Nomological Networks

The nomological network is a tool for construct validation (i.e., gathering evidence about the meaning) of psychological measures. For example, construct validation of job performance ratings by supervisors should indicate what the ratings really mean, or how accurately they reflect actual performance levels. The goal is to link observable measurements to unobservable theoretical constructs.

In 1955, L. J. Cronbach and P. E. Meehl described the nomological network as a system of intertwined laws that make up a theory and stated that the laws in the network should generate testable predictions. Laws could relate measurements to each other (e.g., linking job performance ratings to scores on ability or personality measures), theoretical constructs to observed measurements (e.g., linking a rating of some aspect of job performance, such as effort, to the construct of effort), or constructs to other constructs (e.g., linking the construct of job effort to the personality construct of conscientiousness). Building a nomological network involves thinking about what construct is (or should be) measured by an instrument, what other constructs should be related to that construct, and what other measures should be related to the instrument of interest.

Psychological constructs are generally not directly observable, so it is not usually possible to directly determine how well a measure reflects the intended construct. Research based on a nomological network can provide indirect evidence of validity by demonstrating how well the measure correlates with other measures it should theoretically relate to. Confirmation of relationships predicted by the network sup-ports the construct validity of a measure, whereas failure to confirm predictions leads to doubt about construct validity. (Note: A complication with failures of confirmation is that they could result from poor validity of the measure or from incorrect theory in the nomological network, or both.)

An example of a nomological network involves job performance. In recent years, industrial/organizational psychologists have theorized and researched a distinction between performance on required, job-specific tasks (task performance) and performance of behaviors that are less likely to be required and not specific to particular jobs (e.g., helping coworkers; doing things that need to be done but are not assigned to particular workers); this latter type of behavior has been referred to as organizational citizenship. If supervisors are asked to evaluate workers on both their task performance and citizenship, we might ask whether the ratings really adequately distinguish between the two performance constructs (i.e., we might question the ratings’ construct validity); alternatively, ratings might be subject to a halo effect, in which a general impression of a worker forms the basis for each (supposedly) separate evaluation.

Testing predictions based on a nomological network could help determine the construct validity of the job performance ratings. The research question concerns a link between observed measures of performance and their constructs. This question can be addressed indirectly by examining links between the performance measures and measures of theoretically related constructs. It has been theorized that the task performance construct should relate more strongly to ability constructs, and the citizenship construct should relate more to personality (e.g., conscientiousness) and job satisfaction. There exist measures of ability, personality, and job satisfaction that have been linked to their theoretical constructs. These measures can be used to test predictions from the nomological network. Research has provided some but not overwhelming support for these links, which raises questions about the validity of the performance ratings and/or the theory. For examples of how nomological networks can be applied to personnel selection measures, see the sources in the References: section, which follows, by G. V. Barrett and J. F. Binning.

References:

  1. Barrett, G. V. (1992). Clarifying construct validity: Definitions, processes, and models. Human Performance, 5, 13-58.
  2. Binning, J. F., & Barrett, G. V. (1989). Validity of personnel decisions: A conceptual analysis of the inferential and evidential bases. Journal of Applied Psychology, 74, 476-494.
  3. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281-302.