5.8.3 Tacit knowledge as a general construct

Although the kinds of informal procedural knowledge measured by tacit-knowledge tests do not correlate with traditional psychometric cognition, tacit-knowledge test scores do correlate across domains. Furthermore, the structure of tacit knowledge appears to be represented best by a single, general factor.

Wagner (1987) examined the structure of tacit knowledge inventory for managers. He performed two kinds of factor analyses on the tacit-knowledge scores of these business managers in his study. First, a principal-components analysis yielded a first principal component that accounted for 44 percent of the total variance, and 76 percent of total variance after the correlations among scores were disattenuated for unreliability. The 40 percent variance accounted for by the first principal component is typical of analyses carried out on traditional cognitive-skill subtests. Second, results of a confirmatory factor analysis suggested that a model consisting of a single general factor provided the best fit to the data. The results of both factor analyses suggested a general factor of tacit knowledge.

Similar analyses were performed on a measure of tacit knowledge for academic psychologists. Consistent with the manager study, the factor analytic results suggested a single factor of tacit knowledge within the domain of academic psychology. Wagner (1987) also examined the generalizability of tacit knowledge across domains by administering both tacit-knowledge measures (for business managers and academic psychologists) to undergraduates in his study. He obtained a significant correlation of .58 between the two scores, suggesting that in addition the existence of a general factor of tacit knowledge within a domain, individual differences in tacit knowledge generalize across domains. These findings lend support for a common factor underlying tacit knowledge; a factor that is considered to be an aspect of practical cognition.

5.8.4 Tacit knowledge and performance

Finally, we have shown that tacitknowledge measures are predictive of performance in a number of domains, correlating between .2 to .5 with measures such as rated prestige of business or institution, salary, simulation performance, and number of publications. These correlations, uncorrected for attenuation or restriction of range, compare favorably with those obtained for overall cognitive ability within the range of skills we have tested.