A robust regression model must be evaluated based on multiple criteria: 'goodness of fit' (how well the line fits the data), 'economic plausibility' (whether the relationship makes sense in a business context), and the 'significance of independent variables' (whether the drivers actually impact the cost). All these are essential for reliable analysis.

Serial correlation, also known as autocorrelation, occurs in regression analysis when the residuals (errors) are not independent but show a pattern over time or across observations. This violates the assumption of independent errors in ordinary least squares regression, potentially leading to biased standard errors and unreliable hypothesis tests. Both terms are used interchangeably to describe this phenomenon in statistical modeling.

Regression analysis is a powerful statistical tool used to model the relationship between a dependent variable and one or more independent variables. It quantifies the impact of changes in the independent variable on the dependent variable by calculating the slope coefficient. This method is widely used in accounting and finance to estimate cost behavior, forecast future trends, and evaluate the strength of relationships between various financial and operational metrics.

The disturbance term (or residual) in regression analysis is calculated as the difference between the observed value and the predicted value (Error = Observed - Predicted). In this case, 50 - 7 = 43. This term represents the portion of the dependent variable that the regression model cannot explain, capturing the random variation or noise inherent in the data set.

The coefficient of determination, denoted as R-squared, measures the proportion of the variance in the dependent variable that is predictable from the independent variable. It provides a statistical assessment of how well the regression model fits the observed data. A higher R-squared value indicates that a larger percentage of the variance is explained by the model, suggesting a better fit.

Multiple regression is a statistical technique used to model the relationship between a single dependent variable and two or more independent variables. By including multiple predictors, the model can account for various factors influencing the outcome simultaneously, providing a more comprehensive analysis of complex business scenarios where a single variable is insufficient to explain the variance in the dependent variable.

Homoscedasticity refers to the assumption in regression analysis that the variance of the residuals is constant across all levels of the independent variable. If the variance changes, it is known as heteroscedasticity. Maintaining constant variance is essential for the validity of ordinary least squares (OLS) regression estimates, as it ensures that the standard errors of the coefficients are reliable for hypothesis testing and confidence interval construction.

The coefficient of determination (R-squared) is calculated as 1 minus the ratio of unexplained variation to total variation. Here, R-squared = 1 - (255,050 / 550,505) = 1 - 0.4633 = 0.5367. This value represents the proportion of the variance in the dependent variable that is predictable from the independent variable.

The observed cost is the actual cost incurred, which is the sum of the predicted cost (from the regression model) and the residual error (the difference between actual and predicted). By adding the predicted value of 20 to the residual error of 35, we arrive at an observed cost of 55. This calculation confirms the relationship between the model's output and the actual financial data.

Heteroscedasticity occurs when the variance of residuals or error terms is not constant across different levels of an independent variable. This violates the homoscedasticity assumption required for many statistical tests, such as ordinary least squares regression, potentially leading to inefficient estimates and biased standard errors in cost analysis.