aiorazabala
commented
8 years ago Some notes about using Q method with large P-sets (220 Q-sorts):
- No consensus statements were found, this might be because the threshold to determine dist & cons depends mostly upon the number of flagged Q-sorts, and this number is high when the P-set is high. This may require setting a higher threshold in the equation, or finding alternative ways of comparing the statements across factors
- For the plot functions, in sum, those functions that plot Q-sorts work fine, but the result is quite cluttered. This is obvious, and can't think of any solution straight away (perhaps ), so we may leave it as it is.
- function
q.compplotandq.rotplotwork*, but because there are so many Q-sorts, it's quite cluttered (though I can't think of a solution right away) - ibid. for function
q.corrplot--- @maxheld83 What's the difference between this function and any other correlogram function? (if there isn't any, then we shouldn't include it in the package) q.loaplot()maybe use alternative ways of scatterplots for large datasets? eg http://vis.supstat.com/2012/11/to-see-a-world-in-grains-of-sand/q.scoreplot(my favourite!) works as well. however with many factors, the 'pyramids' are collapsed.
- function
*They work with 3 factors extracted, but with 6 they give an error Error: Results must be all atomic, or all data frames
Job van Exel might give me his 14xx P-set dataset soon, and I'd be nice to make sure that all the functions work and plots make sense (some will not) for such a large P-set, before he tries it himself.