Statistics: endogeneity

Tutoring statistics, definitions are of interest. The tutor mentions endogeneity.

endogeneity: when an explanatory variable affects the error term in a predictable way.

An example would be a model that measures the amount of daily sunlight received, but assumes (wrongly) that the days are the same length throughout the year. Such a model will notice that the sunlight hours will be increased for the summer days, even though it predicts they should have the same mean sunlight. The increase, being unexplained, will go to the error term. Such a flaw in the model results in endogeneity: the time of year influences the error term because the model does not account for its impact.

Source:

towardsdatascience.com