How to Tie a Construct to Indicators: Guidelines for Valid Measurement

Invalid measurement of constructs in information systems research often remains undetected and can lead to false conclusions. The prescriptive literature on measurement has led to a better understanding of the sources of error in various areas, including conceptual modeling, common method bias, and estimation procedures. It has also called for heterogeneity in indicators to overcome sources of error associated with each indicator specifically. It has not led, however, to widespread measurement practice that takes these separate insights into account. This paper aims to facilitate this by integrating insights from the literature. It complements extant guidelines on the development of measurement with a typology of the ways to tie a construct to its indicators. It demonstrates the recommendations with an empirical illustration. This, I hope, will lead researchers adopt more heterogeneous indicators, allowing them to measure their constructs with better confidence in validity.