Assess the Quality of Your Instrument

How good and reliable is your latent variable? Well, that depends on the quality of the questionnaire (and of course, the sample). In many instances we found that our research students tend to take questions from existing literature. This is great as long as the questions come from well established instruments which has been tested and replicated in many other research projects. Sometimes, the (latent) variables you want to investigate is unique to your study and thus, exact questions are not readily available; or you are at an exploratory stage of your work where the light at the end of the tunnel that you are driving into is not yet visible. In such situations, you need to craft your own questions to tap into the dimensions of the latent variables you are investigating. Regardless of the situations you are in, it is always a good idea to test if the instrument is valid and reliable. In this chapter, we explain to you how you can test the reliability of your questionnaires through the Cronbach’s alpha (don’t worry with this technical term. It is not as scary as it sounds). We also demonstrate Fornell-Larcker criterion and average variance extracted (AVE) methods to check for the validity of your instrument. We also mention briefly factor loading and crossloading from factor analysis chapter that you also need to check to add more credence to the validity of the results. It sounds a lot, but it gives you confidence to start driving knowing that your car is reliable.