Using partial least squares (PLS) for digital government research

PLS is a structural equation modeling (SEM) technique similar to covariance-based SEM as implemented in LISREL, EQS, or AMOS. Therefore, PLS can simultaneously test the measurement model (relationships between indicators and their corresponding constructs) and the structural model (relationships between constructs). However, unlike its covariance-based counterpart, PLS allows researchers to working with small samples, too many or too few variables, data from non-normal or unknown distributions, and relatively new theories in which relationships between variables are not well defined. This tutotial shows how to use partial least squares (PLS) and argues that the correct use of this technique could help to incorporate more realistic assumptions and better measurements into digital government research. It does it through a commented example of a digital government research study.