Measurement
This is the landing page for everything related to measuring a process. By measurement we mean the assignment of numeric values to the characteristics of your cases, thereby quantifying them. This quantification enables you to analyze and interpret data systematically. The way you measure your constructs of interests determines the meaning of the values in your data set. This, in turn, implies what analyses are meaningful and what conclusions you can draw based on your study. It is therefore crucial to carefully consider how measurement applies to your constructs of interest and what implications this has for your study design.
Importantly, measurement itself is a process that can be studied. Whether we rely on self-report, behavioral tasks, or passive sensing, our measurements are shaped by the assumptions about how they relate to the psychological process we want to investigate. For example, if you use self-report, the questionnaire design typically is based on your ideas about how people will interpret and respond to your measurement instrument. In contrast, passive sensing methods (e.g., heart rate monitoring) do not involve direct psychological processing by participants during measurement but still require theoretical assumptions about how the observed signal relates to psychological constructs like stress.
Making such assumptions explicit in a measurement theory can help you critically assess them, share them with others, and refine them as needed. You could also consider incorporating measurement models into your data analysis. These can help you evaluate your measurement theory, and, for example, prevent measurement errors from warping your results and conclusions about your constructs of interest.
A first step in your measurement design is deciding how you will select a subset of [cases] (e.g., persons) from your population and a subset of occasions (i.e., temporal lens), that is, creating your sampling design. Another key decision is your choice of measurement instruments and the design of those instruments, for example whether your will use self-reports, observations, or passive sensing. In case you will use a questionnaire, an important part of its design will be selecting and formulating its items, including the response options. Finally, it is essential to validate your measurements, that is, to evaluate if you are indeed measuring what you intended to measure. Without proper validation, data may be misleading or unreliable. Below, we have specified key topics to think about with respect to measurement.
Themes
Measurement theory
The theory that explains how the true scores of your construct lead to the observed scores obtained through your measurement instrument.
Measurement modeling
Modeling and analyzing how the true scores of your construct result in the observations you make with your measurement instrument.
Sampling design
The way you select a subset of cases (e.g., persons) from your population and a subset of occasions (i.e., your temporal lens).
Instrument design
Choosing between self-report, observing or sensing; and designing the instrument, for example, the item order and length of a questionnaire.
Item design
The way your questionnaire items are designed, from the questions you ask to the answer options you provide.
Validation
How to evaluate whether your measurement instrument is measuring what it is supposed to measure.