posting this and spending the time. Given the record in your second reference, I'd imagine the lead author would want to have the data out there and eye-balled by as many as possible to improve reputation.
  - dalke 13y ago
    Some of the data is in link [1] as "Supplementary Material". This can be downloaded without a fee. It isn't much though.