Using fidelity of implementation data in randomized controlled trials: A Primer on using instrumental variables in educational research

Although researchers investigating randomized controlled trials (RCTs) may collect fidelity of implementation data, this information may not be optimally used in analyzing data to yield causal effect estimates. An approach to obtaining causal effects, though underutilized by educational researchers, uses instrumental variable (IV) estimation. Often taught in introductory econometrics classes, the concept of IVs is rarely covered in basic graduate-level quantitative methods classes in education. This tutorial explains why and how instrumental variables work, differentiates compliance types, illustrates how IVs can be used to properly deal with issues related to noncompliance and dosage effects, provides R syntax to estimate models using two-stage least squares regression, structural equation modeling, and Bayesian regression, and we include a sample writeup. We do so in a way without equations or specialized notation to support more widespread understanding and use in RCTs with issues of noncompliance.