On Wed, 8 Aug 2012, MacQueen, Don wrote:
Specifically, since it has only a single detection indicator column (ceneq1), it implies that within any single sample either all the analytes were detected, or all were not. Not what I would expect.
Don, I have been thinking about this and wondered whether the cast format was appropriate just for the reason there's only a single censored indicator. I'm glad you confirmed this as the/one problem.
As to your larger question of which layout is appropriate for use with NADA functions, the answer is that either can be used. The "trick" is to use the appropriate syntax to extract the values needed to pass the data to a NADA function.
And I've not discovered this in the time I spent trying different syntax.
For the long format you subset the rows, then pass the appropriate columns. Here's one way: with(subset(chem, param=='AgDis') , ros(quant,ceneq1))
This makes a lot of sense. I was thinking that I needed to create separate data frames for each parameter but subsetting on the fly is much more efficient and elegant.
Hope this helps.
It certainly does! Thanks.
(p.s., I still think you'll be better off in the long run if you store site, param, and maybe era, as character objects, not factors.)
I need to research this because I thought that factors were character objects used to partition quantities into groups. Regards, Rich ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.