Schult, J., Hell, B., Päßler, K., & Schuler, H. (2013). Sex-Specific Differential Prediction of Academic Achievement by German Ability Tests. International Journal of Selection and Assessment, 21(1), 130-134. doi:10.1111/ijsa.12023
Johannes says: "A cautionary tale of messy syntax and automation..."
Data analysis finished: 2011-05-23
Article published: 2013-02-07
Replication files published: 2013-06-05
Software used: IBM SPSS Statistics 18, R 2.13.1
Below you find the syntax files used for the complete analysis. The individual steps are listed in master.sps.
The analysis itself wasn't that complicated, but I made the attempt to minimize the number of manual steps between SPSS and the manuscript word processor file. While I think I succeeded, the syntax became a monster thanks to the heavy use of the OMS command.
The following syntax files read in three raw data files and create an interim data set for each of three samples. These new data sets contain the same variables in order to run the same differential prediction analsis in each sample without changing commands. Here it gets messy. NB: The raw data is not available online due to data privacy laws.
So what analyses are run? The usual descriptive statistics and group comparisons (t-tests, see Table 1 in the manuscript), correlations (predictive validities) along with the overall r from multiple regressions (Table 2), and finally moderated multiple regressions (see Table 3).
Please visit http://www.jutze.com/research for more materials on publications and possibly newer versions.
Written by Johannes Schult (firstname.lastname@example.org)
Last updated: 2013-06-05