Let's make a list of the variables for inclusion in your analysis that we can use for refining the analysis (#5). For each, we’ll need to know the following:
(1) ForC variables to be included within the broader variable definition (e.g., ANPP may include ANPP_0, ANPP_1, ANPP_2)
(2) data quality criteria to be implemented (i.e., should any data be excluded based on sub-standard methodology?). Examples may include maximum min.dbh, min plot size.
(3) methodological covariates that need to be tested for potential bias (e.g., GPP measured via eddy flux vs scaling up chamber measurements) and potentially included in the model
(4) methodological covariates that need to be included in the model (any that show up as important in step 3 should be included, or some may be automatically included at this stage)
For 2-4, keep track of the sample sizes available under various criteria.
I expect that the process of making this will be an iterative process, and we'll need to discuss issues/ explore the influence of methodological factors/ balance tradeoffs between data quality and quantity.
teixeirak
commented
5 years ago
@beckybanbury,
Let's make a list of the variables for inclusion in your analysis that we can use for refining the analysis (#5). For each, we’ll need to know the following: (1) ForC variables to be included within the broader variable definition (e.g., ANPP may include ANPP_0, ANPP_1, ANPP_2) (2) data quality criteria to be implemented (i.e., should any data be excluded based on sub-standard methodology?). Examples may include maximum min.dbh, min plot size. (3) methodological covariates that need to be tested for potential bias (e.g., GPP measured via eddy flux vs scaling up chamber measurements) and potentially included in the model (4) methodological covariates that need to be included in the model (any that show up as important in step 3 should be included, or some may be automatically included at this stage) For 2-4, keep track of the sample sizes available under various criteria.
I expect that the process of making this will be an iterative process, and we'll need to discuss issues/ explore the influence of methodological factors/ balance tradeoffs between data quality and quantity.