decent work and employees' innovative behaviour
Description
This data set is a revised version of the earlier published one entitled as V1 and V2. Several variables are dropped and the structural models are made more parsimonious. The data set contains 529 supervisor-employee dyads collected from XL, a large state holding company group with over 20 independent business units. The data is used to test the effects of Decent Work (DW) on Employees' Innovative Behaviour (EIB) through multiple mediation of Job Self-Efficacy (JSE), Intrinsic Motivation (IM), and Work Engagement (WE). SEM analysis tested that DW promotes EIB through two paths: the "expertise & self-efficacy" path and the "intrinsic motivation & engagement path". Along the former path, DW promotes EIB via mediation of JSE, with education (EDU) and company tenure (TENURE) conducive to JSE. Along the latter path, decent work promotes EIB via the serial partial mediation of IM and WE. The uploaded data set contains demographic data of the correspondents, value of each manifest items of the variable measures, and processed (e.g. by summation or average) value of variable measures. Following Farmer, Tierney, and Kung-Mcintyre (2003, p. 623), we created three manifest indicators for each latent variable by randomly assigning items to composites to reduce the parameters assessed, thus improving the ratio between sample size and estimator. Thus, for the rearranged manifest items, the letter "r" was added to the end of its variable name. In other words, items ended with "r" are inputs of SEM analysis. Along with the original dataset, we've also uploaded the data coding and major AMOS working models of the SEM analysis. Reviewers can reproduce results by directly running the models with AMOS. For quick reviewers who have not installed AMOS or feel bothered to run the software by themselves, the author also uploaded the AMOS output of hypotheses testing pasted in word documents.
Files
Steps to reproduce
Descriptive Statistics, correlations, and validity of variable measures could be reproduced by the manifest items displayed in dataset. SEM analysis results could be reporduced directly using the uploaded AMOS working models with dataset as the inputs