heteroscedasticity
|het-er-o-sce-das-ti-ci-ty|
🇺🇸
/ˌhɛtəroʊskɪˈdæstɪsɪti/
🇬🇧
/ˌhɛtərəʊskɪˈdæstɪsɪti/
unequal dispersion (nonconstant variance)
Etymology
'heteroscedasticity' originates from Greek roots via modern scientific/Neo-Latin coinage: 'hetero-' from Greek meaning 'different' and a form related to Greek 'skedasis/ skedastos' meaning 'scattering' or 'dispersion', combined with the English suffix '-ity'.
'heteroscedasticity' was formed in 20th-century statistical English from the adjective 'heteroscedastic' (also seen as 'heteroskedastic'), which itself was coined from the Greek elements 'hetero-' + 'skedast-'. The term entered technical use in econometrics and regression analysis to name nonconstant variance.
Initially a literal compound meaning 'different scattering/dispersion', it evolved into the technical statistical sense 'non-constant variance of errors or observations' used in regression and related analyses.
Meanings by Part of Speech
Noun 1
in statistics, the condition in which the variance of the errors (or of a variable) is not constant across observations or across levels of an explanatory variable; unequal dispersion of residuals (commonly encountered in regression analysis).
The regression residuals exhibited heteroscedasticity, so we reported robust standard errors.
Synonyms
Antonyms
Last updated: 2025/12/10 15:17