Having coded and edited the data, we should now have a "clean" datset ready for analysis. The analysis phase is seen by some as a welcome relief after the tedium of editing the data, in that it allows the investigator to show some creative flair. Two types of analysis are possible. First, there may be a simple enumerative or exploratory analysis which seeks to explore the contents of a dataset and to describe the datset in a number of ways (e.g. response rates, means and standard deviations of responses, frequency distributions and cross-classifications). Second, one may proceed to a more complex analysis which seeks to confirm statistical hypotheses and find causal relationships among the variables. This model-building phase is frequently the purpose of many transport-related surveys.
In this Chapter, we shall first concentrate on simple exploratory analysis and then describe some of the basic methods of building causal relationships from the data A more complete description of multivariate analysis and model building may be found in many other texts on transport modelling.
10.1 EXPLORATORY DATA ANALYSIS
10.2 CONFIRMATORY DATA ANALYSIS
10.2.1 Bivariate regression
10.2.2 Multivariate regression
10.2.3 Factor analysis
10.2.4 Discriminate analysis
10.2.5 Maximum likelihood estimation
10.2.6 Logit analysis