Aggregate survey data by quest-year in R

Question

I am trying to aggregate survey data (the unit of analysis is respondent) into marginals: percentage of responses for each question per year. There are 21 questions (columns). This is what I attempted but it does not get the desired output.

#Place each year into a list element
yr.list = by(data=data, INDICES = data$year, FUN = list)
summary(yr.list)

#Within each element, place responses to all questions into one vector
data.form = function(x){
  n = nrow(x)
  k = ncol(x) - 2
  id = rep(1:n, times = k)
  item = sort(rep(1:k, times=n))
  y = c(unlist(x[,3:ncol(x)]))
  out = data.frame("year"=x$year, "id"=id, "item"=item, "y"=y)
  return(out)
}

yr.list = lapply(X = yr.list, FUN = data.form)

yr.vector = do.call(rbind, yr.list)
yr.vector$occurrences = 1
yr.aggregated = aggregate(occurrences ~ year + item, data = yr.vector, FUN =       sum)

summary(yr.aggregated)

Can you be more precise about *"not getting the desired output"*? — Richard-Degenne, Mar 31 '17 at 16:04
This code produces the total number of responses for each question-year; it does not break out frequencies by response. I should have mentioned this initially, but the responses are categorical. — smorg787, Mar 31 '17 at 16:09
Please add a [reproducible example](http://stackoverflow.com/q/5963269/1217536) for people to work with. — gung - Reinstate Monica, Mar 31 '17 at 17:57

Aggregate survey data by quest-year in R

0 Answers0