When developing our measures, we take a number of steps to ensure they are the most representative of the construct under investigation. We begin with the broad construct (characteristic or attribute) that we would like to measure – for example, the concept of Sense of Belonging or Anxiety – then we conduct a thorough examination of the existing literature and consult with subject matter experts. Following that, we narrow these broad concepts into a working definition that will fit the context or environment we are surveying. For instance, how would we define sense of belonging within the school environment? What are the markers of anxiety for a student?
You can read The advantage of using measures to learn more about why measures are important.
A bank of questions is then developed that are intended to reflect the construct. The questions are then rigorously tested, to ensure only those questions that most accurately reflect the construct are included. Questions are selected for inclusion based on a number of statistical analyses, survey taker feedback, as well as subject matter expert feedback.
You can read Items, Measures, and Constructs to learn more.