Further Statistical Methods -HT05

Practical - week 5

MIM and graphical models. Assessed.

1. Start MIM
2. Download this file containing data from a study attitudes towards abortions in the USA. The individuals in the study are cross-classified according to four criteria: race, (white or non-white), sex (male or female), attitude towards legalising abortion (for, against, undecided), and age (6 age groups).
3. Read the file into MIM
4. Specify the saturated model and display its graph.
5. Specify the variable age as ordinal in MIM. Note that this may or may not be justifiable.
6. Test all conditional independences of any pair of variables, given the others. Both by using a standard LR test and the appropriate test which takes into account ordinality of some of the variables. Check also whether the asymptotic results seem to conform with Monte-Carlo p-values. Decompose the tests according to the values of the conditional variables (option z).
7. Based on the above, give a list of apparently reasonable models.
8. Find the best model using Akaike's information criterion (AIC).
9. Find the best model using Schwarz' Bayesian information criterion (BIC).
10. Test the overall fit of the models found under 8 and 9 aginst the saturated model. Comment on the results.
11. Search for acceptable models using stepwise procedures. Both forward from the main effects model (no edges in the graph, specified as o,r,a,s) and the saturated model (oras).
12. Which model is your preferred model? Why? Interpret the model in terms of conditional independence.

Back to course overview.

Last updated: Friday, 18 February 2005 09:14.  Steffen L. Lauritzen