Holistic procedures for contemporary formative construct validation using PLS: a comprehensive example

Sound advice from knowledgeable methodologists synthesised from the literature, has yielded herein a complete and readily useable set of guidelines for best-practice contemporary formative construct validation using partial least squares (PLS). Procedures are logically organised around the indicator level and construct level validation. Indicator level validation is concerned that each indicator contributes to the formative construct by carrying the intended meaning. Tests advocated are in attention to: 1) multicollinearity; 2) indicator weights; 3) indicator loadings; 4) indicator criterion validity. Construct level validation includes tests of: 1) nomological validity; 2) external validity. The referent study conveniently included items whose statistical indications were mixed, inviting illuminating discussion around the basis on which key decisions are made. Guidance compiled herein and demonstrated through example, will be of value to both novice and more experienced researchers working with or considering formative phenomenon. The combination of procedures outlined makes clear the necessary data and tests required in order to facilitate strong formative construct validity testing.