ABSTRACT
The fitting of log linear models is one powerful tool for exploring matters. They are computationally more difficult to handle than the corresponding linear models for approximately normally distributed data, but this is unimportant provided a suitable computer program such as GLIM is available. As with analysis of relatively complex data in general, some choice is needed between starting with fits to the whole data, attempting to achieve reasonable simplification of the most general model, in this case the sections being the different types of housing. Within each type of housing, increased satisfaction is associated with increased feeling of influence upon management. A total of 1681 residents from selected areas living in rented homes built between 1960 and 1968 were questioned on their satisfaction, the degree of contact with other residents and their feeling of influence on apartment management.
