Note:
To protect anonymity, the variables age and gender were deleted from all data files.
Relevant information can be seen in the output file.
Exelsheets contain data on page 1, variable information and/or
the codebook on the following sheet pages.
zip-folder *Study 1*
**01_Pre-Analyses**
Statistical analyses were computed with SPSS 26 (IBM Corporation, Armonk, NY, USA).
The sample was randomly split into two subsamples (subsample 1 and subsample 2).
The two subsamples did not differ with regard to age, gender, level of education and relationship status
(see Output_Study1_pre-analyses).
**02_CFAs**
Further statistical analyses were calculated using R version 4.0.4 (R Core Team, 2019).
Using the Vuong test (Vuong, 1989), we planned to compare the CFAs of the two- and three-scale version,
as well as Halberstadtâ€™s postulated four-scale version against each other within the first subsample (subsample 1).
Then, CFAs were conducted with maximum likelihood estimation with robust (Huber-White) standard errors
and a scaled test statistic equal to Yuan-Bentler test statistic (Raykov, 2012).
The item assignment to the scales followed Halberstadt (1986).
Items with insufficient loadings (i.e. <.32; Worthington & Whittaker, 2006) were deleted in subsample 1.
For subsample 2, we again calculated model fits by means of CFA for the final item solution of the German version.
Please note:
In revision 1, we conducted scalewise CFAs with maximum likelihood estimation and Satorra-Bentler scaled (mean adjusted) test statistics
for testing unidimensionality of the scales (Raykov, 2012).
This approach also favored the three-factor solution with the scales
positive (P), negative-dominant (ND) and negative-submissive family expressiveness (NS).
The items belonging to the P and ND scale remained the same.
But, for the NS scale, item 34 was eliminated. Therefore, all analyses including the NS scale were re-analysed for revision 2.
This applies to ESM 2 showing item parameters from subsample 2 (study 1) and
to all intercorrelations with the NS scale (study 2).
The syntax and output files of revision 1 can be found here: http://dx.doi.org/10.22029/jlupub-11.
zip-folder *Study 2*
Statistical analyses were implemented using SPSS 26 (IBM Corporation, Armonk, NY, USA).
We calculated Pearson's r to investigate the validation of the FEQ-GR scores of
the three scales (positive, negative-dominant and negative-submissive family expressiveness).