Correlation Between The And Path Analysis

777 WordsDec 4, 20144 Pages
With path analysis we are attempting to estimate and describe causal relationships through the use of correlative data. Because of this fact the degree of confidence we can have in the causal inferences drawn from the results of the analysis is bound to be much less than the confidence in inferences drawn from an experimental study. If the model accurately reflects reality, the estimated causal effects are . . . This serves as an appropriate reminder to the readers of our research reports - of the limitations associated with drawing causal inferences from correlative data. There is no empirical test that can tell you us the extent to which we have selected and described the correct model. In order to evaluate these five assumptions, Tate (1992) suggests that we focus our attention on the credibility, reasonableness, and utility of a proposed model. In other words, 1) a model should be plausible to those who are expert in the particular field of inquiry, 2) the results should be reasonable within the context of the current research literature, and 3) a model should be useful in predicting future events. The responsibility for assessing the assumptions in this manner ultimately rests with the researcher and his or her subjective judgments. The assessment of model fit in a path analysis can be accomplished only through the use of hand calculations. The assessment of model fit is conducted by obtaining reproduced correlations and comparing them to the empirical (observed)
Open Document