11.07.2015 Views

statisticalrethinkin..

statisticalrethinkin..

statisticalrethinkin..

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

7.3. CONTINUOUS INTERACTIONS 231is example will also allow me to illustrate two benefits of CENTERING prediction variables.You met centering already (page 111). When a prediction variable is “centered,” it isjust rescaled so that its mean is zero. Why would you ever do that to the data? ere are twocommon benefits. First, centering the prediction variables can make it much easier to leanon the coefficients alone in understanding the model, especially when you want to comparethe estimates from models with and without an interaction. Second, sometimes model fittinghas a hard time with uncentered variables. Centering (and possibly also standardizing) thedata before fitting the model can help you achieve a faster and more reliable set of estimates.Still, even with everything centered and double-checked, it can be hard or impossible toreally understand continuous interactions from numbers alone. So the message here is toalways plot posterior predictions, counterfactual or not, in order to avoid misunderstandingthe model fit.7.3.1. e data. e data in this example are sizes of blooms from beds of tulips grown ingreenhouses, under different soil and light conditions. Load the data with:library(rethinking)data(tulips)d

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!