However, too few primary studies were available to estimate all requisite cells in the meta-analytic matrix. Likewise, the scarcity of primary studies kept us from including another potentially important element of the criterion space: counterproductive behavior (Bennett & Robinson,2000; Collins & Griffin, 1998). We expect future research to demonstrate that these behaviors would fit well under the effectiveness umbrella, showing systematic patterns of covariance with other job behaviors and further improving the predictiveness of overall job attitude. Altogether, the addition of job attitudes and work behaviors to this scheme is indicative of what we would term an attitude-engagement model of personal evaluations and individual actions taken toward one’swork role. Models D and E above are integrative forms of an overarching attitude-engagement model.Another choice we made in building our alternative models was to not estimate the residual correlation between organizational commitment and turnover.
.