Next Steps with Correlations: Scale Development

(Essay found in Nesselroade & Grimm, 2019; pg. 548)

Nesselroade & Grimm, 2019

A common activity for many academic psychologists is the construction of measuring tools. There are literally hundreds of different psychological traits, tendencies, and abilities that psychologists are interested in measuring; from commonly used concepts like extroversion and neuroticism to less frequently-referenced concepts like humility (e.g., Rowatt et al., 2006) and right-wing authoritarianism (e.g., Mirels & Dean, 2006). The scales used to measure these attributes, however, need to be created. They do not appear out of thin air.

Scale development is usually an extensive process.  First, the concept is carefully defined, with a lot of thought given to identifying various sub-components of the concept (e.g., is extroversion marked by being very talkative, striving to be the center of attention, enjoying meeting new people, being very physically demonstrative, all four?). Second, researchers typically start to gather data to see if their understanding of the concept fits well with how people answer questions about themselves. This usually involves the generation of numerous response items, oftentimes presented in the form of a question or a statement to be agreed or disagree with by participants using a Likert scale. These responses are then statistically analyzed.

This is where the correlation concept comes in. By looking at the size and nature of the relationships between pairs of items, researchers can gain feedback information regarding the nature and scope of the concept they are studying. For instance, if high correlations were found between individuals responses regarding questions related to how talkative a person is, how much they enjoy being the center of attention, and how much they like meeting new people, this inter-relatability would provide statistical evidence, in the form of shared variance, that the concept of “extroversion” encapsulates all of these sub-components. If, on the other hand, the responses of individuals to questions related to being physically demonstrative do not tend to correlate highly with the responses regarding these other components, this would provide statistical evidence that this component is not necessarily a part of the concept “extroversion.” Once data has been gathered and analyzed, this two-step process of concept definition/clarification and data gathering/analysis can repeat itself; often several times. The process, as one might imagine, is actually much more sophisticated than what has been presented here. It is described in simple terms to show the relationship between the concept of correlation and this important professional activity. Correlations, by the way, are also the base concept behind other sophisticated analytical techniques (see Box 16.3).

Find this and other similar side-bar discussions in the Nesselroade & Grimm textbook.

Mirels, H. L., & Dean, J. B. (2006). Right-wing authoritarianism, attitude salience, and beliefs about matters of fact. Political Psychology, 27(6), 839-866.

Rowatt, W. C., Powers, C., Targhetta, V., Comer, J. Kennedy, S. & Labouff, J.  (2006). Development and initial validation of an implicit measure of humility relative to arrogance. The Journal of Positive Psychology, 1(4), 198-211.