Chapter 12 Scale Development

LEARNING OUTCOMES

• Define measurement and give several examples of measurement in psychology.
• Explain what is a psychological construct and give several examples.
• Distinguish the four levels of measurement, give examples of each level, and explain why this distinction is important.
• Identify the different classification systems of psychological tests and critically evaluate their implications for scale construction.
• Name the stages and evaluate the iterative process of scale development.

Psychometrics is the science of psychological assessment. It is a subfield of methodology that studies the process of measuring psychological and social phenomena using formal mathematical models.

What do psychologists measure? Do psychologists measure physical objects or hypothetical entities? Although many classic psychological studies have measured overt behaviors (e.g., reaction time, occurrence of a behavioral repertoire), the measurement of behavior has decreased exponentially after the cognitive revolution. Nowadays, psychologists usually ask respondents to complete questionnaires, inventories, and scales by clicking on a computer screen or by ticking a box in paper-and-pencil psychological tests.

Psychologists tend to measure unobservable variables that are operationalized in various ways. Thus, a psychological construct (i.e., a latent variable) is a theoretical psychological entity that cannot be measured directly. To measure latent variables, we need to develop instruments that will include items or tasks devised to measure the underlying psychological construct. Consequently, the observed variables (i.e., measures) are assumed to be probabilistic functions of underlying unobservable variables (i.e., latent variables).

Latent and observed variables

• Guided by a theory, specify what do you want to measure. For example, if we are interested in measuring the psychological construct extraversion, we need to define theoretically the limits of the construct before deciding how to measure it using observed variables
• Make decisions on the purpose of the test and on how you will assign numbers (i.e., scaling) to the observed variables. Should we construct categorical, ordinal, or interval scales? Should we use a knowledge-based or a person-based questionnaire? Should we use norm-referenced or criterion-referenced tests?
• Generate a psychometric model of the data to estimate the latent variables. We can use data-reduction techniques (e.g., exploratory factor analysis) guided by theory. We also need to evaluate how reliable is our psychometric model
• Model the structure of the latent variables to check whether our data fit the model (i.e., the theory) or not. We can also compare between different models to assess which one provides a better fit

How can we measure psychological constructs such as intelligence, neuroticism, psychological well-being, students' satisfaction, academic performance, attitudes towards refugees and migrants, or depression? We must transform observations into data and to apply formal mathematical models to assign numbers to subjects' attributes relying on their responses or on the stimuli observed. In sum, we need to assign numbers to observed psychological phenomena and to unobserved constructs, while assessing the fit of theoretical models to data.