Explain the Gaussian Distribution and Its Limits

Explain the univariate Gaussian distribution to a technical interviewer. Give its density and parameter interpretation, describe standardization and common probability statements, explain why Gaussian models arise so often, and identify cases where assuming normality is unsafe. ### Constraints & Assumptions - Distinguish a population model from the central limit theorem for an estimator. - Explain what changes under linear transformations and sums of independent Gaussian variables. - Include practical diagnostics and robust alternatives. ### Clarifying Questions to Ask - Is the focus mathematical properties, statistical modeling, or both? - Should the discussion include multivariate Gaussians? - Are we modeling raw observations, residuals, or a sample mean? - Is exact normality required for the intended inference? ```hint Separate three claims A variable may be Gaussian, an error model may be approximately Gaussian, or an estimator may be asymptotically Gaussian; these are different statements. ``` ### What a Strong Answer Covers - Density, mean, variance, symmetry, and standardization. - The 68-95-99.7 heuristic with appropriate caution. - Closure under affine transformations and independent sums. - A correct central-limit-theorem explanation. - Skewness, heavy tails, mixtures, bounds, outliers, and diagnostics. ### Follow-up Questions - Why are zero covariance and independence equivalent for jointly Gaussian variables but not generally? - What are the maximum-likelihood estimates of mean and variance? - How would heavy tails affect a confidence interval based on a small sample? - When would a log-normal or Student's (t) model be more appropriate?