MCA and Logit: A Comparison
MetadataShow full metadata
In this paper a technique is developed which shows how both logit and MCA can be adapted for comparison purposes. The technique is developed for the special case of dichotomous dependent variables. The method was empirically tested and the data reveal that the logit probabilities, with few exceptions, were consistent with those generated by the MCA technique. The author advocates the use of this technique to confirm the MCA coefficients. Furthermore, when MCA and logit yield consistent results MCA coefficients should be used for presentation and discussion purposes.