Reliability and validity are concepts commonly used in evaluating psychometric tests. Reliability is a measure of the internal consistency of a test, whilst validity is concerned with the strength of the conclusions and inferences drawn.
Internal consistency is highest where test items are repeated, but this narrows the focus of the test overall. Rather than repeating questions, or introducing items which are virtually identical, e-interplace® (the software which analyses the Belbin Self-Perception Inventory) seeks for clusters of related behaviour. For example, the Shaper cluster refers to an individual who is challenging, competitive, hard driving, tough and outspoken. However, that does not mean to say that everyone who is competitive will necessarily be outspoken.

Most psychometric tests rely on self-reporting. However, the behaviours identified may not correspond with what others observe. The strength of the software, e-interplace®, rests in its emphasis on construct validity: using multiple sources of evidence to draw a conclusion. The system’s outputs are designed to take account of the degree of consensus on observed behaviour. Disparities between self-analysis and the perceptions of others can provide valuable leads for action.
Formal correlations are, however, difficult to calculate, as those providing feedback are not required to make a fixed number of responses. This is because genuine responses are more easily obtained – and more valuable – when forced choices are avoided.