On spurious correlation in geographical problems.
The writer discusses the dangers of spurious correlation in both correlation and regression analysis for researchers in human and physical geography. He argues that spurious correlation occurs when variables used in the calculation of a correlation coefficient are in a functional relationship with...
| Publicado en: | Canadian Geographer Vol. 40; pp. 194 - 203 |
|---|---|
| Autor principal: | |
| Formato: | Artículo |
| Publicado: |
Wiley-Blackwell
Fall 1996
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=512866456&site=ehost-live header: @attributes: shortDbName: ssf uiTerm: 512866456 longDbName: Social Sciences Full Text (H.W. Wilson) uiTag: AN controlInfo: bkinfo: jinfo: jid: 00083658 CGE jtl: Canadian Geographer issn: 00083658 maglogo: N pubinfo: dt: Fall 1996 vid: 40 pid: 480 pub: Wiley-Blackwell artinfo: ui: 512866456 10.1111/j.1541-0064.1996.tb00448.x ppf: 194 ppct: 9 formats: tig: atl: On spurious correlation in geographical problems. aug: au: Kanaroglou, Pavlos S. su: Geography -- Methodology Geography -- Statistical methods Statistical correlation Linear statistical models Regression analysis sug: subj: Geography -- Methodology Geography -- Statistical methods Statistical correlation Linear statistical models Regression analysis ab: The writer discusses the dangers of spurious correlation in both correlation and regression analysis for researchers in human and physical geography. He argues that spurious correlation occurs when variables used in the calculation of a correlation coefficient are in a functional relationship with each other. Through a discussion of simple additive functional relationships between variables, he demonstrates that spurious correlation may change the value of a correlation coefficient in a rather substantial way. In addition, he reveals that spurious correlation between the independent and dependant variable in multiple regression analysis might seriously impact on the coefficient of determination, resulting in incorrect inferences regarding the goodness-of-fit of the regression model. In conclusion, he considers the failure of previous research to deal with spurious correlation when the variables' functional relationship is other than additive and outlines the need for further research in this area. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: N holdings: @attributes: islocal: N |
|---|