The correlation coefficient is commonly used, yet its potential for drawing causal inferences is hardly tapped. Given even quite a simple correlation matrix, any coefficient may be readily decomposed into direct and indirect effects; joint effects with other specified variables; and coordinated error effects due to unspecified causes. A worked example relating to crime rates is given.
Get full access to this article
View all access options for this article.
References
1.
BaldwinJBottomsA E, 1976The Urban Criminal (Tavistock, London)
2.
BlalockH M, 1972Social Statistics2nd edition (McGraw-Hill, New York)
3.
ClarkJ PWenningerE P, 1962“Socio-economic class and area as correlates of illegal behaviour among juveniles”American Sociological Review27826–834
4.
DoganMRokkanS, 1969Quantitative Ecological Analysis in the Social Sciences (MIT Press, Cambridge, Mass)
5.
DuncanO D, 1971“Path analysis: Sociological examples” in Causal Models in the Social Sciences Ed. BlalockH M (Macmillan, London) chapter 7
6.
HarveyD, 1969Explanation in Geography (Edward Arnold, London)
7.
HawkesR K, 1971“The multi-variate analysis of ordinal measures”American Journal of Sociology76908–926
8.
HeiseD R, 1975Causal Analysis (John Wiley, New York)
9.
McDonaldL, 1969Social Class and Delinquency (Faber, London)
10.
ReissA JRhodesA L, 1961“The distribution of juvenile delinquency in the social class structure”American Sociological Review26720–732
11.
SimonH A, 1954“Spurious correlation: A causal interpretation”Journal of the American Statistical Association49467–479
