Linearity in Log-Linear Analysis
Bibliographic Data
| ID | 11010668 |
|---|---|
| Authors | Arthur L Stinchcombe (corresponding author) |
| Year | 1983 |
| Volume | 14 |
| Pages | 104 |
| Publication date | 1983-01-01 |
| Open Access | No |
| Type | ARTICLE |
| Venue | Sociological Methodology (JOURNAL) |
| Journal identifiers | ISSN: 0081-1750 • E-ISSN: 1467-9531 |
| Publisher | JSTOR (PUBLISHER) |
| DOI | 10.2307/270904 |
| OpenAlex | W2125242068 |
| Language | EN |
| Citations received | 5 |
| References cited | 6 |
Quite often when we analyze attitudinal data with log-linear methods, we run into difficulty with groups who are predicted to be toward one extreme or the other of an attitude dimension. Formally, the difficulty appears as a higher-order interaction effect. It sometimes takes the form of extreme groups having a higher probability of answering positively than predicted from the rest of the table (for example, producing an excess of consistent responses; compare Duncan, 1979, who calls this an excess of ideological responses, and Stinchcombe and others, 1980, p. 86). But it fairly often takes the form of even extreme causal forces still not producing 100 percent assent (or denial), so that, for example, only 90 to 95 percent of women in white non-gun-owning families living in large cities near ghettos endorse gun registration
Econometrics · Linear model · Linearity · Log-linear model · Statistics · Economics of Agriculture and Food Markets · Engineering · Mathematics · Social and Intergroup Psychology · Survey Methodology and Nonresponse
| Citations per year | 0.12 |
|---|---|
| Citation span | 1985 - 2010 (26) |
| Highly cited | No |