search

simsem / semTools

Useful tools for structural equation modeling
75 stars 36 forks source link

reliabilityL2() without reference #35

Closed jorgesinval closed 6 years ago

jorgesinval commented 6 years ago

Can someone provide a reference of published paper regarding the realibilityL2 (L1, L2, L1 partial) estimates? Otherwise, I will have to cite the semTools package itself.

TDJorgensen commented 6 years ago

These showed up on the first page of a Google search for "omega reliability higher order factor", so you might be able to turn up more with a deeper search or looking through their reference list:

http://mjesec.ffzg.hr/revija.psi/vol%2021%20no%202%202014/Wahyu%20Widhiarso%201.pdf

http://personality-project.org/revelle/publications/zinbarg.revelle.pmet.05.pdf

They discuss hierarchical omega, but I don't know specifically about the partial vs. others. Not sure what reference Sunthud had read when he wrote this function, but the differences are just a matter of what is in the numerator and what is in the denominator. Are you asking how much of the total observed variance is explained (L1, L1-partial) or how much of the first-order factor variance is explained (L2) by the higher-order factor? If the former, are you asking how much is explained by all factors (L1) or uniquely by the higher-order factor (L1-partial)?

It's similar to expressing multiple types of ICC at each level of analysis when you have more than 2 levels (e.g., students nested in classes nested in schools): You can ask how much class variance is explained by school differences (like L2), or how much student variance is explained by school and class differences (like L1), or how much student variance is explained by either school or class differences alone (like L1-partial).

If you decide to cite the package, there will be a new version on CRAN soon, with updated citation("semTools") output.

jorgesinval commented 6 years ago

Thanks! I think that I will opt for the semTools citation, I used it already in my publications. So there will be no more "semTools Contributors" as authors? I saw that the citation now suggests to you, Pornprasertmanit, Schoemann, and Rosseel.

TDJorgensen commented 6 years ago

Yes, I felt that "semTools contributors" unfairly equated the level of contributions between authors/maintainers and occasional outside contributions (which are very much appreciated, but often need further revision by maintainers before inclusion in semTools). All contributors are still listed in the DESCRIPTION page, on CRAN, and on the semTools wiki.