R help - Jul 2006 - Counting observations split by a factor when there are NAs in the data

I am a very novice R user, a social scientist (linguist) who is trying
to learn to use R after being very familiar with SPSS. Please be kind!

My concern:
I cannot figure out a way to get an accurate count of observations of
one column of data split by a factor when there are NAs in the data.

I know how to use commands like tapply and summaryBy to obtain other
summary statistics I am interested in, such as the following:
tapply(RLWTEST, list(STATUS), mean, na.rm=T)
summaryBy(RLWTEST~STATUS, data=lh.forgotten, FUN=c(mean, sd, min, max),
na.rm=T)

However, with tapply I know I cannot use length to get a count where
there are NAs. summaryBy appears to work the same way. I do know how to
get a count of the entire column using sum:
sum(!is.na(lh.forgotten$RLWTEST))

However, this does not give me a count split up by my factor (STATUS). I
have looked through Daalgard (2002) and Verzani (2005), and have
searched the help files, but with no luck.

Thank you in advance for your help. I love R and am interested in making
it more accessible to social scientist types like me. I know it can do
everything SPSS can and more, but sometimes the very simplest things
seem to be a lot harder in R.

Jenifer

Dr. Jenifer Larson-Hall
Assistant Professor of Linguistics
University of North Texas
(940)369-8950

"Jenifer Larson-Hall" <jenifer at unt.edu> writes:
> I am a very novice R user, a social scientist (linguist) who is trying
> to learn to use R after being very familiar with SPSS. Please be kind!
> 
> My concern:
> I cannot figure out a way to get an accurate count of observations of
> one column of data split by a factor when there are NAs in the data.
> 
> I know how to use commands like tapply and summaryBy to obtain other
> summary statistics I am interested in, such as the following:
> tapply(RLWTEST, list(STATUS), mean, na.rm=T)
> summaryBy(RLWTEST~STATUS, data=lh.forgotten, FUN=c(mean, sd, min, max),
> na.rm=T)
> 
> However, with tapply I know I cannot use length to get a count where
> there are NAs. summaryBy appears to work the same way. I do know how to
> get a count of the entire column using sum:
> sum(!is.na(lh.forgotten$RLWTEST))
> 
> However, this does not give me a count split up by my factor (STATUS). I
> have looked through Daalgard (2002) and Verzani (2005), and have
                      ^^^^^^^^
                      Ahem!....
> searched the help files, but with no luck.
How about

with(lh.forgotten,
  tapply(!is.na(RLWTEST), STATUS, sum) 
)

or maybe just 

table(STATUS[!is.na(RLWTEST)]) 
 
> Thank you in advance for your help. I love R and am interested in making
> it more accessible to social scientist types like me. I know it can do
> everything SPSS can and more, but sometimes the very simplest things
> seem to be a lot harder in R.

-- 
   O__  ---- Peter Dalgaard             ?ster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907