Calculation of coefficients in word-variable co-occurrence network

[andraselina]

Hi all,

I've used KH Coder for word-word cooccurrence networks quite often and I understand how the relationship coefficients between words are calculated. Now I created a network to understand the association between words and an external variable (with two different values, let's say a and b). However for some words, I do not understand why they have been assigned to only a, only b or both. I investigated the total occurrence of a word with a or b, the occurrence relative to the number of documents with a or b etc. but for some words I would assign them differently than KH Coder did.

Could you maybe explain with an example how the coefficient here is calculated / how the assignment to either one of the variables or all of them works? Unfortunately I couldn't find more information about this particular network in previous discussions or the manual.

Thank you!

[ko-ichi-h]

Hi,

Please uncheck "Standardize coef" and check "Show coef" option. Then click OK to make the network and look into it.

Do you get expected results with this procedure?

With "Standardize coef" option checked, KH Coder try to "standardize" coefficients so that "a" and "b" will have roughly the same number of connected words.

Currently, this functionality is only documented in the Japanese manual. Sorry.