Path coefficient

From Wikipedia, the free encyclopedia

Path coefficients are standardized versions of linear regression weights which can be used in examining the possible causal linkage between statistical variables in the structural equation modeling approach. The standardization involves multiplying the ordinary regression coefficient by the standard deviations of the corresponding explanatory variable: these can then be compared to assess the relative effects of the variables within the fitted regression model. The idea of standardization can be extended to apply to partial regression coefficients.

[edit] External links and references

• www2.chass.ncsu.edu/garson/pa765/path.htm
• Bill Shipley Cause and Correlation in Biology Cambridge University Press (2000)

---