Substantive nonadditivity in social science research A note on induced collinearity and measurement and testing of effects
Theories of social behavior often refer to nonlinear or nonadditive relationships among relevant variables. While nonadditive features are theoretically important, the inclusion of quadratic or multiplicative terms in structural equation models can cause significant methodological problems, which in...
Saved in:
| Published in: | Quality & quantity Vol. 22; no. 3; pp. 221 - 237 |
|---|---|
| Main Authors: | , |
| Format: | Journal Article |
| Language: | English |
| Published: |
01-01-1988
|
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Theories of social behavior often refer to nonlinear or nonadditive relationships among relevant variables. While nonadditive features are theoretically important, the inclusion of quadratic or multiplicative terms in structural equation models can cause significant methodological problems, which include collinearity, ill-conditioning of input data, & precision of parameter estimates; these are illustrated with a model of differential mortality estimated on data from 3,069 US counties. Methods are presented for measuring & testing the statistical significance of effects of explanatory variables in nonlinear models. Collinearity may not be as big a problem for linear structural models in the social sciences as is often believed; while it is increased by adding quadratic or multiplicative terms, its effects are localized & entail only variables with a common base. These findings suggest that the substantive insights gained from including theoretically appropriate nonlinear & nonadditive terms in a model may outweigh the methodological problems they create. 3 Tables, 16 References. Modified HA |
|---|---|
| Bibliography: | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ISSN: | 0033-5177 1573-7845 |
| DOI: | 10.1007/BF00183538 |