[R-sig-eco] post-treatment measurements as covariable?

[Nicholas Lewin-Koh]

Hi Melissa,
I think the problem is confounding.
If you do not have pre and post measurements
of soil moisture, you cannot really fully model
the effect of warming on on growth vs the effect of drying, without 
making some assumptions. Though it would seem to me, rusty though my 
botany is, that increasing temp, would increase transpiration rate 
of your plants, and hence growth. In general you need longitudinal data
to disambiguate these causal factors, with pre-treatment baseline 
measurements. If this was a several year experiment, than maybe
there are some assumptions you can make and get adequate 
results, such as all plants start with the same moisture level. 
But if it is not too onerous I would repeat the experiment with a
cleaner design.
And use your current data to plan the experiment in terms of power
and treatment levels. This sounds like a good candidate for a split plot
experiment.

Nicholas






> ------------------------------
> 
> Message: 5
> Date: Wed, 19 Jan 2011 11:24:57 +0100
> From: melissa.dawes at slf.ch
> To: r-sig-ecology at r-project.org
> Subject: [R-sig-eco] post-treatment measurements as covariable?
> Message-ID:
> 	<OF6EEDFA2C.AB268F73-ONC125781D.00393781-C125781D.00393784 at wsl.ch>
> Content-Type: text/plain
> 
> Dear All,
> 
> I 
> am trying to analyze the effect of a warming treatment on plant growth. 
> The warming treatment is associated with some drying of the soil, and I 
> would like to determine if/how much of the growth response is due to 
> drying. My original idea was to include soil moisture as a covariable in
>  the model (which would include all variation from plot to plot, 
> including that due to the warming treatment) and see how this changed 
> the warming effect. However, I recently looked at Gelman and Hill 
> (2007), where they explain in chapter 9 that including post-treatment 
> variables is highly problematic. I would greatly appreciate any 
> suggestions (or references) for a more appropriate way to distinguish 
> between direct effects of warming and indirect effects due to reduced 
> soil moisture.
> 
> original model format:
> model <- lm(growth ~ cov + warming, data = x)
> 
> (probably incorrect) model with soil moisture added as covariable:
> model <- lm(growth ~ cov + soilmoisture + warming, data = x)
> 
> variables:
> growth = response variable (annual shoot length measured in x individual
> plots)
> cov = pre-warming covariable (annual shoot length in years before
> warming)
> soilmoisture = soil moisture measured in each individual plot (continuous
> variable, measured during years with warming)
> warming = categorical variable (warmed or unwarmed)
> 
> Thank you in advance for your help!
> Melissa
> 
> ----
> Melissa Dawes, PhD
> Research Group Alpine Ecosystems
> WSL Institute for Snow and Avalanche Research SLF
> Flüelastrasse 11
> 7260 Davos Dorf
> Switzerland
> 
> 
> 	[[alternative HTML version deleted]]
> 
> 
> 
> ------------------------------
> 
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
> 
> 
> End of R-sig-ecology Digest, Vol 34, Issue 12
> *********************************************
>