THE ROLE OF DATA AND THEORY IN COVARIATION ASSESSMENT - IMPLICATIONS FOR THE THEORY-LADENNESS OF OBSERVATION

The issue of the theory-ladenness of observation has long troubled phi losophers of science, largely because it seems to threaten the objecti vity of science. However, the way in which prior beliefs influence the perception of data is in part an empirical issue that can be investig ated by cognitive psychology. This point is illustrated through an exp erimental analogue of scientific data-interpretation tasks in which su bjects judging the covariation between personality variables based the ir judgments on (a) pure data, (b) their theoretical intuitions about the variables, or (c) both data and prior theoretical beliefs. Results showed that the perceived magnitude of correlations was greatest when subjects relied solely on theoretical intuitions; that data-based jud gments were drawn in the direction of those prior beliefs; but that ex posure to data nonetheless moderated the strength of the prior theorie s. In addition, prior beliefs were found to influence judgments only a fter a brief priming interval, suggesting that subjects needed time to retrieve their theoretical intuitions from memory. These results sugg est ways to investigate the processes mediating theory laden observati on, and, contrary to the fears of positivist philosophers, imply that the theory-ladenness of observation does not entail that theoretical b eliefs are immune to data.