If the conclusions of the previous chapter are accepted, it might be wondered why estimation would be attempted under such an analytically intractible and generally unidentified model. The explanation may be simple. In effect, there has been estimation without a model. Recall that Schmidt and Hunter (1977) sketched only an idea: no model was presented, but they nonetheless provided an estimation scheme. This same trend has continued. In fact, there are a variety of related, generally complex, estimation schemes (cf. Burke, 1984), and almost all of them are computer intensive—cut loose as it were from any formal model to guide interpretation. The consequences of this state of affairs are easily predicted: the problem both practically and conceptually is to make sense out of the estimates provided.
KeywordsEstimation Scheme Bayesian Method Negative Expectation Validity Coefficient Bayesian Procedure
Unable to display preview. Download preview PDF.