allow two 0 columns in import.q.sorts

[maxheld83]

@aiorazabala just suggested:

The import.q.sorts function to allow two central columns with zero. I see that this is not optimal from a research design point of view, but some people might do it too, and it's not strictly incorrect (plus this restriction requires to code some tweaks to convert the CSV files to the format that the function wants!)

[maxheld83]

just to be clear, @aiorazabala, because I don't quite understand why people would use two central columns with zero – this is merely for because data entry is easier (b/c the column height is lower), correct?

[maxheld83]

three comments:

• I agree that it is not "wrong", and there is no technical reason not to do it.
• However, import.q.sorts() currently does not look at any row to figure out the rank, but merely infers it from the index (left to right, from negative to positive). So we'd need to add a) a convention for a row in which the column value (-2, -1, 0, 0, 1, 2) etc is stored and b) an alternative way to read in q sorts in that way, which, I worry, might make things more complex and buggier.
• Unless people actually have people Q-sort items in two zero columns (which might be questionable methodologically?), I am somewhat philosophically opposed to accommodating non-raw data (which this would be).

So bottom line: I probably won't spend time doing this, but sure, why not, if people find it helpful.

[aiorazabala]

Hummm, I see the logic of the import.q.sorts() function.

1. So basically, having two 0 columns wouldn't matter as long as the bottom row indicates which are the column values, Correct?
2. if (1 == T) then the problem would occur when someone enters these Q-sorts but excludes the last line indicating column values, Correct?

[aiorazabala]

Here's the code snippet I used to fix this issue when using our 'faulty' data. mydata is a raw matrix like the one you read from a Q-sort csv.

  # Merge the 0 columns
  central <- (ncol(mydata-1))/2
  top.part <- matrix(data=NA, nrow=dim(mydata)[1], ncol=dim(mydata)[2]-1)
  top.part[,central] <- mydata[ ,central+1] 
  colnames(top.part) <- colnames(mydata[ ,-(central+1)])
  mydata <- rbind(top.part, mydata[ ,-(central+1)])

This could be either integrated inside the function (e.g. as an optional argument) or simply point at this snippet whenever someone has this issue.