Full score invariance of employee commitment across countries and times
Social sciences data archives Universitiy of Ljubljana Slovenia, http://www.adp.fdv.uni-lj.si/
Mesner Andolsek, Dana
Chair of Human Resources and Social Management University of Ljubljana Ljubljana, Ljubljana
Employee organisational commitment has been long and extensively studied until now (Meyer & Allen, 1997; Jaussi, 2007). An emphasis of current analysis was to verify its measurement characteristics, for the purpose of comparisons of levels of commitment across time and countries. A limited set of countries was chosen among those available in a sample from the data on Work Orientations II, ISSP 1997, purposefully selected to reflect cultural and structural differences that was expected to affect change in levels of organisational commitment. With the use of structural equations models we first confirmed that a model for configural invariance for two factors measuring conceptually distinct components of Affective commitment (AC) and Continuance commitment (CC) respectively has better support than of one factor model. Metric and error term invariance was subsequently confirmed. Scalar equivalence, needed for valid comparison of mean levels of both components of commitment, was confirmed as well, with the exception of two country specific Tau coefficient. Finally, a model thus established was applied additionally on data from 2005 ISSP. Acceptable fit was achieved for a common model containing both points in time and all countries, which allowed making more firm conclusions about the changes in AC and CC in different countries.
A procedure for testing measurement characteristics of items from comparative survey proved useful. In particular, it makes possible to compare multidimensional concept in situation when a limited number of items (4 altogether) was available, of which in 2005 data one was omitted. A robustness of procedure has to be tested by replicating analysis on extended set of countries.