Goodness-of-fit is a statistical analysis technique used to determine how well a statistical model fits a set of observations. It evaluates the discrepancy between observed data and the values expected under the model in question, helping to assess the model's validity and predictive power.