Glossary

Polychoric Correlation

An estimate of the correlation between two continuous latent variables underlying a pair of observed ordinal variables. It assumes that the observed ordinal categories arise from thresholding bivariate normal latent variables. Polychoric correlations are used in factor analysi...

Definition

An estimate of the correlation between two continuous latent variables underlying a pair of observed ordinal variables. It assumes that the observed ordinal categories arise from thresholding bivariate normal latent variables. Polychoric correlations are used in factor analysis and structural equation modelling when inputs are ordinal.

Why It Matters

Standard Pearson correlation assumes continuous, normally distributed data and can underestimate associations between ordinal variables. Polychoric correlation recovers the underlying continuous relationship, producing more accurate estimates that improve downstream modelling — particularly in psychometrics, survey analysis, and any field relying on Likert-scale data.

Example

A researcher investigating the relationship between job satisfaction (measured on a 5-point Likert scale) and organisational commitment (also on a 5-point scale) computes a polychoric correlation. The resulting coefficient (0.62) is higher than the Pearson correlation (0.48), more accurately reflecting the latent continuous association.

Related Terms

Software Notes

  • SPSS: Not built-in; use the R plugin or the POLYCOR extension
  • R: polychor() from the polycor package
  • Stata: polychoric command (install via ssc install polychoric)