search

aiorazabala / qmethod

R package to analyse Q methodology data
GNU General Public License v2.0
34 stars 18 forks source link

check standardization for FREE distributions #325

Closed maxheld83 closed 7 years ago

maxheld83 commented 8 years ago

This is just something I want to get back to later and check.

So if we have free distributions it is possible, that individual (or all) distributions may be :

At several places in the package, we do z-transforms, and it would be helpful to think about what that means, and how it can be appropriately done for free distros.

For example, if we wanted to retain (individual or collective) skew and asymmetry, we might want to standardize across all columns (participants), not per column (which would negate the asymmetry).

Question of course remains how this can be done in the context of PCA which requires single-centered variables (though I'm not sure whether this prejudices how the z-scoring is to be done).

it seems to me that if you're serious about a free distribution, and assume that these differences can be compared between individuals (otherwise: why bother?), you'd want to retain them in the analysis.

maxheld83 commented 8 years ago

duplicate of https://github.com/maxheld83/pensieve/issues/100