z-logo
open-access-imgOpen Access
Making good use of numerical predictors: alternatives to categories or straight lines
Author(s) -
Wells J. Elisabeth
Publication year - 1998
Publication title -
international journal of methods in psychiatric research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.275
H-Index - 73
eISSN - 1557-0657
pISSN - 1049-8931
DOI - 10.1002/mpr.35
Subject(s) - categorization , spline (mechanical) , psychology , novelty , regression , anxiety , novelty seeking , regression analysis , personality , clinical psychology , econometrics , big five personality traits , psychiatry , mathematics , statistics , social psychology , computer science , psychotherapist , artificial intelligence , engineering , structural engineering
Numerical predictors such as age, scores or scales are often used in analyses that assume linearity, which may be unjustified, or they are categorized, which results in loss of power. Restricted cubic spline regression is an alternative technique without linearity assumptions or the problems associated with categorization. This paper describes restricted cubic spline regression models, following Durrleman and Simon. Two examples are given, one from psychiatric epidemiology and one from clinical research. The first example models the probability of marriage in relation to age, for women with or without antisocial personality disorder. The second example looks at the relationship between anxiety disorder and novelty seeking in depressed patients. Curvilinear effects do occur in psychiatry and it is important to know how to detect them. Copyright © 1998 Whurr Publishers Ltd.

The content you want is available to Zendy users.

Already have an account? Click here to sign in.
Having issues? You can contact us here