Loading...
Please wait, while we are loading the content...
Similar Documents
Log-linear models and goodness-of-fit statistics for categorical data in contingency table analysis
| Content Provider | Semantic Scholar |
|---|---|
| Author | Brouwer, Floor Nijkamp, Peter |
| Copyright Year | 1983 |
| Abstract | Categorical data may be classified into a dichotomous and polychotomous data. In the present paper the use of such qualitative data will be discussed in the context of contingency table analysis. A log-linear model and underlying statistical hypotheses are related to each other, especially in regard to nested or hierarchical models. The goodness-of-fit of different hierarchical models is judged by means of some asymptotic chi-squared test-statistics, viz. Pearsons test-statistic and the likelihood ratio-test. A new issue in this paper is the discussion of the problem of a priori or structural zeros-in addition to sampling zeros-in log-linear analysis. This paper proposes a solution to deal with this problem. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://dare.ubvu.vu.nl/bitstream/handle/1871/11800/Scanjob%20198300003.pdf?isAllowed=y&sequence=1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
